VMS - What's next?

[Mack Altman III]

Each time I come to this group, I find myself less optimistic than I came here. Usually, it's to find how others have approached the same situations. However, I couldn't help but ask myself on this visit, "Will VMS be a part of the future?"

Personally, I started learning Linux around the age of 13 as that's when I started designing (not much to develop then) web sites. However, I didn't hear about VMS until 2008. There has only been one company that I have worked for that had VMS; and I never heard about VMS in college. I am aware that many of the staple industries have used the OS (e.g. finance, medical, transportation), but I am also aware that many have been converting.

It seems VSI is holding onto keeping it a closed source, and I can't help but wonder why. I say this with the greatest respect, but I feel it's truthfully killing the OS. The only reason the WWW became what it is today is the open format. Millennials are rising and they all want to "do it themselves" because they feel they can "Google it" and know everything. What's working for Linux is the ability for all of them to find that information quickly, test it out, and tinker with it.

So, why should one use VMS? Why not use Linux over VMS?

I believe the answer in the past was to choose VMS because the on staff programmer already had everything working for them and that was something they were used to. However, it's easy to prove how quick you can install a third-party API or programming language on a cloud VPS and start programming an application within minutes thanks to companies, like DigitalOcean, Linode, and Vultr.

Personally, I don't understand what benefit the VMS OS offers over Linux or Windows. I feel at least Windows provides a GUI; although, I'd enjoy Linux any day over Windows. And VMS is behind in almost every version of programming language unless you utilize a third-party installation (so again, why not Linux if there's no difference?).

For the VMS-lovers out there, we run InterSystems Cache on VMS every day. And I've been pondering for years why we haven't switched it to a Linux system. The only reason I've ever gotten is, well that's just always what we have had. So, please don't think I'm bashing the OS at all. For those that have it, I've heard their passionate about it, but that's not an atmosphere for growth. Unless the client using it grows (which in my experience people are using Linux more often now), the revenue streams I would imagine are stagnating.

[David Froble]

Mack Altman III wrote:
> Each time I come to this group, I find myself less optimistic than I came
> here. Usually, it's to find how others have approached the same situations.
> However, I couldn't help but ask myself on this visit, "Will VMS be a part of
> the future?"
>
> Personally, I started learning Linux around the age of 13 as that's when I
> started designing (not much to develop then) web sites. However, I didn't
> hear about VMS until 2008. There has only been one company that I have worked
> for that had VMS; and I never heard about VMS in college. I am aware that
> many of the staple industries have used the OS (e.g. finance, medical,
> transportation), but I am also aware that many have been converting.

At one time VMS was in education. But schools are, well, CHEAP. Offer them
something a few cents cheaper, and they'll go for it. DEC didn't help. They
were so afraid of losing revenue that they shot theirselves in the "education
foot". As a result, students learned something else.

> It seems VSI is holding onto keeping it a closed source, and I can't help but
> wonder why.

They have no choice. They do not "own" the OS. They have been licensed to
develop new versions. I'm guessing they will own what they write, but, I don't
know if that is a question that has been answered.

> I say this with the greatest respect, but I feel it's truthfully
> killing the OS.

And I respectfully disagree. You don't need to have the sources of an OS in
order to use it. Even if something you need is missing, you can add it yourself.

> The only reason the WWW became what it is today is the open
> format. Millennials are rising and they all want to "do it themselves"
> because they feel they can "Google it" and know everything. What's working
> for Linux is the ability for all of them to find that information quickly,
> test it out, and tinker with it.

I must admit that people I would have never guessed are using smart phones and
tablets to get information, do things, and such. But, they are for the most
part "users". They use what's available, they do not develop anything. And,
that's probably a good thing. Having the resources available definitely is a
good thing. Now if only I could learn to be a decent user ....

> So, why should one use VMS? Why not use Linux over VMS?

Reverse the question. Why use Linux? Only decent reason I can think of is that
it's free. Well, not if you want supported Linux.

> I believe the answer in the past was to choose VMS because the on staff
> programmer already had everything working for them and that was something
> they were used to. However, it's easy to prove how quick you can install a
> third-party API or programming language on a cloud VPS and start programming
> an application within minutes thanks to companies, like DigitalOcean, Linode,
> and Vultr.

I can start programming an application even quicker. Nor do I have to worry
about the cloud going "poof". Yes, I have hardware. Nothing wrong with that.
What do you use to get to "the cloud"?

> Personally, I don't understand what benefit the VMS OS offers over Linux or
> Windows. I feel at least Windows provides a GUI; although, I'd enjoy Linux
> any day over Windows. And VMS is behind in almost every version of
> programming language unless you utilize a third-party installation (so again,
> why not Linux if there's no difference?).

Oh, someone else has something that can hold a candle to VAX/DEC/Compaq/HP
Basic? I think I'm going to start calling it VSI Basic. Less typing.

> For the VMS-lovers out there, we run InterSystems Cache on VMS every day. And
> I've been pondering for years why we haven't switched it to a Linux system.
> The only reason I've ever gotten is, well that's just always what we have
> had. So, please don't think I'm bashing the OS at all. For those that have
> it, I've heard their passionate about it, but that's not an atmosphere for
> growth. Unless the client using it grows (which in my experience people are
> using Linux more often now), the revenue streams I would imagine are
> stagnating.

Perhaps one reason is in your words. "on VMS every day". Every day, huh? No
problems, it just works, huh? If it ain't broke, then don't fix it. What would
be the cost of at best a lateral move to Linux? If it's more than $0.00, then
what is the business case for doing so?

[JF Mezei]

On 2015-10-23 18:29, Mack Altman III wrote:
> Each time I come to this group, I find myself less optimistic than I came here.

Considering VMS was officially killed by HP and then brought back to
life by VSI, the "optimistic" meter should be on the rise.

VMS is still in intensive care, kept alive by old antiques (Itanium,
Alpha, VAX). So yeah, not much visible progress, but the engineers are
busy with a serious transplant to modern 64 bit 8086s that will allow
VMS to run on modern commodity servers.

But one has to be a bit patient before we see an 8086 heartbeat on VMS.

> It seems VSI is holding onto keeping it a closed source, and I can't help but wonder why.

Because HP has refused to hand VMS over to VSI, they only granted them a
sort of "right to use". So VSI cannot decide to open source something
which is still property of HP.

Hpefully one day, HP will just hand VMS over to VSI lock stock and
barrel and then VSI can deal directly with the IP owners for proprietary
parts of VMS that are left, or just open source the portions of VMS that
VSI has control over.

> So, why should one use VMS? Why not use Linux over VMS?

Right now ? Because one has large applications running on VMS which rely
on unique services provided by VMS and rewriting them for Unix would
require too much work.

The hope is that in the future, VMS's strengths can be modernized and
VMS will make a compelling competitor to Linux with capabilities such as
clustering where Linux is clearly inferior.

> Personally, I don't understand what benefit the VMS OS offers over Linux or Windows.

Right now, VMS is being handled by a very experienced crew of OS
software engineers. I would say that it is its biggest long term asset
because they can create quality and extremely reliable software going
forward.

These guys have shown they can architect software that doesn't have "oh
I forgot about THAT" 0 day vulnerabilities.

Whether VMS will become a niche OS like VXworks with extremely talented
crew, or whether it becomes more mainstream remains to be seen.

[rcyoung]

One of the biggest appeals to VMS is clustering. One system goes down, another cluster member takes over. Minimal loss of time. Extremely important for things like stock exchanges, emergency backup in case of disasters, etc. If necessary, you can program in a naive GUI, but its not VMS's main area of application since it came up BEFORE the GUI era. GUI is largely relegated to a user's desktop or mobile system. OpenVMS usually hosts the core database ( Oracle and Rdb are popular). They were also among the first to come up with what we refer to as RAID and disk mirroring today.

[Richard Maher]

On 10/24/2015 6:29 AM, Mack Altman III wrote:
Very good questions!

Before anything VSI have to sure up the installed base and I think
they're doing a very good job. Without that beach-head we'd have nothing.

Commodity, standard hardware, virtualization, Apache, JAVA, Oracle 12G,
Oracle Rdb, DR clusters and all the 3GLs, is a formidable platform!

I've been harder on VSI than most about getting it done but first to
acknowledge that it's not as easy as it looks.

But those not at the coal face must be asking the same questions as Mack?

Who's going to build a search engine on VMS? Who is wanting to lease a
VMS VM in the Cloud? Who can backup the data, monitor the alarms,
patch-it, cater for the idiosyncrasies, and more importantly why would they?

Manufacturing? Real-Time? Wide-Area disaster recovery? I'm sure Windows
and Linux have "offerings" in this space.

What will differentiate VMS in the market for green-field customers.

I've previously suggested in getting onboard with the push-messaging
RFCs and being able to offer something the others (Yes, Google, Yahoo,
Microsoft, Amazon) do not.

[JF Mezei]

On 2015-10-23 19:45, David Froble wrote:

> And I respectfully disagree. You don't need to have the sources of an OS in
> order to use it. Even if something you need is missing, you can add it yourself.

And without respect, I disagree :-)

PSION shut down. Out of the PDA business. Proprietary system shifted to
Symbian which never really did much with it. Gone.

APPLE pulled out of servers, leaving anyone with a Apple server used as
a real server stranded. Apple went out of its way to disable server
management clients to prevent management of its last true server from
more modern clients. (Server monitor which is the ILO gui etc).

Digital/Compaq/HP pulled out of VMS, and HP killed ot outright (no
surprise).

The story of my life: choosing dead end OS.

When you rely on a proprietary system, you are at the mercy of the
owners who can pull the rug from under you with little to no warning.
PSION exited the PDA business when Microsoft announced it would produce
Windows-CE for PDAs. It took many years before the iPaq came out, the
only PDA from "windows" with some success, and it was killed by HP.

With Linux, or other "open source" software, it is much harder for any
one party to kill the software. And it is harder for anyone to remove
functionality (as Apple and Microsoft have done with OS-X a number of
times as it moves from one framework to the next and eventually
widthdraws the old one)


What Apple now calls a "server" is just an application. And it can
choose to remove stuff that used to come with the server such as MySQL,
or Squirrel Mail web interface, Dovecot IMAP/POP servers etc etc.

And when youa re proprietary, you are also often dependent on the
commercial provider to give your "their" version of open source software
such as Apache, Postfix etc, instead of allowing you to build from the
latest/patched open sourced versions.

[Arne Vajhøj]

On 10/23/2015 7:45 PM, David Froble wrote:

> Mack Altman III wrote:
>> So, why should one use VMS? Why not use Linux over VMS?
>
> Reverse the question. Why use Linux? Only decent reason I can think of
> is that it's free. Well, not if you want supported Linux.

What about availability of applications. In my book that is a very
decent reason.

Be sure that the platform will stick around.

be sure that you can easily get people that that know the platform.

>> I believe the answer in the past was to choose VMS because the on staff
>> programmer already had everything working for them and that was something
>> they were used to. However, it's easy to prove how quick you can
>> install a
>> third-party API or programming language on a cloud VPS and start
>> programming
>> an application within minutes thanks to companies, like DigitalOcean,
>> Linode,
>> and Vultr.
>
> I can start programming an application even quicker. Nor do I have to
> worry about the cloud going "poof". Yes, I have hardware. Nothing
> wrong with that. What do you use to get to "the cloud"?

Undoubtedly a computer.

But does that really have anything to do with the argument?

The argument is that you can scale up or scale down the number of
servers in minutes/hours and only pay for what you actually use.

Important for some businesses, but not for all. And cloud hosting
do come with its own issues as well.

>> For the VMS-lovers out there, we run InterSystems Cache on VMS every
>> day. And
>> I've been pondering for years why we haven't switched it to a Linux
>> system.
>> The only reason I've ever gotten is, well that's just always what we have
>> had. So, please don't think I'm bashing the OS at all. For those that
>> have
>> it, I've heard their passionate about it, but that's not an atmosphere
>> for
>> growth. Unless the client using it grows (which in my experience
>> people are
>> using Linux more often now), the revenue streams I would imagine are
>> stagnating.
>
> Perhaps one reason is in your words. "on VMS every day". Every day,
> huh? No problems, it just works, huh? If it ain't broke, then don't
> fix it. What would be the cost of at best a lateral move to Linux? If
> it's more than $0.00, then what is the business case for doing so?

You are assuming that the company:
- has no need for stuff not available on VMS
- do not see VMS long term viability (not just VMS itself but also third
party products support for VMS) as a risk
- are confident that they can continue finding people with VMS skills
- will not have economies of scale by reducing the number of platforms
to support
etc.

I am sure that that there are companies where that indeed is the case.

But there are also a lot where it is not the case. That is rather
obvious if you look at the number of companies that have moved
off VMS.

Arne

[Arne Vajhøj]

On 10/23/2015 8:54 PM, JF Mezei wrote:
> On 2015-10-23 19:45, David Froble wrote:
>> And I respectfully disagree. You don't need to have the sources of an OS in
>> order to use it. Even if something you need is missing, you can add it yourself.
>
> And without respect, I disagree :-)
>
> PSION shut down. Out of the PDA business. Proprietary system shifted to
> Symbian which never really did much with it. Gone.
>
> APPLE pulled out of servers, leaving anyone with a Apple server used as
> a real server stranded. Apple went out of its way to disable server
> management clients to prevent management of its last true server from
> more modern clients. (Server monitor which is the ILO gui etc).

> When you rely on a proprietary system, you are at the mercy of the
> owners who can pull the rug from under you with little to no warning.

And with open source you are at the mercy of those wanting
to work on that open source.

Sure Google and similar can take over the open source project themselves
if the need it, but the average small or mid size business does not
really have that option.

> And when youa re proprietary, you are also often dependent on the
> commercial provider to give your "their" version of open source software
> such as Apache, Postfix etc, instead of allowing you to build from the
> latest/patched open sourced versions.

No.

Open source software build just as well on proprietary OS.

Proprietary OS with not enough volunteers to maintain the
open source support for that OS have to rely on what the
vendor is willing to spend money on.

Different thing.

Arne

[Arne Vajhøj]

On 10/23/2015 8:47 PM, rcyoung wrote:
> One of the biggest appeals to VMS is clustering. One system goes
> down, another cluster member takes over. Minimal loss of time.
> Extremely important for things like stock exchanges, emergency backup
> in case of disasters, etc.

Cluster technology was rare 30 years ago.

Today almost everybody have it on some form.

> OpenVMS usually hosts the core database ( Oracle and
> Rdb are popular).

In the mid 90's Digital got the idea that VMS belonged in database tier
while client tier and server application should be Windows NT.

In my opinion one of the most disastrous decisions for VMS.

> They were also among the first to come up with what
> we refer to as RAID and disk mirroring today.

DEC was very much a technology leader.

In the 1980's.

Arne

[Arne Vajhøj]

On 10/23/2015 6:29 PM, Mack Altman III wrote:
> Each time I come to this group, I find myself less optimistic than I
> came here. Usually, it's to find how others have approached the same
> situations. However, I couldn't help but ask myself on this visit,
> "Will VMS be a part of the future?"

That is a good question.

:-)

> It seems VSI is holding onto keeping it a closed source, and I can't
> help but wonder why. I say this with the greatest respect, but I feel
> it's truthfully killing the OS.

It would not change anything.

There are not enough people to support even the most basic
open source applications in VMS.

It would be unrealistic to expect much support for open
sourced VMS itself.

> The only reason the WWW became what
> it is today is the open format.

Yes.

But standard interfaces and open source are two different things.

WWW became popular with proprietary browsers (NetScape and
Internet Explorer) .

The open standards HTML and HTTP made the difference.

> Millennials are rising and they all
> want to "do it themselves" because they feel they can "Google it" and
> know everything. What's working for Linux is the ability for all of
> them to find that information quickly, test it out, and tinker with
> it.

There are a few that like that.

But the vast majority just want something that works.

Try look at what Linux distros are most popular. It is the
ones that are geared towards people without too much IT
experience.

> So, why should one use VMS? Why not use Linux over VMS?
>
> I believe the answer in the past was to choose VMS because the on
> staff programmer already had everything working for them and that was
> something they were used to.

25 years ago the quality of VMS was much better than the
predecessors of todays popular OS.

> For the VMS-lovers out there, we run InterSystems Cache on VMS every
> day. And I've been pondering for years why we haven't switched it to
> a Linux system. The only reason I've ever gotten is, well that's just
> always what we have had.

It cost to switch platform.

Arne

[David Froble]

Arne Vajhøj wrote:
> On 10/23/2015 7:45 PM, David Froble wrote:
>> Mack Altman III wrote:
>>> So, why should one use VMS? Why not use Linux over VMS?
>>
>> Reverse the question. Why use Linux? Only decent reason I can think of
>> is that it's free. Well, not if you want supported Linux.
>
> What about availability of applications. In my book that is a very
> decent reason.

Well, in this case, it appears from what the OP wrote that they are running
their application(s) "every day". Guess they have what they need.

> Be sure that the platform will stick around.

One way to insure it won't is to not use it.

> be sure that you can easily get people that that know the platform.

A company trains people in what it needs them to do.

>>> I believe the answer in the past was to choose VMS because the on staff
>>> programmer already had everything working for them and that was
>>> something
>>> they were used to. However, it's easy to prove how quick you can
>>> install a
>>> third-party API or programming language on a cloud VPS and start
>>> programming
>>> an application within minutes thanks to companies, like DigitalOcean,
>>> Linode,
>>> and Vultr.
>>
>> I can start programming an application even quicker. Nor do I have to
>> worry about the cloud going "poof". Yes, I have hardware. Nothing
>> wrong with that. What do you use to get to "the cloud"?
>
> Undoubtedly a computer.

My point exactly, therefore, in either case some local HW is required. Now, if
you want to use outside resources, and way back in the day it was called "time
sharing" and there were entities offering such services, that is an option. Now
it's called "the cloud". So what?

> But does that really have anything to do with the argument?

There is no argument. Using outside resources is not limited to Linux, or
weendoze, or ????

> The argument is that you can scale up or scale down the number of
> servers in minutes/hours and only pay for what you actually use.
>
> Important for some businesses, but not for all. And cloud hosting
> do come with its own issues as well.

Oh, yes, definitely issues.

Well, in this case:
- it appears they have what they need, on VMS
- it's still running, long term
- you train what you need
- it appears their economies of scale are on VMS

> I am sure that that there are companies where that indeed is the case.
>
> But there are also a lot where it is not the case. That is rather
> obvious if you look at the number of companies that have moved
> off VMS.

Any time you stare a discussion with the "given" that "Z" is worthless, and "A"
is perfect, it is a flawed discussion. It becomes opinion and preferences.
Totally flawed.

Everything you say can be turned around. I can hold opinions totally opposite
yours.

[Phillip Helbig (undress to reply)]

In article <562acb27$0$19750$c3e8da3$3388...@news.astraweb.com>, JF

Mezei <jfmezei...@vaxination.ca> writes:

> > It seems VSI is holding onto keeping it a closed source, and I can't
> help but wonder why.
>
> Because HP has refused to hand VMS over to VSI, they only granted them a
> sort of "right to use". So VSI cannot decide to open source something
> which is still property of HP.

Even if VSI could decide this, why should they? It would cut off their
own revenue, as they could earn money only through maintenance. This
would make sense only if there were enough users willing to pay for
maintenance and also a large enough community able and willing to keep
VMS going, essentially without being paid for it. Neither is the case.

[David Froble]

Cut off revenue? I doubt it. People are not going to pay for an OS when they
can get one for free. If an OS vendor wants to stay in the game, them's the
current rules.

Don't see Red Hat not having revenue.

If the success of VSI is tied to one time license fees, then VMS is dead, and so
is VSI. A one time sale does not provide a steady revenue stream, and a steady
revenue stream is what a company needs to exist.

You could try the Microsoft way. Keep introducing new versions, hoping for new
sales. Got to ask, how many weendoze users have the latest version, what, 10 or
so? How many are still running XP? Microsoft has a rough road ahead of
themselves, at least for OS software.

[Scott Dorsey]

David Froble <da...@tsoft-inc.com> wrote:
>Phillip Helbig (undress to reply) wrote:
>> Even if VSI could decide this, why should they? It would cut off their
>> own revenue, as they could earn money only through maintenance. This
>> would make sense only if there were enough users willing to pay for
>> maintenance and also a large enough community able and willing to keep
>> VMS going, essentially without being paid for it. Neither is the case.
>>
>
>Cut off revenue? I doubt it. People are not going to pay for an OS when they
>can get one for free. If an OS vendor wants to stay in the game, them's the
>current rules.

Opening the system does two things: first of all it means that you lose
control over it almost completely, and secondly it means you lose the
revenue stream from sales.

Now, the honest truth is that the real money is in maintenance, and the
sales revenue is much smaller than the maintenance revenue for a commercial
system in general. However, as the product becomes more and more mature,
the need for maintenance is reduced.

There are good compromises; you can make the system available with source,
you can have free hobbyist licenses. You can in fact make the OS available
for free without making it open source. All of these help get more people
familiar with the system and build your installed base so that you can start
making more money on maintenance and support.

There are a lot of different alternatives, and I am sure VSI is considering
many of them, and I am sure that at some point once they have an OS that
they can sell which runs on modern hardware and which has few legal
encumberances that they will consider them still more seriously. But it
may take a few years for all of this to happen.

>Don't see Red Hat not having revenue.

It took Red Hat many, many years to have revenue. They were running in
the red for a long time until they could build their market up, but it
was okay because at the time people were pouring money into anything that
said 'linux' or 'internet' on the box. VSI does not have that option
open to them.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Jan-Erik Soderholm]

Den 2015-10-24 kl. 16:11, skrev David Froble:
> Phillip Helbig (undress to reply) wrote:
>> In article <562acb27$0$19750$c3e8da3$3388...@news.astraweb.com>, JF
>> Mezei <jfmezei...@vaxination.ca> writes:
>>>> It seems VSI is holding onto keeping it a closed source, and I can't
>>> help but wonder why.
>>> Because HP has refused to hand VMS over to VSI, they only granted them a
>>> sort of "right to use". So VSI cannot decide to open source something
>>> which is still property of HP.
>>
>> Even if VSI could decide this, why should they? It would cut off their
>> own revenue, as they could earn money only through maintenance. This
>> would make sense only if there were enough users willing to pay for
>> maintenance and also a large enough community able and willing to keep
>> VMS going, essentially without being paid for it. Neither is the case.
>>
>
> Cut off revenue? I doubt it. People are not going to pay for an OS when
> they can get one for free. If an OS vendor wants to stay in the game,
> them's the current rules.
>
> Don't see Red Hat not having revenue.
>
> If the success of VSI is tied to one time license fees, then VMS is dead,
> and so is VSI. A one time sale does not provide a steady revenue stream,
> and a steady revenue stream is what a company needs to exist.
>
> You could try the Microsoft way.

Like switching Office and the OS to an yearly fee modell?
Yes, we'll see how that will work.

> Keep introducing new versions, hoping for new sales.

According to MS themself, Windows 10 is the last new "version".
Everything after Win10 will be smaller or larger updates.

> Got to ask, how many weendoze users have the latest version,
> what, 10 or so?

According to the reports, it's over 100.000.000.
Most of these are free updates from earlier verisons,
but you asked how many they where.

Not that bad in an declaining PC market that has seen
over 10% drop in new PC sales over 2015.

> How many are still running XP?

https://www.netmarketshare.com/operating-system-market-share.aspx

[Neil Rieck]

IMHO, having a computer which does not play nice with the internet is a lot like having a telephone which does not play nice with the telephone network (in this case you move to a better phone). Okay so if you want run a web-server on OpenVMS then you can either 1) write one from from scratch 2) get some open-source software then compile-link-run 3) curate open-source software by hardening it for use on OpenVMS. If you go with items 2 or 3 the stuff will mostly be written in C whether you like it or not.

Okay so now consider other products like: a TCP/IP stack, OpenSSL, MySQL/MariaDB, gSOAP, to only name a few of many. The current version of "TCPIP Services" is so bad that VSI has announced ( http://vmssoftware.com/pdfs/VSI_Roadmap_20150924.pdf ) they would rather produce a replacement product. If I were a betting man I'd bet that they'd start by acquiring an opensource stack codebase then curate it for use on OpenVMS.

But all this is moot if the C and C++ compilers on OpenVMS are not up-to-date, and they are not. For example, just this week I needed to use these two C library functions (isblank and strnstr) but they were missing so I wrote my own which adds time to any project.

So I applaud the efforts of people at VSI who have taken up the crusade to modernize OpenVMS software development tools especially C and C++

Neil Rieck
Kitchener / Waterloo / Cambridge,
Ontario, Canada.
http://www3.sympatico.ca/n.rieck/links/openvms_demos.html

[Craig A. Berry]

On 10/24/15 10:43 AM, Neil Rieck wrote:

> The current version of
> "TCPIP Services" is so bad that VSI has announced (
> http://vmssoftware.com/pdfs/VSI_Roadmap_20150924.pdf ) they would rather
> produce a replacement product. If I were a betting man I'd bet that
> they'd start by acquiring an opensource stack codebase then curate it
> for use on OpenVMS.

I have no idea how they are doing it (as far as I can tell they haven't
said) but do note that traditional TCP/IP Services already is an open
source port. See the technical journal article that states, "The OpenVMS
TCP/IP kernel, the heart of the OpenVMS TCP/IP architecture, was ported
from BSD UNIX."[1]

Since they haven't said what they are doing, one can only speculate
whether VSI is replacing the "cradle," the "kernel" or both, and whether
this has any impact on user-facing APIs, utilities, and management and
configuration capabilities.


[1]
<http://h71000.www7.hp.com/openvms/journal/v4/tcp_ip_scalable_kernel.html>

[Phillip Helbig (undress to reply)]

In article <n0g3fe$fl3$1...@dont-email.me>, David Froble

<da...@tsoft-inc.com> writes:

> > Even if VSI could decide this, why should they? It would cut off their
> > own revenue, as they could earn money only through maintenance. This
> > would make sense only if there were enough users willing to pay for
> > maintenance and also a large enough community able and willing to keep
> > VMS going, essentially without being paid for it. Neither is the case.
>
> Cut off revenue? I doubt it. People are not going to pay for an OS
> when they can get one for free.

People certainly do pay for OS's. Maybe some are free, some are worth
paying for, like anything else.

> If an OS vendor wants to stay in the
> game, them's the current rules.

One size doesn't fit all.

> Don't see Red Hat not having revenue.

Do you honestly think that the Red-Hat revenue model would work for VMS?

> If the success of VSI is tied to one time license fees, then VMS is
> dead, and so is VSI. A one time sale does not provide a steady revenue
> stream, and a steady revenue stream is what a company needs to exist.

False dichotomy: it's not "only revenue is one-time license fee" or "it
must be open source".

> You could try the Microsoft way. Keep introducing new versions, hoping
> for new sales. Got to ask, how many weendoze users have the latest
> version, what, 10 or so? How many are still running XP? Microsoft has
> a rough road ahead of themselves, at least for OS software.

Again, false dichotomy: It is not either the Microsoft way or open
source.

[li...@openmailbox.org]

On Sat, 24 Oct 2015 17:19:28 +0000 (UTC)
Phillip Helbig undress to reply via Info-vax <info...@rbnsn.com> wrote:

> In article <n0g3fe$fl3$1...@dont-email.me>, David Froble
> <da...@tsoft-inc.com> writes:
>
> > > Even if VSI could decide this, why should they? It would cut off
> > > their own revenue, as they could earn money only through
> > > maintenance. This would make sense only if there were enough users
> > > willing to pay for maintenance and also a large enough community able
> > > and willing to keep VMS going, essentially without being paid for
> > > it. Neither is the case.
> >
> > Cut off revenue? I doubt it. People are not going to pay for an OS
> > when they can get one for free.

They're not going to pay for any OS on Intel. Nobody does that except
Windows victims and that appears to be coming to an end.

> > Don't see Red Hat not having revenue.
>
> Do you honestly think that the Red-Hat revenue model would work for VMS?

It might. Purchases and operational expenses are different types of budget
items. Sometimes it's harder to buy something than it is to spend the exact
same amount of money or more on something as a recurring expense. Given
people are now very used to paying for Linux that way and Linux is going to
be the main challenger given that it has eaten away and fouled the
landscape to the extent it has, and that it runs on Intel.

I think it's worth a try.

> > If the success of VSI is tied to one time license fees, then VMS is
> > dead, and so is VSI. A one time sale does not provide a steady revenue
> > stream, and a steady revenue stream is what a company needs to exist.

Agreed.

> False dichotomy: it's not "only revenue is one-time license fee" or "it
> must be open source".

Also agreed. But this doesn't appear to have anything to do with the
previous statement.

--
Please DO NOT COPY ME on mailing list replies. I read the mailing list.
RSA 4096 fingerprint 7940 3F02 16D3 AFEE F2F8 ACAA 557C 4B36 98E4 4D49

[JF Mezei]

On 2015-10-24 06:55, Phillip Helbig (undress to reply) wrote:

> Even if VSI could decide this, why should they? It would cut off their
> own revenue, as they could earn money only through maintenance.

If I were a large corporation, I would be a tad worried about uncertain
future for VMS/VSI. (which is better than having no future). I am not
sure how real/serious such fears are.

Open sourcing is a PR way to alleviate such fears. It doesn't really
matter whether VMS would continue to be developped by the community
after a theoretical sinkhole swallowed VSI office and all its engineers
into an abyss.

[JF Mezei]

On 2015-10-24 10:11, David Froble wrote:

> If the success of VSI is tied to one time license fees, then VMS is dead,

Perhaps VSI can do both. Sell perpetual VMS licence for a specific
version, or sell support which includes free upgrades.

Some customers may prefer one time capital expenditure and 0 cost after
that, other may prefer 0 captial exopenditure and opertational costs
after. Why not keep both options open ?

[John Reagan]

On Saturday, October 24, 2015 at 1:45:09 PM UTC-4, li...@openmailbox.org wrote:

>
> They're not going to pay for any OS on Intel. Nobody does that except
> Windows victims and that appears to be coming to an end.
>

Well, so far, the customers buying NonStop X are proving otherwise.

[John Reagan]

On Saturday, October 24, 2015 at 1:45:09 PM UTC-4, li...@openmailbox.org wrote:

>

> They're not going to pay for any OS on Intel. Nobody does that except
> Windows victims and that appears to be coming to an end.
>

[Bill Sorenson]

I think an apt comparison here is Solaris. Sun started open-sourcing Solaris in 2005, The rationale was that the "cost of software" is rapidly trending towards $0. People aren't paying for software anymore, they are paying for support. services, hardware, etc.

If nobody is going to be paying to use your software, you might as well get people using it so you can keep them in the eco-system. Notice Apple is giving away OS X now.

Solaris was, although slowly, on the rise until Oracle bought sun and squashed OpenSolaris. However, OpenSolaris lives on with Joyent, Delphix, OmniTI, to some extent iXSystems. The leftovers of OpenSolaris are probably used more now than Oracle Solaris.

I also wouldn't say you lose control of your open codebase. The official repo of record maintains under your control. You just lose the ability to prevent forks of the code. But with the Solaris example, Delphix, Joyent etc converged on Illumos. A single repo of record under their control. Yes there are forks but they are few and far between and with minimal changes.

If VSI could open VMS (which I know is an uphill battle) I think they'd be in a better position than they will with VMS staying proprietary. 100 companies paying for support is better than 0 companies paying for licensing.

I come from a mixed Solaris, BSD and VMS background. I'd love to see that combination displace Linux in the private cloud / enterprise space. But My opinion is that the future of software is open source. Proprietary software is in an ever shrinking position. VSI is currently chasing a shrinking market. There is growth potential in the near term but unless they start to pick up on the Illumos or BSD model in the future, I think the long run picture is a bit bleak.

[David Froble]

Bill Sorenson wrote:

> If VSI could open VMS (which I know is an uphill battle) I think they'd be in
> a better position than they will with VMS staying proprietary. 100 companies
> paying for support is better than 0 companies paying for licensing.

Here is where I got to ask "why?"

What does it really matter if the OS code is open source?

How many people have the expertise to work on the OS, and why would they do so?
Ok, if the vendor goes away, or something is missing, maybe, but if VMS is
developed to the extent neither of those are in play, what purpose would open
source be?

Now, as for licenses vs support, I've been saying that recurring revenue is the
way to go. If license fees are selected for revenue, either there won't be
enough and VSI will fail, or, they will raise prices to where nobody will buy,
and VSI will fail.

Short term, with the captive user base, license fees might work. But a captive
user base will always be a shrinking user base. Just the way it is.

[David Froble]

Phillip Helbig (undress to reply) wrote:

Where did I ever mention "open source". What I'd advocate is make the software
have no license fees, but if used for commercial purposes, then it must have
support.

[David Froble]

Those people would pay regardless of the method. I doubt many (if even one)
doesn't pay for support.

And has been mentioned again and again, support comes from a different budget,
that does not need upper management approval for a "purchase".

[JF Mezei]

On 2015-10-24 14:58, John Reagan wrote:

> Well, so far, the customers buying NonStop X are proving otherwise.

Something to consider:

Nonstop has extremely proprietary applications. For instance, in Canada,
the banks connect to the Interac inter-bank network using Tandem apps.
It would be very costly for 1 bank to commission a duplicate of that app
running on another platform and have it maintained/updated at exactly
the same speed as the "official" Tandem app.

Does VMS still have such apps ? It did when it was a popular platform
for SWIFT funds tranfer, but then your buddy Bob GQ Palmer told SWIFT to
get off VMS ASAP because VMS had no future.


Tandem was never as dead as VMS. In fact, HP went out of its way to
single out VMS and HP-UX when it promised they would never be ported
beyond IA64. (BTW, what is the current status of HP-UX ? have they
announced EOL for it yet ? or still pretending Itanium has a bright life
ahead of it ?)

[mcle...@gmail.com]

On Sunday, October 25, 2015 at 2:43:08 AM UTC+11, Neil Rieck wrote:
...

"Okay so if you want run a web-server on OpenVMS then you can either 1) write one from from scratch 2) get some open-source software then compile-link-run 3) curate open-source software by hardening it for use on OpenVMS. If you go with items 2 or 3 the stuff will mostly be written in C whether you like it or not."

WASD.

[Bill Sorenson]

To quote Bryan Cantrill, formerly of Sun:

"Ten years ago, I wrote that open source is "a loss leader -- minus the loss, of course"; after a decade of experience with open source business models, I would add that open source also serves as sales outreach without cold calls, as a channel without loss of margin, and as a marketing campaign without advertisements."

[David Froble]

Neil Rieck wrote:
> IMHO, having a computer which does not play nice with the internet is a lot
> like having a telephone which does not play nice with the telephone network
> (in this case you move to a better phone). Okay so if you want run a
> web-server on OpenVMS then you can either 1) write one from from scratch 2)
> get some open-source software then compile-link-run 3) curate open-source
> software by hardening it for use on OpenVMS. If you go with items 2 or 3 the
> stuff will mostly be written in C whether you like it or not.

Well, if I was to write a web server, it wouldn't be in C. Not that I have any
plans on writing a web server.

However, I do produce what I call "web services", even if my Swedish friend may
not allow them that title. I don't see any major differences in the concepts.
Basically, a listener that when it grants a connection, and the incoming data is
something it knows how to deal with, it does a job for the client. Basic, of
course.

:-)

[Jan-Erik Soderholm]

Den 2015-10-26 kl. 04:57, skrev David Froble:
> Neil Rieck wrote:
>> IMHO, having a computer which does not play nice with the internet is a lot
>> like having a telephone which does not play nice with the telephone network
>> (in this case you move to a better phone). Okay so if you want run a
>> web-server on OpenVMS then you can either 1) write one from from scratch 2)
>> get some open-source software then compile-link-run 3) curate open-source
>> software by hardening it for use on OpenVMS. If you go with items 2 or 3 the
>> stuff will mostly be written in C whether you like it or not.
>
> Well, if I was to write a web server, it wouldn't be in C. Not that I have
> any plans on writing a web server.
>
> However, I do produce what I call "web services", even if my Swedish friend

> may not allow them that title...

Childish. And I'm sure you now know why also. "Web Services" is a standard
name for a specific way of exposing interfaces to services. As you have
described your "services", it looks as just an open port to connect to.

> I don't see any major differences in the concepts.

There are. You do not have any WSDL files, do you? You do not have
any HTTP wrapper protocol, do you? That you do not understand
the difference is not my fault.

[Johnny Billquist]

On 2015-10-26 04:57, David Froble wrote:

David, you might be amused by my web-server, written in BASIC+2. Sources
at http://mim.update.uu.se/sources

Of course, this is served by that very same web-server... :-)

Johnny

--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: b...@softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol

[Bob Koehler]

In article <n0g3fe$fl3$1...@dont-email.me>, David Froble <da...@tsoft-inc.com> writes:
>

> Cut off revenue? I doubt it. People are not going to pay for an OS when they
> can get one for free. If an OS vendor wants to stay in the game, them's the
> current rules.

People will pay for a Lambourghini when they can get Fiat 500.

Folks will pay for an OS when the free OS doesn't meet their needs.

I have a few Windows-only apps that I must use at my current job.
What free OS is goind to meet my needs?

[David Froble]

Ah, yes. The guaranteed response ....

:-)

[David Froble]

Highly amused. Never thought of this one ...

"ON ERROR GOTO Hell"

LMAO

[David Froble]

Regardless of your nit picking, if you were the vendor, which would you rather
have, a one time license fee, or, recurring support fees, forever ....

I'm sure greed would want both, but there are people out there who won't go
along with that. If writing them off, is Ok, fine, but, customers are hard to
recruit.

[Bob Koehler]

In article <n0li71$k83$1...@dont-email.me>, David Froble <da...@tsoft-inc.com> writes:
>
> Regardless of your nit picking, if you were the vendor, which would you rather
> have, a one time license fee, or, recurring support fees, forever ....

That sounds like a different question.

Having dealt with both types of vendors, I'll go with one time
licensing for any one version. Updates by one time additional
license or by recurring maintenance fee are fine with me.

I find the recurring fees for one version are things customers often
try to cheat on.

[David Froble]

Being on both sides of the issue, things become more clear, at least to me.

All of our customers is what allows us to exist. So, revenue is very important.
The customers also recognize this, as we are constantly giving them new
things, and what would they do without us?

Similarly, we advise all of our customers to maintain a support contract for
VMS. Why? Not because any of them will ever use it. If there is an issue, it
goes through us, and normally stops right there. There is also the issue of new
versions. Ok, why? Because we depend on VMS being there. We recognize that
they need revenue to exist.

We do not sell our applications to our customers. They pay a monthly fee to use
them. Works out well for both. We get needed revenue. They get continuing
work on the applications. Staying up to date with new requirements is a big
part of the whole relationship.

I could see VMS for commercial use doing something similar. Actually, that is
what support fees do. It's just not addressed in the same language.

So, you want VMS to remain available, and with required new features? Pay for
it! Work out whatever structure VSI sets up for such a relationship.

[Keith Parris]

On 10/25/2015 12:23 AM, JF Mezei wrote:
> HP went out of its way to single out VMS and HP-UX when it promised
> they would never be ported beyond IA64.

Thanks to HP's deal with VSI, that temporary roadblock has been removed.

> (BTW, what is the current status of HP-UX ? have they
> announced EOL for it yet ? or still pretending Itanium has a bright life
> ahead of it ?)

HP's HP-UX Roadmap
( http://www8.hp.com/h20195/v2/GetPDF.aspx/4AA3-9071ENW.pdf ) shows
Standard Support (which includes new features and Engineering fixes)
for HP-UX 11i v3 through at least the end of 2025 (and could be
extended, of course).

Kittson is still on the official Roadmaps from Intel (page 12 at
http://www.intel.com/content/dam/www/public/us/en/documents/roadmaps/public-roadmap-article.pdf
)
and from HP
(
http://h17007.www1.hp.com/docs/enterprise/servers/integrity/project-odyssey/Project_Odyssey_Roadmap.pdf
).
Leo Demers' slides from Boot Camp indicated when HP expects
Kittson-based servers to become available for sale.

[JF Mezei]

On 2015-10-27 13:12, Keith Parris wrote:

> HP's HP-UX Roadmap
> ( http://www8.hp.com/h20195/v2/GetPDF.aspx/4AA3-9071ENW.pdf ) shows
> Standard Support (which includes new features and Engineering fixes)
> for HP-UX 11i v3 through at least the end of 2025 (and could be
> extended, of course).

So, no new version for HP-UX, just some fixes to support new disks.
Sounds a lot like what VMS was like just before HP formally announced
end of development.

[Keith Parris]

On 10/24/2015 12:04 PM, JF Mezei wrote:
> If I were a large corporation, I would be a tad worried about uncertain
> future for VMS/VSI. (which is better than having no future). I am not
> sure how real/serious such fears are.

VSI OpenVMS 8.4-1H1 is now selling, and VSI now has a revenue stream.
Some of that revenue is coming from HP for VSI products that HP is
reselling. Knowing this fact should help alleviate fears.

> Open sourcing is a PR way to alleviate such fears.

VSI has a license to the HP OpenVMS technology. The license doesn't
allow VSI to release OpenVMS code to open source, even if they
wanted to.

[Johnny Billquist]

On 2015-10-27 19:04, Keith Parris wrote:
> On 10/24/2015 12:04 PM, JF Mezei wrote:
>> If I were a large corporation, I would be a tad worried about uncertain
>> future for VMS/VSI. (which is better than having no future). I am not
>> sure how real/serious such fears are.
>
> VSI OpenVMS 8.4-1H1 is now selling, and VSI now has a revenue stream.
> Some of that revenue is coming from HP for VSI products that HP is
> reselling. Knowing this fact should help alleviate fears.

I wonder how much VSI gets, and how much HP keeps... I sure hope VSI can
make enough money to keep working at it. But I somehow don't trust HP...

>> Open sourcing is a PR way to alleviate such fears.
>
> VSI has a license to the HP OpenVMS technology. The license doesn't
> allow VSI to release OpenVMS code to open source, even if they
> wanted to.

Oh boy. Do this sound like Mentec all over again?

[Simon Clubley]

On 2015-10-28, Johnny Billquist <b...@softjar.se> wrote:
> On 2015-10-27 19:04, Keith Parris wrote:
>>
>> VSI has a license to the HP OpenVMS technology. The license doesn't
>> allow VSI to release OpenVMS code to open source, even if they
>> wanted to.
>
> Oh boy. Do this sound like Mentec all over again?
>

That's _exactly_ what I thought earlier on. Long term however, I really
hope VSI works out better than Mentec did.

While there are additional issues for VSI, such as porting to a new
architecture, which Mentec didn't have to worry about, there are also
positives in that the VMS ecosystem as it exists today is probably far
larger than the PDP-11 ecosystem was when Mentec took it over.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world

[Craig A. Berry]

On 10/30/15 2:23 PM, Simon Clubley wrote:
> On 2015-10-28, Johnny Billquist <b...@softjar.se> wrote:
>> On 2015-10-27 19:04, Keith Parris wrote:
>>>
>>> VSI has a license to the HP OpenVMS technology. The license doesn't
>>> allow VSI to release OpenVMS code to open source, even if they
>>> wanted to.
>>
>> Oh boy. Do this sound like Mentec all over again?
>>
>
> That's _exactly_ what I thought earlier on. Long term however, I really
> hope VSI works out better than Mentec did.

Did Mentec have a "grant of rights in perpetuity" as VSI has? Were there
other interested buyers when Mentec got access the PDP-11 code as there
were when VSI got access to VMS? Did Mentec have the right to port to
new hardware as VSI has?

[Simon Clubley]

On 2015-10-30, Craig A. Berry <craig...@nospam.mac.com> wrote:
> On 10/30/15 2:23 PM, Simon Clubley wrote:
>> On 2015-10-28, Johnny Billquist <b...@softjar.se> wrote:
>>> On 2015-10-27 19:04, Keith Parris wrote:
>>>>
>>>> VSI has a license to the HP OpenVMS technology. The license doesn't
>>>> allow VSI to release OpenVMS code to open source, even if they
>>>> wanted to.
>>>
>>> Oh boy. Do this sound like Mentec all over again?
>>>
>>
>> That's _exactly_ what I thought earlier on. Long term however, I really
>> hope VSI works out better than Mentec did.
>

Interesting questions below. Here's my attempt at answering them
but I'm sure Johnny and co can do better.

> Did Mentec have a "grant of rights in perpetuity" as VSI has?

Someone, (I can't remember who) apparently still has the rights to the
PDP-11 IP (although they are not doing anything with it) so my answer
is probably yes.

> Were there
> other interested buyers when Mentec got access the PDP-11 code as there
> were when VSI got access to VMS?

I don't know for sure.

> Did Mentec have the right to port to
> new hardware as VSI has?
>

They didn't actually port to new architectures (that probably didn't
make sense for the PDP-11 OS options) but they did produce a range of
new hardware options to run those operating systems.

[IanD]

+100

I'm going to give WASD a real plug here...

Pros
- Does websockets NOW
- Awesome support
- Runs on Vax, Alpha and Itanium
- Is cluster controllable, configurable
- Outperforms Apache on VMS
- Has bloody awesome support
- Proven itself on large sites, such as University of Malaga (The 4.1TB network traffic were responses from some 72 million network connections and 300 million requests), http://wasd.vsm.com.au/ht_root/other/uma_4tb.html (these are stats from 2009)
- Open source (C)

Cons
- One person looks after it
(however, others have contributed to the source over time for specific problems)

Has some great inbuilt stuff out of the box that enables one to search VMS systems / help. Makes for a much better interface into VMS help over terminal based help

I'm sure if VSI approached Mark Daniel (custodian of WASD) he would be more than happy to have WASD released with VMS as at least a help interface and per-installed from which a customer could easily build upon (obviously Mark would have to speak for himself on this front however)

Here's a taste of how easy access to VMS help stuff is

http://wasd.vsm.com.au/wasd_root/doc/misc/vmsscripts.html

or more specifically, here's the VMS help library exposed

http://wasd.vsm.com.au/Help

Can have tools like hyperspi active, which allows for easy graphic presentation of some monitor related functions. http://wasd.vsm.com.au/wasd_root/src/hyperspi/hyperspi_example.png

Or alamode system monitor, http://wasd.vsm.com.au/wasd_root/src/alamode/readmore.html#png

There is even a VT terminal available called dclinabox via websockets :-)

And if you still use vaxmail, there's a web interface for that was well, called soymail

Development is active and Daniel actually listens to what people want to do with WASD and before you know it, after some tinkering, he typically announces he's created functionality to support it

The above links are hosted on systems provided by VSM.COM (Thanks)

Why do I push WASD? Because many years ago, the company I worked for looked at web servers on VMS. I put my hat into the Apache camp because it was simply Apache (not objective I know).
Daniel was invited to give a presentation and I sat there grilling him with questions and he let the webserver do the talking instead of trying to push a sale pitch. He left the room and people looked at me and asked me what I thought.
What could I say! I had not been objective, he had. I had a per-concieved idea that Apache was it, he let WASD do the talking.
I simply answered that on VMS this 'WASD thing' kicks arse and I see no reason not to go with it...

The best thing about WASD is that it is driven by the VMS users, not some committee trying to appease people

[IanD]

Dam it. My previous post missed the comment of which I was replying to

Also a few typos; meh...

I still maintain VMS clusters are nice but what advantage do they hold over other linux clusters (apart from more stability)

The other clusters are catching and in some instances gone past VMS

I still think cheap grid computing back-boned off a VMS cluster might win some future sales

Having a system that can spin off a sort/cluster for example means you use all the systems in your cluster

Having a cluster present it's as a single large system means you can use all of your cluster versus having around for redundancy

As I said in another post, just what market is VMS going to target? No vision means no goals, no goals means going nowhere

I work in a fairly large IT company that is headquartered in the part of the world that see's lots of IT outsourcing

Young grads will not touch VMS. Why? Because in their eye's, why should they invest their time in something that will not get them into the new job, that is why

Linux = cheap, do it yourself in the education market. If you want to get marks on a project at uni, you need to build something with a team and present. Where is VMS in this equation?
If students cannot get their hands on a VMS system for turning in a project and if they cannot easily get hold of the bit of software they need to turn that project into a successful one, then they will not bother with it and then you have those same grads exiting the university system no-the-wiser about VMS and not asking for it or recommending it in their workplaces

Just how are we going to attract students to VMS? What incentive are we going to give them to be attracted to the platform?

x86 is a start but it's 2+ years away. What is being done at the grass roots level to communicate that VMS is available, it is being modernized and here is why you should be interested in it again. The education sector is important IMO, too important to ignore. Yes, it's a long term investment and not perhaps the immediate focus but the message needs to be getting out there

Anyone want to do a edX course on platform modernisation as a study around VMS? Might win some VMS future talent this way but you'd have to work with a university to do this and if they don't see a real reworking of VMS from the ground up as a new platform, they may not want to back the course

I'm beginning to think that VSI are interested in VMS for the current revenue stream and if it takes off again, then so be it but as long as it covers costs then they will be happy. I'm going by comments posted in another thread about an interview with the backer for VSI. I didn't really see anything in that interview that made me think they are backing this project on a do or die paradigm, like their life depended on it. Maybe I'm just skeptical. Just boxing along isn't good enough for VMS, it's going to take people willing to talk it up at every opportunity because the likes of windooze and linux certainly are, even to the point of BS with a mix of bravado. Get the gloves off VSI :-)

[David Froble]

IanD wrote:
> Dam it. My previous post missed the comment of which I was replying to
>
> Also a few typos; meh...
>
> I still maintain VMS clusters are nice but what advantage do they hold over
> other linux clusters (apart from more stability)

Could not answer this, as I have no knowledge of any clustering capabilities on
anything other than VMS. I've read that IBM has something similar, Sysplex, or
something like that.

However, Stephen Hoffman recently mentioned something about someone using a file
for locking a resource. Now, that's totally unacceptable, as far as I'm
concerned, but I'm sure there are those who will accept the practice.

> The other clusters are catching and in some instances gone past VMS

Don't know. As for shared everything, I've read that others don't agree with
this concept, so how could they "go past" VMS?

> I still think cheap grid computing back-boned off a VMS cluster might win
> some future sales

Not sure what this might be, but, not all holes are round, and not all pegs are
square. Not every application needs what VMS clustering provides.

> Having a system that can spin off a sort/cluster for example means you use
> all the systems in your cluster
>
> Having a cluster present it's as a single large system means you can use all
> of your cluster versus having around for redundancy

I do believe that in a VMS cluster, all systems can be used, none are waiting
around for others to fail, so no idle systems for redundancy. Not saying VMS
clusters don't provide redundancy, just saying they don't have systems idle and
waiting to be needed.

> As I said in another post, just what market is VMS going to target? No vision
> means no goals, no goals means going nowhere

This is a concept that can be rather hard. When you provide a general purpose
widget, it's not your job to "target" markets. It's the users who decide your
widget can help them. Yes, if you are aware of potential uses, you can market
to that usage, but, targeting specific markets sounds too much like limiting
yourself to just those markets. Goals could be to produce a better widget.

> I work in a fairly large IT company that is headquartered in the part of the
> world that see's lots of IT outsourcing

Maybe it's just me, but my opinion is that sometimes that doesn't work too well.

> Young grads will not touch VMS. Why? Because in their eye's, why should they
> invest their time in something that will not get them into the new job, that
> is why

The only thing that will allow them to get and keep a job is their ability.
Yes, some people hiring may think they're getting something special. Not too
many student projects are much good in the real world.

> Linux = cheap, do it yourself in the education market. If you want to get
> marks on a project at uni, you need to build something with a team and
> present. Where is VMS in this equation?

It's not. Students work toward good grades. To do so, you better do as the
instructors tell you to do. It's the instructors who have no use for VMS.
Those are the ones who would need to be convinced there is reasons to use VMS.
As Bill Gunshannon once observed, these instructors work from lesson plans,
reusing the same plan year after year. They are not going to do anything they
don't have to do. Changing to VMS would cause them work, and time away from the
golf course, the co-eds, and other important things.

> If students cannot get their hands on
> a VMS system for turning in a project and if they cannot easily get hold of
> the bit of software they need to turn that project into a successful one,
> then they will not bother with it and then you have those same grads exiting
> the university system no-the-wiser about VMS and not asking for it or
> recommending it in their workplaces

What companys do you know about, that hire grads, and then ask them how to run
the company? Do you really think some grad can come in and get a company to
change what it does? Not too often. Most of the time, such won't keep the job
long.

As for getting their hands on a VMS system, for student work, SimH is more than
enough, and they can run it on their notebook, tablet, and maybe their phones.

> Just how are we going to attract students to VMS? What incentive are we going
> to give them to be attracted to the platform?

An instructor who will give them good grades for using VMS? Only thing I can
think of.

> x86 is a start but it's 2+ years away. What is being done at the grass roots
> level to communicate that VMS is available, it is being modernized and here
> is why you should be interested in it again. The education sector is
> important IMO, too important to ignore. Yes, it's a long term investment and
> not perhaps the immediate focus but the message needs to be getting out there

You are correct. Whether there is a good answer, well, maybe there isn't.

> Anyone want to do a edX course on platform modernisation as a study around
> VMS? Might win some VMS future talent this way but you'd have to work with a
> university to do this and if they don't see a real reworking of VMS from the
> ground up as a new platform, they may not want to back the course
>
> I'm beginning to think that VSI are interested in VMS for the current revenue
> stream and if it takes off again, then so be it but as long as it covers
> costs then they will be happy. I'm going by comments posted in another thread
> about an interview with the backer for VSI. I didn't really see anything in
> that interview that made me think they are backing this project on a do or
> die paradigm, like their life depended on it. Maybe I'm just skeptical. Just
> boxing along isn't good enough for VMS, it's going to take people willing to
> talk it up at every opportunity because the likes of windooze and linux
> certainly are, even to the point of BS with a mix of bravado. Get the gloves
> off VSI :-)

What's wrong with supporting the current VMS users. They are sort of captive,
since anybody that could easily get off VMS left long ago. Pretty reliable
customers, huh?

Now, as for new users, VMS isn't totally unknown. There may be some who would
need what VMS provides. Nor will you win them all. Doesn't happen.

VSI needs to start somewhere, and the current customer base is their best
choice. If they get things more up to date, then VMS will sink or swim on it's
own merits. If somebody really needs it, and is for example a Microsoft or
Linux bigot, and their application fails, then they get fired, and perhaps VMS
gets a chance. Or perhaps not. But, regardless, at this time, the x86 port
isn't everything, it's the ONLY thing. [1]


[1] - credits to Vince Lombardi, who told his team "winning isn't everything,
it's the only thing".

[Johnny Billquist]

On 2015-10-30 20:23, Simon Clubley wrote:
> On 2015-10-28, Johnny Billquist <b...@softjar.se> wrote:
>> On 2015-10-27 19:04, Keith Parris wrote:
>>>
>>> VSI has a license to the HP OpenVMS technology. The license doesn't
>>> allow VSI to release OpenVMS code to open source, even if they
>>> wanted to.
>>
>> Oh boy. Do this sound like Mentec all over again?
>>
>
> That's _exactly_ what I thought earlier on. Long term however, I really
> hope VSI works out better than Mentec did.
>
> While there are additional issues for VSI, such as porting to a new
> architecture, which Mentec didn't have to worry about, there are also
> positives in that the VMS ecosystem as it exists today is probably far
> larger than the PDP-11 ecosystem was when Mentec took it over.

True. I do think that VMS have a larger ecosystem, which is good. And
yeah, Mentec adopted RSX for one or two new cpu models, but they were
still PDP-11s, so the work was fairly small. Most work went into Y2K
issues, and a few enhancements, and not much else. I guess it was partly
an effect also of existing customers mostly not really needing much else.

But the sad downside is that the software is still locked away because
of HP having control over things, and yet you will have a hard time
finding anyone at HP today who would be willing to even try to figure
out if they could release it, or what the situation exactly looks like.

[Johnny Billquist]

On 2015-10-30 22:02, Simon Clubley wrote:
> On 2015-10-30, Craig A. Berry <craig...@nospam.mac.com> wrote:
>> On 10/30/15 2:23 PM, Simon Clubley wrote:
>>> On 2015-10-28, Johnny Billquist <b...@softjar.se> wrote:
>>>> On 2015-10-27 19:04, Keith Parris wrote:
>>>>>
>>>>> VSI has a license to the HP OpenVMS technology. The license doesn't
>>>>> allow VSI to release OpenVMS code to open source, even if they
>>>>> wanted to.
>>>>
>>>> Oh boy. Do this sound like Mentec all over again?
>>>>
>>>
>>> That's _exactly_ what I thought earlier on. Long term however, I really
>>> hope VSI works out better than Mentec did.
>>
>
> Interesting questions below. Here's my attempt at answering them
> but I'm sure Johnny and co can do better.

I can try and add a few things I know or can guess. Unfortunately some
information is just not known to me, and it would appear some details of
the agreement was kept rather secret between DEC and Mentec,

>> Did Mentec have a "grant of rights in perpetuity" as VSI has?
>
> Someone, (I can't remember who) apparently still has the rights to the
> PDP-11 IP (although they are not doing anything with it) so my answer
> is probably yes.

Yes. However, the tricky word here is IP. Just as with VSI, HP kept the
IP rights. Mentec were allowed to develop and sell new versions, but
they could not make the sources available. Also, as far as I understand,
HP have a right to a fee for each license sold, and I believe it's not
possible to get around by "selling" for free, as that fee to HP is not a
percentage thing.
And the right to develop and sell new versions now lie with XX2247 LLC,
which bought the rights from Mentec. And they inherited the agreement
between DEC and Mentec as a part of taking over the software.
(And to make it even sadder, the layered products for the PDP-11 seems
to have been lost somewhere at Mentec, so they are not around anymore.
Only the OSes and DECnet.)

>> Were there
>> other interested buyers when Mentec got access the PDP-11 code as there
>> were when VSI got access to VMS?
>
> I don't know for sure.

I don't know either.

>> Did Mentec have the right to port to
>> new hardware as VSI has?
>>
>
> They didn't actually port to new architectures (that probably didn't
> make sense for the PDP-11 OS options) but they did produce a range of
> new hardware options to run those operating systems.

Right. Mentec made some new PDP-11 models, and they added support for
those into the OSes, along with some other peripherals.

Which is why I'm saying the deal with VSI sounds so similar to the
Mentec deal.
But bear in mind what I've written here is what I understand is the
deal. I have not seen the actual agreement in writing.
I also have not seen the actual agreement between HP and VSI in writing,
so all this is hearsay on my part.

[Jan-Erik Soderholm]

Den 2015-10-31 kl. 02:20, skrev IanD:


>
> I'm beginning to think that VSI are interested in VMS for the current
> revenue stream and if it takes off again, then so be it but as long as
> it covers costs then they will be happy.

At the talk I linked to by Johan Gedda (the main invester in VSI) he
was quite clear tht the first priority was the high end users that
asks for the i4 support and so on, 21% of the customers. Then comes
the lower-end folks that might still be running an older IA64 or
Alpha system. 79% of the customers. Third comes what could be
new sales/customers.

This was also backed by the PPT slides.

> I'm going by comments posted in
> another thread about an interview with the backer for VSI. I didn't
> really see anything in that interview that made me think they are
> backing this project on a do or die paradigm, like their life depended
> on it.

Johan Gedda was quite clear on that point also. He is also the main
owner of Rocket Software with 1.400 emp, so as he said, he was not
depending on VSI to get food on the table, or something similar.

[Stephen Hoffman]

On 2015-10-31 04:08:06 +0000, David Froble said:

> IanD wrote:
>> Dam it. My previous post missed the comment of which I was replying to
>>
>> Also a few typos; meh...
>>
>> I still maintain VMS clusters are nice but what advantage do they hold
>> over other linux clusters (apart from more stability)

Some of these include (reliable) host-based Volume Shadowing (HBVS)
integrated software RAID-1, and — this is a plus and sometimes a minus
— file locking enabled by default, and online volume expansion.

Biggest liability to these discussions is that these aren't features
that end-users usually buy into. It's the ISVs that have to want and
to use these (and other) features.

> Could not answer this, as I have no knowledge of any clustering
> capabilities on anything other than VMS. I've read that IBM has
> something similar, Sysplex, or something like that.

If you're interested in one of the approaches for operating a large
number of systems, see the Apache tools; Hadoop, Mesos, Kafka,
Zookeeper, etc.

There are other approaches and other sets of tools.

> However, Stephen Hoffman recently mentioned something about someone
> using a file for locking a resource. Now, that's totally unacceptable,
> as far as I'm concerned, but I'm sure there are those who will accept
> the practice.

OpenVMS doesn't have a way to use DLM system-wide without privileges;
noth without something somewhere granting SYSLCK.

On some other systems, a developer can digitally sign — a security
mechanism that OpenVMS isn't even remotely good at — an application
bundle — a concept that OpenVMS completely lacks — and part of that
signing and provisioning process also provides coordinated access to
system-wide mechanisms — which is again an approach that OpenVMS lacks.

OpenVMS security was state-of-the-art back in the 1990s. In this era,
consider how many of us could have compromised zip and unzip images
installed in our systems, secondary to some security breach at one of
the main distribution sites? PCSI utterly insanely lacks a way to
audit and verify what's installed, for that matter. One add-on cert
blows away the whole PCSI signing process, too — y'all do know how to
verify you don't have a rogue cert, right?

Why can't an application or an application bundle have its own
system-wide coordination via DLM? Is the old model of isolating stuff
by UICs and groups model really that useful, for that matter? UICs
and groups really aren't all that flexible, even years ago — ACLs were
in response to these limitations.

>> The other clusters are catching and in some instances gone past VMS
>
> Don't know. As for shared everything, I've read that others don't
> agree with this concept, so how could they "go past" VMS?

Usually by having far larger configurations available and operational,
for starters.

>
>> I still think cheap grid computing back-boned off a VMS cluster might
>> win some future sales
>
> Not sure what this might be, but, not all holes are round, and not all
> pegs are square. Not every application needs what VMS clustering
> provides.

Quite true. Conversely, whether the OpenVMS applications and designs
using what OpenVMS clustering provides might also be deployed in other
clusters, too.

>> Having a system that can spin off a sort/cluster for example means you
>> use all the systems in your cluster
>>
>> Having a cluster present it's as a single large system means you can
>> use all of your cluster versus having around for redundancy
>
> I do believe that in a VMS cluster, all systems can be used, none are
> waiting around for others to fail, so no idle systems for redundancy.
> Not saying VMS clusters don't provide redundancy, just saying they
> don't have systems idle and waiting to be needed.

More than a few folks with clusters do use hosts as a sort of warm
standby. This because distributing the activity over the hosts
involved can spin up the DLM and that can end up leading to
host-to-host lock traffic. That can involve links with more latency
or less bandwidth than might be desired, too. There are also
applications which aren't themselves clustered, but are using HBVS or
other features.

>> As I said in another post, just what market is VMS going to target? No
>> vision means no goals, no goals means going nowhere
>
> This is a concept that can be rather hard. When you provide a general
> purpose widget, it's not your job to "target" markets. It's the users
> who decide your widget can help them. Yes, if you are aware of
> potential uses, you can market to that usage, but, targeting specific
> markets sounds too much like limiting yourself to just those markets.
> Goals could be to produce a better widget.

OpenVMS Engineering has long had market segments they have worked with
— more than a few presentations showed these — and where engineering
added features and mechanisms useful for those folks, and reduced
bottlenecks that targeted customers were encountering. This is both a
very great strategy, and an utterly awful strategy. It's great for
the sorts of things that customers have seen and know and tell you
about, and it's awful for developing the sorts of things that customers
don't yet know they want — once they see or use those new features,
then they do want them. This strategy gets you incremental and
evolutionary changes, and far fewer revolutionary changes. The latter
are the sorts of features that tend to attract new applications and new
vendors.

>> I work in a fairly large IT company that is headquartered in the part
>> of the world that see's lots of IT outsourcing
>
> Maybe it's just me, but my opinion is that sometimes that doesn't work
> too well.

There's no panacea. I've seen competent outsourcing work well, and
I've seen insourcing work well, and I've seen both approaches fail
miserably. It's all in how it's managed and funded and monitored.
Biggest issue with outsourcing: the vendors usually won't fight with
the customer, when the customer is doing something stupid. This is the
flip side of the "the customer is always right" mantra. Sometimes the
customer is being an utter idiot. If a manager or a leader have good
staffers, they'll tell you you're being an idiot. Leaders that
surround themselves with sycophants hopefully have their personal exit
strategies completed before the organization imploded. (Outsourcing
your sycophants is cheaper than paying for in-house ones, too.)

> ...

> What's wrong with supporting the current VMS users. They are sort of
> captive, since anybody that could easily get off VMS left long ago.
> Pretty reliable customers, huh?

It's an inherently-declining market.

> ...

> VSI needs to start somewhere, and the current customer base is their
> best choice. If they get things more up to date, then VMS will sink or
> swim on it's own merits. If somebody really needs it, and is for
> example a Microsoft or Linux bigot, and their application fails, then
> they get fired, and perhaps VMS gets a chance. Or perhaps not. But,
> regardless, at this time, the x86 port isn't everything, it's the ONLY
> thing. [1]

And it may well not be enough.


--
Pure Personal Opinion | HoffmanLabs LLC

[Richard Maher]

On 10/31/2015 7:31 AM, IanD wrote:
> +100
>
> I'm going to give WASD a real plug here...

> Cons - One person looks after it (however, others have contributed to
> the source over time for specific problems)
>
>

> I'm sure if VSI approached Mark Daniel (custodian of WASD) he would
> be more than happy to have WASD released with VMS as at least a help
> interface and per-installed from which a customer could easily build
> upon (obviously Mark would have to speak for himself on this front
> however)
>

>

> The best thing about WASD is that it is driven by the VMS users, not
> some committee trying to appease people
>

If it is "Just Mark" still doing most of the work then it is a herculean
effort! Does WASD support HTTP/2?

[Jan-Erik Soderholm]

No. Currently (since 2004) full HTTP/1.1 compliance (RFC2616).

For details:

wasd.vsm.com.au/wasd_root/DOC/features/WASD_FEATURES.PDF

Chapter 2, Package Overview has a summary lists of the current
features.

[Jan-Erik Soderholm]

Actualy, just found out that I myself asked if there had been any
discussions around HTTP/2 on the WASD mail-list this summer... :-)

Mark Daniel replied like below.


> No not yet. But it has been topical with a couple of subscribers
> in private communication.

And about HTTP/2 and WASD as such Mark wrote:

> Not without its critics, it does look like becoming the next
> instantiation of "the Web protocol" and so deserves attention.
>
> A WASD implementation actively is being pursued (though at a very
> early stage) and if runs to completion would constitute WASD v11.
> However the protocol and associated header compression RFC are
> non-trivial and the developmental timeline probably puts any
> release somewhere in late 2016. I looked at its progenitor
> development, SPDY, a couple of years ago but understandably was
> unprepared to put in the effort on an "experiment". Whether or
> not the average WASD site will benefit from HTTP/2 is moot as it
> doesn't fundamentally change the underlying semantics of HTTP/1.1
> (it could be thought of as "wrapper" of sorts) and is targeted
> at reducing perceived networking bottlenecks in the client-server
> relationship.
> Cheers, Mark.

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
> Stephen Hoffman
> Sent: 31-Oct-15 11:29 AM
> To: info...@info-vax.com
> Subject: Re: [New Info-vax] VMS - What's next?
>
> On 2015-10-31 04:08:06 +0000, David Froble said:
>
> > IanD wrote:
> >> Dam it. My previous post missed the comment of which I was replying
> to
> >>
> >> Also a few typos; meh...
> >>
> >> I still maintain VMS clusters are nice but what advantage do they hold
> >> over other linux clusters (apart from more stability)
>
> Some of these include (reliable) host-based Volume Shadowing (HBVS)
> integrated software RAID-1, and — this is a plus and sometimes a minus
> — file locking enabled by default, and online volume expansion.
>
> Biggest liability to these discussions is that these aren't features
> that end-users usually buy into. It's the ISVs that have to want and
> to use these (and other) features.
>
> > Could not answer this, as I have no knowledge of any clustering
> > capabilities on anything other than VMS. I've read that IBM has
> > something similar, Sysplex, or something like that.

True - the z/OS (previously called MVS) clustering uses the same
design concepts i.e. shared everything as OpenVMS. They also
support multi-site clustering albeit quite a bit less distances than
OpenVMS.

Used to be a statement that the 2 best OS clustering technologies
on the planet had the same 3 letters. (VMS and MVS)

:-)

Having stated this, in today's world, one needs to architect new solutions
from the ground up. Hence, depending on the target strategy and goals,
there are pro's and con's with both shared nothing and shared everything
designs.

>
> If you're interested in one of the approaches for operating a large
> number of systems, see the Apache tools; Hadoop, Mesos, Kafka,
> Zookeeper, etc.
>
> There are other approaches and other sets of tools.

While stateless and highly distributed shared nothing solutions has been
the norm for the past decade+, it's interesting that some developers
are calling for a change back to stateful shared everything solutions as a
means to address all of the issues associated with high distributed shared
nothing solutions.

Reference: (thanks to Steve H for this pointer)
October 12, 2015 - Making the Case for Building Scalable Stateful Services
in the Modern Era
http://tinyurl.com/pylf5mg

Here is Caitie’s talk (Tech Lead at Twitter and former HALO4 game
developer). Listen to the headaches developers go through to develop
and support scalable and highly distributed solutions. Then consider how
a stateful solution based on a common clustered file system with an
integrated DLM might help to address stateful solutions.

Reference:
https://www.youtube.com/watch?v=H0i_bXKwujQ

Imho, after reading all of the above, a stateful architecture based on a
common clustered file system that contains an integrated distributed lock
manager design means that many of the developers concerns raised in
the video about workload distribution, keeping track of cluster members,
Sticky Sessions, Data Shipping Paradigm, Function Shipping Paradigm,
Data Locality, CAP (consistency, availability, persistence), Gossip
Protocols, Consistent Hashing etc. are taken care of at the operating
system/cluster level.

Such an approach would allow developers to concentrate more on where
their real added value is i.e. their creativity in adding new and better
code and functions in the Apps themselves.

[yes, it's back to the future time again ..]

:-)

[snip ..]

Regards,

Kerry Main
Kerry dot main at starkgaming dot com

[Stephen Hoffman]

On 2015-11-01 15:09:55 +0000, Kerry Main said:

> October 12, 2015 - Making the Case for Building Scalable Stateful
> Services in the Modern Era

direct URL:
http://highscalability.com/blog/2015/10/12/making-the-case-for-building-scalable-stateful-services-in-t.html

> Here is Caitie’s talk (Tech Lead at Twitter and former HALO4 game
> developer). Listen to the headaches developers go through to develop
> and support scalable and highly distributed solutions. Then consider
> how a stateful solution based on a common clustered file system with
> an integrated DLM might help to address stateful solutions.
>
>
>
> Reference:
>
> https://www.youtube.com/watch?v=H0i_bXKwujQ
>
>
>
> Imho, after reading all of the above, a stateful architecture based on
> a common clustered file system that contains an integrated distributed
> lock manager design means that many of the developers concerns raised
> in the video about workload distribution, keeping track of cluster
> members, Sticky Sessions, Data Shipping Paradigm, Function Shipping
> Paradigm, Data Locality, CAP (consistency, availability, persistence),
> Gossip Protocols, Consistent Hashing etc. are taken care of at the
> operating system/cluster level.

The disadvantage being the limitations on available scaling that the
coordination entails — same general morass as scaling cache coherency
hardware and protocols in larger shared-memory SMP or NUMA systems.
The more sockets you add, the more expensive and the slower your gear
inherently gets. There's almost always a sweet spot in scaling and
performance and price — for many applications in the Alpha era, that
was the AlphaServer ES4x boxes, for instance.

The biggest OpenVMS clusters are tiny little things in comparison to
the scale of current-generation application clusters too, which puts
you right back into the mix of approaches. OpenVMS clusters just can't
get that big, due to the coordination overhead.

Some application designs then head toward using HPE Superdome X boxes
or the yet-bigger SGI UV boxes, some toward grids, some toward
OpenVMS-style clusters, some toward lots and lots of small boxes.

As has long been stated... One size does not fit all.

More often than not, the biggest designs involve the application
developers making decisions that allow for that scaling, too — the OS
just doesn't have any clue what the application requirements really
are, and the OS — with "generic" tasks like memory paging or the
management of data caches or process scheduling — can make some Really
Bad Decisions for an application. For various designs, scaling might
also or will also involve a mix of Linux or BSD servers for some
functions, and OpenVMS servers for other tasks. Then there's that the
ratio of local to remote access to a DLM-protected resource is around
7:1, for instance. This relative unfairness of access then leads folks
to migrate applications, and to segment, and to shard. But I digress.

There are common issues for many, though — means of rapid provisioning
and deployment, monitoring and related tasks are all necessities.
These and other capabilities are often included in the much-maligned
"cloud" moniker. These become particularly necessary for folks dealing
with rapid changes in loads or in application mixes, or dealing with
patches and basic maintenance across an ever-increasing number of
servers for the available staff, too.


But then VSI also needs to make a compelling case for the inclusion of
OpenVMS in these sorts of designs, because OpenVMS is not a very
popular choice. This empirically, and for various good and pragmatic
reasons, too.

[David Froble]

What I take from that is that unless your bandwidth is maxed out, HTTP/2 doesn't
do much for you.

That sound about right?

[Stephen Hoffman]

On 2015-11-01 17:01:22 +0000, David Froble said:

> What I take from that is that unless your bandwidth is maxed out,
> HTTP/2 doesn't do much for you.
>
> That sound about right?

Web servers such as Apache can and often do zip the data with HTTP/1.1,
using a tool such as mod_deflate.

HTTP/2 gets you header compression, as well as integrated push support,
and some other additions.

Basically, the web applications now have a full duplex communications
path with the server, given current tools. A web browser or some other
arbitrary client can use HTTP as the backhaul link to the server, and
the web server can now initiate state changes as needed. The host
server can be running Apache 2.4.17 or later, some other web server, or
some other HTTP/2 application server.

I'd expect the VSI Apache port will likely pick up HTTP/2 support,
though the version discussed for the port was Apache 2.4.12 and not
2.4.17. (2.4.17 is the first version with fully integrated the
necessary HTTP/2 support into Apache. Prior to that, you needed some
patches.) Hopefully, VSI manages the source code in a way that's
reasonably maintainable, too; some VSI github project, for instance.

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
> Stephen Hoffman
> Sent: 01-Nov-15 10:49 AM
> To: info...@info-vax.com
> Subject: Re: [New Info-vax] VMS - What's next?
>

Agree that there is trade-off between SMP scaling, App scale-out, NUMA
and node in-memory caching considerations and the final pricing.

Part of the challenge is the current pricing model for OpenVMS is that
the costs for its biggest features are currently way to expensive. This
was true way back in the Digital days and has not improved with Compaq
or HP. It's also what has led to the downfall of other OS's like HP-UX,
Solaris, AIX and Non-Stop.

Very few Cust's went to legacy commodity systems because they were
technically better than what the Cust's were currently using. In most
(granted, not all) situations, it was because the costs were exponentially
Cheaper and the target commodity OS was viewed as "good enough".

With OpenVMS, clustering is one of its biggest features, but due to
the current pricing model, the only ones that were able to take advantage
of these features are higher end customers.

> The biggest OpenVMS clusters are tiny little things in comparison to
> the scale of current-generation application clusters too, which puts
> you right back into the mix of approaches. OpenVMS clusters just can't
> get that big, due to the coordination overhead.
>

That conclusion assumes a single cluster vs multiple clusters - each with
dedicated functions. As an example, one could use a dedicated Proxy / Web
services cluster (likely using WASD), another cluster for accounting, another
for APP1 services, another for App2 services etc.

A key advantage in each case is the common file system, DLM and other
cluster related features so that the clustering related issues do not have
to be designed into the App and the app coders can focus more on what
is their core responsibility.

In the WASD case, if the Proxy/Web service cluster started hitting the
75% mark, then one could simply drop a few more blades into the cluster.

In terms of scalability, I know of one ISP Cust that uses a small 8 node
OpenVMS completely homogeneous cluster to manage a customer base
of approx. 750,000 (likely quite a bit larger now) + many Help Desk + field
service reps all with heavily integrated workflows.

Yes, 750,000 cust base might be small, but as their loads increase, they
simply add another server or two and their load is automatically balanced.
Their app teams do not need to be involved in this type of activity.

> Some application designs then head toward using HPE Superdome X boxes
> or the yet-bigger SGI UV boxes, some toward grids, some toward
> OpenVMS-style clusters, some toward lots and lots of small boxes.
>
> As has long been stated... One size does not fit all.
>

Agree - really big nodes is one end of the spectrum and is good for OS's
like HP-UX because they do not scale out. Scale up is their only solution.

On the other end of the spectrum is the really distributed approach which
then becomes a mgmt and App support nightmare. In today's world, many
Custs are dealing with VM sprawl and that is a nightmare are on its own.

Imho, the better answer is to architect application solutions that use
dedicated clustered solutions with medium sized systems, perhaps with
larger memory for in memory caching, SSD's for system and hot files, and
a common file system so that things like maintaining state does have to be
left for silly and archaic solutions like cookies on client systems. One would
also focus on minimizing latency overhead by using things like Infiniband
based interconnects.

> More often than not, the biggest designs involve the application
> developers making decisions that allow for that scaling, too — the OS
> just doesn't have any clue what the application requirements really
> are, and the OS — with "generic" tasks like memory paging or the
> management of data caches or process scheduling — can make some
> Really
> Bad Decisions for an application. For various designs, scaling might
> also or will also involve a mix of Linux or BSD servers for some
> functions, and OpenVMS servers for other tasks. Then there's that the
> ratio of local to remote access to a DLM-protected resource is around
> 7:1, for instance. This relative unfairness of access then leads folks
> to migrate applications, and to segment, and to shard. But I digress.
>
> There are common issues for many, though — means of rapid provisioning
> and deployment, monitoring and related tasks are all necessities.
> These and other capabilities are often included in the much-maligned
> "cloud" moniker. These become particularly necessary for folks dealing
> with rapid changes in loads or in application mixes, or dealing with
> patches and basic maintenance across an ever-increasing number of
> servers for the available staff, too.
>

What you are talking about is capacity and infrastructure management.

"Public Cloud" solutions are simply outsourcing the capacity & infrastructure
management to an external vendor. Like any outsourcing solution, there
are good things and some very bad things associated with outsourcing.

The more you outsource, the less control you have over your solution. As
an example, how do you customize your "cloud" solution event's and alarms
to fit into your overall support model so that you can proactively take steps
to fix things before they impact your business?

"Private Clouds" are imho, the better approach because you control the
environment and only outsource those components which are tactical
and have low strategic value.

>
> But then VSI also needs to make a compelling case for the inclusion of
> OpenVMS in these sorts of designs, because OpenVMS is not a very
> popular choice. This empirically, and for various good and pragmatic
> reasons, too.
>

We all know the past issues associated with lack of marketing, lack of
vision, excessive pricing vs. competition and lack of support in HP, Compaq
and Digital.

I am sure the HP-UX and Non-Stop folks have similar complaints. For them,
they are under the same restrictions.

Yes, there is work to be done to catch up for the years of neglect, but the
good news is there is a new driver at the wheel, so OpenVMS is heading
in a new direction.

A diamond in the rough looking for some polish.

:-)

[Jan-Erik Soderholm]

That sounds about what I've understood. There are a number of features
in HTTP/2, one of them is that the web server identifies items that
it knows the browser will ask for, such as embeded icons or images
that the browser will fetch in separate HTTP GET transactions.

In HTTP/2, the server can identify these and pack them up at once
and send them over together with the base page in fewer transfers.

Then, of course, local browser cache will to a degree prevent
those additional fetches, if the user has "been there before".

See under "Goals" on https://en.wikipedia.org/wiki/HTTP/2.

[Mark Daniel]

Not what I intended to say David.

HTTP/2 is intended to improve browser page rendering by reducing a
number of latencies; those associated with maintaining multiple TCP
connections, the blocking request-response-request-response paradigm,
inability for the server to indicate relative resource importance,
amongst others.

It does these by multiplexing multiple request, client-server data over
a single TCP connection, using binary representations for HTTP/1.1
protocol metadata, compressing headers and using binary tokens to reduce
redundancy in metadata, allowing the server to "push" resources it
considers more important, amongst others.

My comment regarding whether or not the average WASD site will benefit
from HTTP/2, was more that (in my limited experience) VMS sites tend to
use Web services as glue in data processing (and so tend to be "simple"
request-responses), or the Web pages provided tend to be more prosaic,
fairly basic lumps of information rather than complex browser-based
applications where fine-grained, low-latency interactions dictate the
user experience (and the sort of Web environment the likes of Google
have particular interests in).

BTW. The quoted email suggests late 2016 as a possible WASD v11.0
release. The development code currently has all the basics of HTTP/2
functioning with browsing of a WASD site using HTTP/2 indistinguishable
from using HTTP/1.1. Not without some wrinkles to be ironed, it is
currently undergoing a refinement and corner-case testing phase, and may
well be available earlier than originally estimated - as a point-zero
release ;-)

[Richard Maher]

On 11/2/2015 5:29 AM, Jan-Erik Soderholm wrote:

> See under "Goals" on https://en.wikipedia.org/wiki/HTTP/2.
>
>
>

"Server push technologies"

[dunno]

Mack Altman III <mack.al...@gmail.com> wrote:
> Each time I come to this group, I find myself less optimistic than I came
> here. Usually, it's to find how others have approached the same
> situations. However, I couldn't help but ask myself on this visit, "Will
> VMS be a part of the future?"
>
> Personally, I started learning Linux around the age of 13 as that's when
> I started designing (not much to develop then) web sites. However, I
> didn't hear about VMS until 2008. There has only been one company that I
> have worked for that had VMS; and I never heard about VMS in college. I
> am aware that many of the staple industries have used the OS (e.g.
> finance, medical, transportation), but I am also aware that many have been converting.
>
> It seems VSI is holding onto keeping it a closed source, and I can't help
> but wonder why. I say this with the greatest respect, but I feel it's
> truthfully killing the OS.

I wonder how many Linux users nowadays actually need to read and/or alter
source code of OS. I did some Linux programming before, and I never had to
mess with OS source code. Windows have closed source, but I don't see
Windows programmers complaining about it.

> The only reason the WWW became what it is today is the open format.

There is a difference between open format and open source code. Many of
early web browsers had closed source, some are still closed.

> Millennials are rising and they all want to "do it themselves" because
> they feel they can "Google it" and know everything. What's working for
> Linux is the ability for all of them to find that information quickly,
> test it out, and tinker with it.

You probably thinking of something else. Most people don't need to read OS
source code to create application for that OS. Manuals and system calls
documentation work much better.

You probably thinking of everything that is shipped with OS, e.g. shell,
utilities, GUI, et cetera. However, those programs definitely can be under
different license than core of OS. Look at Mac OS X, almost everything
there is open source, including kernel, except Cocoa GUI and some other
smaller parts.

> So, why should one use VMS? Why not use Linux over VMS?
>
> I believe the answer in the past was to choose VMS because the on staff
> programmer already had everything working for them and that was something
> they were used to. However, it's easy to prove how quick you can install
> a third-party API or programming language on a cloud VPS and start
> programming an application within minutes thanks to companies, like
> DigitalOcean, Linode, and Vultr.
>
> Personally, I don't understand what benefit the VMS OS offers over Linux
> or Windows. I feel at least Windows provides a GUI; although, I'd enjoy
> Linux any day over Windows. And VMS is behind in almost every version of
> programming language unless you utilize a third-party installation (so
> again, why not Linux if there's no difference?).
>
> For the VMS-lovers out there, we run InterSystems Cache on VMS every day.
> And I've been pondering for years why we haven't switched it to a Linux
> system. The only reason I've ever gotten is, well that's just always what
> we have had. So, please don't think I'm bashing the OS at all. For those
> that have it, I've heard their passionate about it, but that's not an
> atmosphere for growth. Unless the client using it grows (which in my
> experience people are using Linux more often now), the revenue streams I
> would imagine are stagnating.
>



--
dunno

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
> dunno
> Sent: 04-Nov-15 2:01 PM
> To: info...@info-vax.com
> Subject: Re: [New Info-vax] VMS - What's next?
>

> Mack Altman III <mack.al...@gmail.com> wrote:
> > Each time I come to this group, I find myself less optimistic than I came
> > here. Usually, it's to find how others have approached the same
> > situations. However, I couldn't help but ask myself on this visit, "Will
> > VMS be a part of the future?"
> >
> > Personally, I started learning Linux around the age of 13 as that's when
> > I started designing (not much to develop then) web sites. However, I
> > didn't hear about VMS until 2008. There has only been one company that
> I
> > have worked for that had VMS; and I never heard about VMS in college. I
> > am aware that many of the staple industries have used the OS (e.g.
> > finance, medical, transportation), but I am also aware that many have
> been converting.
> >
> > It seems VSI is holding onto keeping it a closed source, and I can't help
> > but wonder why. I say this with the greatest respect, but I feel it's
> > truthfully killing the OS.
>
> I wonder how many Linux users nowadays actually need to read and/or
> alter
> source code of OS. I did some Linux programming before, and I never had
> to
> mess with OS source code. Windows have closed source, but I don't see
> Windows programmers complaining about it.
>

Agree 100%. There is a growing movement among many medium to large
companies to have their senior programmers focussed on adding strategic
value to the business by using their creativity to add functions and better
code to their applications to make their companies more competitive.

Spending a great deal of time in the weeds with OS patching and browsing
Web forums for support is not what most company IT execs would call
adding value to their business.

That's why even with RH, most companies buy support agreements from
RH. They would rather have a vendor do the low level support stuff &
focus their senior resources on making their company more competitive.

While open source will continue to have a place at the App/Utility layer,
my view is that it will continue to be (and likely become even less) of a
consideration at the OS layer.

> > The only reason the WWW became what it is today is the open format.
>
> There is a difference between open format and open source code. Many of
> early web browsers had closed source, some are still closed.
>
> > Millennials are rising and they all want to "do it themselves" because
> > they feel they can "Google it" and know everything. What's working for
> > Linux is the ability for all of them to find that information quickly,
> > test it out, and tinker with it.
>

What's driving Linux in medium to large companies is not their tinkering
capability but their very smart licensing model. They have very low CAPEX
(up front licensing free) which gets very high visibility with other platforms.
At the same time, they get loads of revenue from their monthly subscription
model since most med to large companies always buy support.

The support revenue is OPEX which, in most companies, is never seen by
anyone outside of the OPS manager level.

> You probably thinking of something else. Most people don't need to read
> OS
> source code to create application for that OS. Manuals and system calls
> documentation work much better.
>

Correct.

Remember the old expression - "When the only tool you know is a hammer,
every problem is a nail."

Many Customers are now running into VM sprawl with both Windows
and Linux. Their overall IT costs are rising as they need to manage, license
(agents, 3rd party SW), backup, manage users etc. for every OS - VM
or physical.

VM sprawl is a much, much tougher issue to resolve than server hardware
sprawl.

A big challenge will be the Windows/Linux culture which does not easily
adapt to running more than one business app on the same OS instance
(typically called App stacking).

App stacking is part of the OpenVMS culture (has always been done), so
it will be interesting how companies with commodity OS's address this
in the future. Docker is one technical way, but there is still the culture
issues to overcome i.e. "no way any other app is going to run on the
same OS as my App".

Another thought -

The statement "why would anyone run any OS other than Linux?" reminds
me of past statements like this:

"No product will ever replace WordPerfect or Lotus 123. They are used
everywhere and are much to engrained to ever be replaced."

"Why would anyone use Microsoft AD when they also have Novell NDS
In place already? NDS is years ahead of AD!"

[Stephen Hoffman]

On 2015-11-04 19:00:39 +0000, dunno said:

> Mack Altman III <mack.al...@gmail.com> wrote:
>> Each time I come to this group, I find myself less optimistic than I
>> came here. Usually, it's to find how others have approached the same
>> situations. However, I couldn't help but ask myself on this visit,
>> "Will VMS be a part of the future?"

For the vast majority of the IT market, no.

>> Personally, I started learning Linux around the age of 13 as that's
>> when I started designing (not much to develop then) web sites. However,
>> I didn't hear about VMS until 2008. There has only been one company
>> that I have worked for that had VMS; and I never heard about VMS in
>> college. I am aware that many of the staple industries have used the OS
>> (e.g. finance, medical, transportation), but I am also aware that many
>> have been converting.

Ports off of OpenVMS have been happening for years particularly
starting when VAX fell off the price and performance curve and as Unix
had more than a few users, and DEC itself starting recommended moving
chunks of applications off of OpenVMS back in ~1995, with their Windows
NT Affinity project, and more recently HP (now HPE) customer events
were increasingly highlighting migrations, in the years prior to the
the VSI announcement.

>> It seems VSI is holding onto keeping it a closed source, and I can't
>> help but wonder why. I say this with the greatest respect, but I feel
>> it's truthfully killing the OS.

Ignoring that VSI reports they don't have the rights to do release the
source code...

> I wonder how many Linux users nowadays actually need to read and/or
> alter source code of OS. I did some Linux programming before, and I
> never had to mess with OS source code. Windows have closed source, but
> I don't see Windows programmers complaining about it.

If your doesn't line up with evidence, maybe there's a different
explanation? Like maybe that there are different considerations.
Maybe the customers don't foresee Microsoft walking away from Windows
any time soon, but maybe it's that HPE has predicted that they will end
OpenVMS standard support for all but OpenVMS I64 V8.4 this year, end
prior version support for all but OpenVMS I64 V8.4 in 2018, and OpenVMS
I64 V8.4 exits standard in 2020 and all support in 2025. There's also
some understandable concern that VSI may well disappear.

>> The only reason the WWW became what it is today is the open format.

That TBL and CERN didn't decide to monetize the HTML work.

> There is a difference between open format and open source code. Many of
> early web browsers had closed source, some are still closed.

Some of the early components were, however, open-source.

>> Millennials are rising and they all want to "do it themselves" because
>> they feel they can "Google it" and know everything. What's working for
>> Linux is the ability for all of them to find that information quickly,
>> test it out, and tinker with it.

Folks accustomed to actually using the Internet being quite adept about
figuring out how to do things from web search engines and the Stack
sites and other available resources, after all.

Folks just don't have the time to read a whole manual when they're
trying to figure out specific tasks. This is related to the increase
in the availability of task-oriented documentation, and online help and
project examples in IDEs. It's quicker.

> You probably thinking of something else. Most people don't need to read
> OS source code to create application for that OS. Manuals and system
> calls documentation work much better.

The operating system itself usually contains tested examples of how to
do many of the operations, whether as part of the test suite or as part
of performing other functions. OpenVMS is filled with examples of
FABs and RABs and XABs and the rest, as well as zillions of system
calls.

Then there's that some of the available OpenVMS source code examples
can be hit or miss. More than a few that just don't compile now, and
that's without any discussions of the (lack of) error handling or
security or other issues within the code. We're not in the same
computing world that many of the OpenVMS examples were originally
written for, either.

There's that the major vendor-provided support resources (e.g. AskQ,
Ask James) went offline a decade or so ago, too.

Here's a list of what I've encountered
<http://labs.hoffmanlabs.com/node/203>, and — with the salient
exception of sites like 8³ — it's not readily indexed or searched, nor
is there a means to comment on or update what's posted.

> You probably thinking of everything that is shipped with OS, e.g.
> shell, utilities, GUI, et cetera. However, those programs definitely
> can be under different license than core of OS. Look at Mac OS X,
> almost everything there is open source, including kernel, except Cocoa
> GUI and some other smaller parts.

So... is this a suggestion that other hunks of OpenVMS be placed under
different licenses? Because otherwise, that OS X is a hybrid with
open-source and some have proprietary licenses doesn't seem to make
much sense in the context of learning and reading about OpenVMS.
Oddly, there are some examples of OpenVMS programming posted in the OS
X source pool, too — I've ended up there with some research. But
those examples are probably not what most folks are looking for when
they're trying to figure out the OpenVMS APIs.

[Stephen Hoffman]

On 2015-11-04 21:25:23 +0000, Kerry Main said:

>>
>> -----Original Message-----
>> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
>> dunno
>>

>> I wonder how many Linux users nowadays actually need to read and/or
>> alter source code of OS. I did some Linux programming before, and I
>> never had to mess with OS source code. Windows have closed source, but
>> I don't see Windows programmers complaining about it.

See my previous post.

> Agree 100%. There is a growing movement among many medium to

> largecompanies to have their senior programmers focussed on adding
> strategicvalue to the business by using their creativity to add
> functions and bettercode to their applications to make their companies
> more competitive.

Though some of those same folks are necessarily finding themselves also
managing risks; of what happens when a resource or a supplier fails, or
exits a market.

For its many problems, open source projects — those with a relatively
open license — can be picked up by another team.

> Spending a great deal of time in the weeds with OS patching and
> browsing Web forums for support is not what most company IT execs would
> call adding value to their business.

Ayup. How OpenVMS itself gets patched is positively baroque. That,
and some of the security patches are slow to arrive for OpenVMS —
Apache, OpenSSL, etc. I'm getting questions on how to change IP
addresses, too — hopefully the new VSI release fixes that.

> That's why even with RH, most companies buy support agreements fromRH.
> They would rather have a vendor do the low level support stuff &focus
> their senior resources on making their company more competitive.

It's sometimes called "buying somebody to blame", too. Whether this
approach works and is appropriate depends on the situation. There are
some vendors that are far more vertically-integrated
<https://en.wikipedia.org/wiki/Vertical_integration>, and either
reaping benefits from the associated control, or avoiding having some
other entity that can control your business.

> While open source will continue to have a place at the App/Utility
> layer, my view is that it will continue to be (and likely become even
> less) of aconsideration at the OS layer.

True. Most folks have settled on the Linux kernel, or on Windows.

But then the OpenVMS applications and libraries and frameworks are
going to be a subject for VSI work too, after they dig out from their
roadmap and from the port.

> ...

> What's driving Linux in medium to large companies is not their
> tinkering capability but their very smart licensing model. They have
> very low CAPEX (up front licensing free) which gets very high

> visibility with other platforms.At the same time, they get loads of
> revenue from their monthly subscriptionmodel since most med to large

> companies always buy support. The support revenue is OPEX which, in

> most companies, is never seen byanyone outside of the OPS manager level.

That, and Linux having tools and features and capabilities that can
solve many of the problems folks have, and in a relatively
cost-effective manner, and more than a few folks familiar with the
platform, and more than a few support resources available.

> Remember the old expression - "When the only tool you know is a
> hammer,every problem is a nail."

Spending time working on other platforms and using different approaches
and different tools is laudable. That can most definitely provide a
perspective into OpenVMS, too.

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of

> Stephen Hoffman
> Sent: 04-Nov-15 5:45 PM
> To: info...@info-vax.com
> Subject: Re: [New Info-vax] VMS - What's next?
>

[snip]

>
> > I wonder how many Linux users nowadays actually need to read and/or
> > alter source code of OS. I did some Linux programming before, and I
> > never had to mess with OS source code. Windows have closed source, but
> > I don't see Windows programmers complaining about it.
>
> If your doesn't line up with evidence, maybe there's a different
> explanation? Like maybe that there are different considerations.
> Maybe the customers don't foresee Microsoft walking away from Windows
> any time soon, but maybe it's that HPE has predicted that they will end
> OpenVMS standard support for all but OpenVMS I64 V8.4 this year, end
> prior version support for all but OpenVMS I64 V8.4 in 2018, and OpenVMS
> I64 V8.4 exits standard in 2020 and all support in 2025. There's also
> some understandable concern that VSI may well disappear.
>

Let's not forget that OpenVMS v8.4 is over 5 years old now (released in
June 2010). Similar to Microsoft, Apple, and other software companies,
VSI has been pretty clear that they do not plan to offer decades long
support contracts like has been the practice before with OpenVMS.

Re: VSI stability .. sure, any company might come to an end these days.

Just ask EMC, Compaq, Digital etc .. In the case of EMC, they had a close
relationship with Cisco as part of their VCE partnerships. Regardless of
official positioning, how long do you think Dell, who just bought EMC
will support new Cisco blade technologies as part of this VCE technology
agreement?

Heck, at one point even Apple almost went down the tube before they
woke up and brought Steve Jobs back to lead the company again.

Summary - There is no such thing these days as zero long term risk for
any company.

[snip...]

[Norm Raphael]

> On 11/04/15, Kerry Main<kerry...@backtothefutureit.com> wrote:
>
> ---8<snip>8---

>
> Summary - There is no such thing these days as zero long term risk for
> any company.
>

...or any bus.

> [snip...]
>
> Regards,
>
> Kerry Main
> Kerry dot main at starkgaming dot com

Norman F. Raphael
"Everything worthwhile eventually
degenerates into real work." -Murphy

[David Froble]

Kerry Main wrote:

> Summary - There is no such thing these days as zero long term risk for
> any company.

IBM ????????????????

[Bob Koehler]

A while back, IBM succeeded in making major changes that it needed to
survive. It could have failed.

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of Bob
> Koehler
> Sent: 06-Nov-15 9:22 AM
> To: info...@info-vax.com
> Subject: Re: [New Info-vax] VMS - What's next?
>

> In article <n1ftu0$rh2$1...@dont-email.me>, David Froble <davef@tsoft-

> inc.com> writes:
> > Kerry Main wrote:
> >
> >> Summary - There is no such thing these days as zero long term risk for
> >> any company.
> >
> > IBM ????????????????
>
> A while back, IBM succeeded in making major changes that it needed to
> survive. It could have failed.
>

Yep - while I have no doubts IBM will be relevant player for many years
to come, I do believe its sheer size is going to be a challenge for it to be
anywhere near as successful as in the past.

http://www.theregister.co.uk/2015/01/09/ibm_biggest_ever_reorg/
Article from Jan this year:
"IBM CEO Ginni Rometty is instigating the biggest global re-org in the
history of the corporation in a bid to carve out a clearer future in a cloudy
world."

"All the technology old guard are trying to ramp up their cloud business
to remain relevant in the 21st century, and not become the next Digital
Equipment Corporation"

[Stephen Hoffman]

On 2015-11-05 03:13:55 +0000, Kerry Main said:

>>
>> -----Original Message-----
>>
>> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
>>
>> Stephen Hoffman
>>
>> Sent: 04-Nov-15 5:45 PM
>>
>> To: info...@info-vax.com
>>
>> Subject: Re: [New Info-vax] VMS - What's next?
>>
>>
>>
>
>
> [snip]
>
>
>
>>
>>
>>> I wonder how many Linux users nowadays actually need to read and/or
>>> alter source code of OS. I did some Linux programming before, and I
>>> never had to mess with OS source code. Windows have closed source, but
>>> I don't see Windows programmers complaining about it.
>>>
>>
>>
>> If your doesn't line up with evidence, maybe there's a different
>> explanation? Like maybe that there are different considerations.
>> Maybe the customers don't foresee Microsoft walking away from Windows
>> any time soon, but maybe it's that HPE has predicted that they will end
>> OpenVMS standard support for all but OpenVMS I64 V8.4 this year, end
>> prior version support for all but OpenVMS I64 V8.4 in 2018, and OpenVMS
>> I64 V8.4 exits standard in 2020 and all support in 2025. There's also
>> some understandable concern that VSI may well disappear.
>>
>
>
> Let's not forget that OpenVMS v8.4 is over 5 years old now (released in
> June 2010).

OpenVMS V8.4 might be chronologically five or six years old, but — in
terms of its feature set, and associated tools and supported standards
— it's arguably older than that.

> Similar to Microsoft, Apple, and other software companies, VSI has been
> pretty clear that they do not plan to offer decades long support
> contracts like has been the practice before with OpenVMS.

Microsoft is still supporting a subset of certain Windows XP
configurations through ~2019. The extended support for Windows XP
went on for ~13 years.

While the Windows PC market isn't on a growth trend, the PC market is
massive and not going away for decades. Short of TEOTWAWKI.

As for OpenVMS, getting long-term support down to five years will be
interesting for everybody involved — for both positive and negative
reasons — and particularly if VSI starts to deprecate and remove
certain interfaces. But then call back in a decade, when VSI actually
has decade-old software, too.

> Re: VSI stability .. sure, any company might come to an end these days.
>
>
>
> Just ask EMC, Compaq, Digital etc .. In the case of EMC, they had a
> close relationship with Cisco as part of their VCE partnerships.

Which is part of why commodity gear can — while variously neither
distinguished or distinguishable — interesting. Which is why POWER,
SPARC and Itanium are a tougher sell, too.

> Regardless of official positioning, how long do you think Dell, who
> just bought EMC will support new Cisco blade technologies as part of
> this VCE technology agreement?

In three to five years, most commodity server gear installed will be
replaced by the then-current Dell, HPE, SuperMicro or other blade gear,
or by some generic OpenCompute blade or other such.

Getting tied to a proprietary management stack or tools can make that
migration more of a problem. Variations of discontinued tools and
products arise with even a stable vendor, too.

Or in five or seven years, maybe HPE The Machine will be available,
cheap and fast — and we'll all be porting our applications to photonics
and RRAM. Or we'll be working toward AArch64 servers. Mobile became
huge in the last five to seven years, after all.

> Heck, at one point even Apple almost went down the tube before they
> woke up and brought Steve Jobs back to lead the company again.

Having bought some of it, 1990s-vintage Apple hardware and software
gear was not competitive. Apple then started producing gear that
folks wanted, first with OS X, and then with reliable hardware that
could run Windows, BSD and Linux natively when (or if) that was
required. DEC in the 1990s... was acquired. For those that even
remember them, the BUNCH is long gone.

> Summary - There is no such thing these days as zero long term risk for
> any company.

Which is why open source, or source code escrow, or having multiple
vendors — something which is planned to be available for OpenVMS I64
for the next ~five years — can be requested and can be preferred by
various organizations. With OpenVMS starting in 2021 — which really
isn't all that long — we're all single-sourced again. With Linux or
BSD, the options are rather wider.

Beyond servicing and supporting the installed base, attracting new
applications and wholly new customers to VSI OpenVMS won't be easy.

[Kerry Main]

Those who buy Windows XP support pay exorbitant amounts of money
to Microsoft for custom support contracts with all sorts of qualifiers that
limit MS's exposure.

HP / Compaq did the same for Alpha HW. What they charged for support
of older Alpha HW would shock most people, but there were Cust's that
paid as they had few alternatives.

> While the Windows PC market isn't on a growth trend, the PC market is
> massive and not going away for decades. Short of TEOTWAWKI.
>
> As for OpenVMS, getting long-term support down to five years will be
> interesting for everybody involved — for both positive and negative
> reasons — and particularly if VSI starts to deprecate and remove
> certain interfaces. But then call back in a decade, when VSI actually
> has decade-old software, too.
>

Part of the solution to long support cycles is making it much easier to
upgrade with zero to minimal impact on Cust applications.

A good example is what VMware does with its ability to move OS's
dynamically from one server to another transparently to Apps. You
do this, then you can upgrade the old servers OS, FW, whatever.

One component with OpenVMS can be addressed with clusters and
good app design. Workloads can be migrated with zero App service
availability impact. This allows the old server to be upgraded
transparently.

Course, with cluster licensing being so expensive, many Cust's choose
to not implement clusters, so this further aggravates the long term
support challenge

> > Re: VSI stability .. sure, any company might come to an end these days.
> >
> > Just ask EMC, Compaq, Digital etc .. In the case of EMC, they had a
> > close relationship with Cisco as part of their VCE partnerships.
>
> Which is part of why commodity gear can — while variously neither
> distinguished or distinguishable — interesting. Which is why POWER,
> SPARC and Itanium are a tougher sell, too.
>

Well, in the case of Power, it has a pretty significant lead in single
process performance (some cores >4Ghz), so while it is not likely to have
the same volume as X86-64, when you get companies like Google
investigating the Power platform, that is a pretty good indication Power
is going to be around for some time.

> > Regardless of official positioning, how long do you think Dell, who
> > just bought EMC will support new Cisco blade technologies as part of
> > this VCE technology agreement?
>
> In three to five years, most commodity server gear installed will be
> replaced by the then-current Dell, HPE, SuperMicro or other blade gear,
> or by some generic OpenCompute blade or other such.
>
> Getting tied to a proprietary management stack or tools can make that
> migration more of a problem. Variations of discontinued tools and
> products arise with even a stable vendor, too.
>

There will always be a challenge between using std stacks or tools that
have limited customization to fit the Cust environment vs. a custom
stack or tool that is heavily integrated into the Cust workflows.

This is why Google is looking at using Power in its future environments
i.e. heavily custom solutions that have maximum added value for their
specific requirements.

> Or in five or seven years, maybe HPE The Machine will be available,
> cheap and fast — and we'll all be porting our applications to photonics
> and RRAM. Or we'll be working toward AArch64 servers. Mobile became
> huge in the last five to seven years, after all.
>

Applications, Marketing, ISV acceptance, Compatibility with existing
environments etc. will all have to be addressed with cool new stuff.

The Machine, Big Blue (IBM) are interesting ideas, but are solutions
looking for problems to solve and Apps that run on them.

> > Heck, at one point even Apple almost went down the tube before they
> > woke up and brought Steve Jobs back to lead the company again.
>
> Having bought some of it, 1990s-vintage Apple hardware and software
> gear was not competitive. Apple then started producing gear that
> folks wanted, first with OS X, and then with reliable hardware that
> could run Windows, BSD and Linux natively when (or if) that was
> required. DEC in the 1990s... was acquired. For those that even
> remember them, the BUNCH is long gone.
>
> > Summary - There is no such thing these days as zero long term risk for
> > any company.
>
> Which is why open source, or source code escrow, or having multiple
> vendors — something which is planned to be available for OpenVMS I64
> for the next ~five years — can be requested and can be preferred by
> various organizations. With OpenVMS starting in 2021 — which really
> isn't all that long — we're all single-sourced again. With Linux or
> BSD, the options are rather wider.
>

Let's not forget that a significant amount of companies PREFER proprietary
solutions. They would rather a vendor handle the weeds type OS issues
while their staff focus on added value applications & business integration.

Having senior support staff cruise web support forums looking for and/or
reviewing potential patches is, for many Cust's, not a good use of their staff
resources. It also means they do not have the single throat to choke when
there are issues.

Hence, as long as there is a formal support roadmap, AND the costs are
not through the roof, the single source vendor platform is not an issue
with most SMB companies.

> Beyond servicing and supporting the installed base, attracting new
> applications and wholly new customers to VSI OpenVMS won't be easy.
>

Agree, but being on the same X86-64 platform, having a more cost
competitive offering, and additional marketing will no doubt be part
of what VSI is planning for V9.X and V10.++ versions.

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
> Arne Vajhøj
> Sent: 23-Oct-15 9:45 PM
> To: info...@info-vax.com
> Subject: Re: [New Info-vax] VMS - What's next?
>

> On 10/23/2015 8:47 PM, rcyoung wrote:
> > One of the biggest appeals to VMS is clustering. One system goes
> > down, another cluster member takes over. Minimal loss of time.
> > Extremely important for things like stock exchanges, emergency
> backup
> > in case of disasters, etc.
>
> Cluster technology was rare 30 years ago.
>
> Today almost everybody have it on some form.
>
> > OpenVMS usually hosts the core database ( Oracle and
> > Rdb are popular).
>
> In the mid 90's Digital got the idea that VMS belonged in database tier
> while client tier and server application should be Windows NT.
>
> In my opinion one of the most disastrous decisions for VMS.
>
> > They were also among the first to come up with what
> > we refer to as RAID and disk mirroring today.
>
> DEC was very much a technology leader.
>
> In the 1980's.
>
> Arne
>

Well, since DEC is no longer around, this is a woulda, coulda,
shoulda type discussion, but having stated this, I have some internal
VHS recordings of DEC Engineering presentations from the 80's (yes,
using foils on overhead projector for those youngsters reading this).

First one was about "NAS" (Network Application Services). This was
all about integrating distributed vendor independent SW services on
the network.

When you see what they are talking about, todays strategists would
describe it as SOA - Service Oriented Architecture.

A second internal presentation was on "Software as a Strategy"

Unfortunately, senior DEC management and the boards long term vision
was measured in months. They gradually sold off critical assets to prop up
various quarter's results which further devalued the longer term value
of the company and then eventually sold the farm.

What DEC needed back then was a visionary like a Steve Jobs.

Ah well .. looking through the rear view mirror is no way to plan the
future, so onward and upward.

[abrsvc]

Kerry,

I still have some of those "foils" available. Even some that can run through a Xerox to get created. If you want to dazzle some of your future clients, let me know...

Dan

[JF Mezei]

On 2015-12-21 16:47, Kerry Main wrote:

> What DEC needed back then was a visionary like a Steve Jobs.

When they ousted Olsen and brought in Palmer, there was initially lost
of hope he would be able tourn the place around, market products, lowr
prices and push Alpha big time.

It's been too long to know whether this was wishful thinking or if there
were hints he would do that.

When he decided to make subtle colour change to the DIgital logo and
change the dot on the i, hope quickly faded and this was confirmed when
his only action was the regular quarterly reshuffling of execs in his
game of musical chairs.


Note that when Pfeiffer bought Digital, there was also hope he would
leverage stuff. Compaq even acknoledge us by having the gas pup ad where
it was written VMS. Imagine, not only was VMS mentioned in advertising,
but as "VMS" not that stuppid OpenVMS.

Pfeiffer also wanted to leverage Alpha. But he was ousted before they
could find a replacement and never found a replacement. The accountant's
job was to find a buyer for Compaq. (which he did when LaCarly needed
some diversion to delay her firing).

At the point VMS is today, (low point), it is much easier to look up and
ahead than to look behind because I don't think VMS can be harmed
anymore. HP put it on life support. Hopefully VSI wil have paddles set
to 15,000 volts and give VMS quite a jolt !

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of JF
> Mezei via Info-vax
> Sent: 21-Dec-15 8:23 PM
> To: info...@info-vax.com
> Cc: JF Mezei <jfmezei...@vaxination.ca>
> Subject: Re: [New Info-vax] VMS - What's next?
>

> On 2015-12-21 16:47, Kerry Main wrote:
>
> > What DEC needed back then was a visionary like a Steve Jobs.
>
> When they ousted Olsen and brought in Palmer, there was initially lost
> of hope he would be able tourn the place around, market products, lowr
> prices and push Alpha big time.
>
> It's been too long to know whether this was wishful thinking or if there
> were hints he would do that.
>

[snip...]

>
> At the point VMS is today, (low point), it is much easier to look up and
> ahead than to look behind because I don't think VMS can be harmed
> anymore. HP put it on life support. Hopefully VSI wil have paddles set
> to 15,000 volts and give VMS quite a jolt !
>

The "low" point was when HP announced VMS EOL with no buyer.

Today, with all that has transpired with VSI, VMS potential is higher than
any time in the last 10+ years (perhaps even longer).

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
> abrsvc via Info-vax
> Sent: 21-Dec-15 6:45 PM
> To: info...@info-vax.com
> Cc: abrsvc <dansabr...@yahoo.com>
> Subject: Re: [New Info-vax] VMS - What's next?
>

Dan - wow ... and I thought I was a pack rat from the good ole days.

:-)

I do have some of these VHS videos converted to .mpg files, so If these
are of interest, let me know.

[peter.l...@stromasys.com]

Hi,

I'm interested in those DEC engineering recordings.

Regards, Peter

[Simon Clubley]

One idea:

If Kerry puts them on Youtube (assuming there are no copyright issues)
then everyone can watch them (and others might stumble across them as
well).

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world

[IanD]

On Monday, December 28, 2015 at 10:14:41 PM UTC+11, Simon Clubley wrote:

> On 2015-12-28, peter.ljungberg@xxxxx <peter.ljungberg@xxxx> wrote:
> > On Tuesday, December 22, 2015 at 3:00:07 PM UTC+1, Kerry Main wrote:
> >>
> >> Dan - wow ... and I thought I was a pack rat from the good ole days.
> >>
> >> :-)
> >>
> >> I do have some of these VHS videos converted to .mpg files, so If these
> >> are of interest, let me know.
> >>
> >> Regards,
> >>
> >> Kerry Main
> >> Kerry dot main at starkgaming dot com
> >
> > Hi,
> >
> > I'm interested in those DEC engineering recordings.
> >
>
> One idea:
>
> If Kerry puts them on Youtube (assuming there are no copyright issues)
> then everyone can watch them (and others might stumble across them as
> well).
>
> Simon.
>
> --

> Simon Clubley, clubley@xxxxxx

> Microsoft: Bringing you 1980s technology to a 21st century world

+1 good idea

The only catch would be I think that YouTube limits an upload to 10 mins video duration unless you either have an old account that is pre this restriction or if you have a paid /business account

[Ken Robinson]

At 02:32 AM 1/5/2016, IanD via Info-vax wrote:

[Big Snip]

>+1 good idea
>
>The only catch would be I think that YouTube limits an upload to 10
>mins video duration unless you either have an old account that is
>pre this restriction or if you have a paid /business account

That restriction was lifted long ago. You now just have to verify
your account with them. I did that a few years ago and can upload
videos longer than 15 minutes. See
https://support.google.com/youtube/answer/71673?hl=en

Ken

[peter.l...@stromasys.com]

In that case you can either cut them up in smaller pieces (movie editing) or render the videos with a lower resolution.

^P

[Neil Rieck]

I noticed that HTTP2 is already supported in Apache-2.4 via plugin "mod_http2".

https://httpd.apache.org/docs/2.4/mod/mod_http2.html

However, the latest version of CSWS-2.2-1 from HPE is still based upon Apache-2.0

Perhaps HPE can back-port mod_http2 or perhaps they will move CSWS to Apache-2.4 or perhaps they will engage VSI where Apache-2.4 is rumored to already be working on OpenVMS-8.4-1H1

Neil Rieck
Kitchener / Waterloo / Cambridge,
Ontario, Canada.
http://www3.sympatico.ca/n.rieck/OpenVMS.html

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
> Simon Clubley via Info-vax
> Sent: 28-Dec-15 6:12 AM
> To: info...@info-vax.com
> Cc: Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP>
> Subject: Re: [New Info-vax] VMS - What's next?
>

> On 2015-12-28, peter.l...@stromasys.com
> <peter.l...@stromasys.com> wrote:
> > On Tuesday, December 22, 2015 at 3:00:07 PM UTC+1, Kerry Main
> wrote:
> >>
> >> Dan - wow ... and I thought I was a pack rat from the good ole days.
> >>
> >> :-)
> >>
> >> I do have some of these VHS videos converted to .mpg files, so If
> these
> >> are of interest, let me know.
> >>
> >> Regards,
> >>
> >> Kerry Main
> >> Kerry dot main at starkgaming dot com
> >
> > Hi,
> >
> > I'm interested in those DEC engineering recordings.
> >
>
> One idea:
>
> If Kerry puts them on Youtube (assuming there are no copyright issues)
> then everyone can watch them (and others might stumble across them
> as
> well).
>
> Simon.
>

Just in case not everyone saw the separate thread, the first DEC
video is now up on YouTube (20mins).

https://youtu.be/G2Rs-u55XUc
Digital Equipment - Future Directions in Computing (Mike Horner)

"Otherwise known as "DEC", there were numerous products and
strategies that DEC created which largely went unnoticed after
Compaq Computer bought the company back in 1998. This
Customer marketing VHS video was created my Mike Horner (DEC
Engineering Manager) and team back in Oct 1987. The video
content shows that some IT guiding principles and strategies are
timeless. As you listen to Mike's description of DEC's NAS (Network
Application Services), consider the similarity to what is now known
as SOA (service oriented architecture).
- Learning from the past to plan for the future ...Kerry Main"

Posting did not require me to have a business account. I plan
to post at least a couple more shortly.

[Richard Maher]

On 1/10/2016 1:49 AM, Neil Rieck wrote:
> Perhaps HPE can back-port mod_http2 or perhaps they will move CSWS to
> Apache-2.4 or perhaps they will engage VSI where Apache-2.4 is
> rumored to already be working on OpenVMS-8.4-1H1
>
> Neil Rieck Kitchener / Waterloo / Cambridge, Ontario, Canada.
> http://www3.sympatico.ca/n.rieck/OpenVMS.html
>
>
>

Good question.

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
> Richard Maher via Info-vax
> Sent: 09-Jan-16 8:05 PM
> To: info...@info-vax.com
> Cc: Richard Maher <maher_rj...@hotmail.com>
> Subject: Re: [New Info-vax] VMS - What's next?
>

>From the Dec 2015 roadmap published by VSI:
http://www.vmssoftware.com/pdfs/VSI_Roadmap_20151201.pdf

As part of V8.4-2 (March 2016)
- Apache 2.4.12

[Richard Maher]

On 1/10/2016 10:45 PM, Kerry Main wrote:
>
>
>>From the Dec 2015 roadmap published by VSI:
> http://www.vmssoftware.com/pdfs/VSI_Roadmap_20151201.pdf
>
> As part of V8.4-2 (March 2016)
> - Apache 2.4.12
>
> Regards,
>
> Kerry Main
> Kerry dot main at starkgaming dot com
>

Thanks Kerry. Missed it in that weighty tome.

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax...@info-vax.com] On Behalf Of
> Richard Maher via Info-vax
> Sent: 10-Jan-16 7:24 PM
> To: info...@info-vax.com
> Cc: Richard Maher <maher_rj...@hotmail.com>
> Subject: Re: [New Info-vax] VMS - What's next?
>

Perhaps they need a Coles Notes version?

:-)

[Jan-Erik Soderholm]

Den 2015-11-01 kl. 22:29, skrev Jan-Erik Soderholm:
> Den 2015-11-01 kl. 18:01, skrev David Froble:
>> Jan-Erik Soderholm wrote:
>>> Den 2015-11-01 kl. 13:22, skrev Jan-Erik Soderholm:
>>>> Den 2015-11-01 kl. 02:40, skrev Richard Maher:
>>>>> On 10/31/2015 7:31 AM, IanD wrote:
>>>>>> +100
>>>>>>
>>>>>> I'm going to give WASD a real plug here...
>>>>>
>>>>>> Cons - One person looks after it (however, others have contributed to
>>>>>> the source over time for specific problems)
>>>>>>
>>>>>>
>>>>>> I'm sure if VSI approached Mark Daniel (custodian of WASD) he would
>>>>>> be more than happy to have WASD released with VMS as at least a help
>>>>>> interface and per-installed from which a customer could easily build
>>>>>> upon (obviously Mark would have to speak for himself on this front
>>>>>> however)
>>>>>>
>>>>>
>>>>>>
>>>>>> The best thing about WASD is that it is driven by the VMS users, not
>>>>>> some committee trying to appease people
>>>>>>
>>>>>
>>>>>
>>>>> If it is "Just Mark" still doing most of the work then it is a herculean
>>>>> effort! Does WASD support HTTP/2?
>>>>>
>>>>
>>>> No. Currently (since 2004) full HTTP/1.1 compliance (RFC2616).
>>>>
>>>> For details:
>>>>
>>>> wasd.vsm.com.au/wasd_root/DOC/features/WASD_FEATURES.PDF
>>>>
>>>> Chapter 2, Package Overview has a summary lists of the current
>>>> features.
>>>>
>>>
>>> Actualy, just found out that I myself asked if there had been any
>>> discussions around HTTP/2 on the WASD mail-list this summer... :-)
>>>
>>> Mark Daniel replied like below.
>>>
>>>
>>> > No not yet. But it has been topical with a couple of subscribers
>>> > in private communication.
>>>
>>> And about HTTP/2 and WASD as such Mark wrote:
>>>
>>> > Not without its critics, it does look like becoming the next
>>> > instantiation of "the Web protocol" and so deserves attention.
>>> >
>>> > A WASD implementation actively is being pursued (though at a very
>>> > early stage) and if runs to completion would constitute WASD v11.
>>> > However the protocol and associated header compression RFC are
>>> > non-trivial and the developmental timeline probably puts any
>>> > release somewhere in late 2016. I looked at its progenitor
>>> > development, SPDY, a couple of years ago but understandably was
>>> > unprepared to put in the effort on an "experiment". Whether or
>>> > not the average WASD site will benefit from HTTP/2 is moot as it
>>> > doesn't fundamentally change the underlying semantics of HTTP/1.1
>>> > (it could be thought of as "wrapper" of sorts) and is targeted
>>> > at reducing perceived networking bottlenecks in the client-server
>>> > relationship.
>>> > Cheers, Mark.
>>>
>>>
>>
>> What I take from that is that unless your bandwidth is maxed out, HTTP/2
>> doesn't do much for you.
>>
>> That sound about right?
>
> That sounds about what I've understood. There are a number of features
> in HTTP/2, one of them is that the web server identifies items that
> it knows the browser will ask for, such as embeded icons or images
> that the browser will fetch in separate HTTP GET transactions.
>
> In HTTP/2, the server can identify these and pack them up at once
> and send them over together with the base page in fewer transfers.
>
> Then, of course, local browser cache will to a degree prevent
> those additional fetches, if the user has "been there before".
>
> See under "Goals" on https://en.wikipedia.org/wiki/HTTP/2.
>
>
>


Mark just posted the following on the WASD mail list

-------------------------------------------------------------------
WASD v11.0 is in the final stages of development before release.

The main new feature is HTTP/2 support. This has required a
significant refactoring of network I/O and the interface between
the HTTPs (now plural at the transport level) and server internals.
The rework was also an opportunity to rationalise other aspects of
server code and to excise *some* of the cruft accumulated over the
last twenty years.

To provide the WASD community with an opportunity to "touch and
feel" the operating server, as well as have access to the relevant
documentation prior to release, a demonstration site has been
arranged. This is hosted on an HP rx2660 (1.40GHz/6.0MB) with
2 CPUs and 8191MB running VMS V8.4 (cut-and-paste from the Server
Admin Statistics report :-) with Process Software MultiNet.

https://haven.vsm.com.au/

Poke around the server (pretty-much an out-of-the-box WASD) and
documentation. Please feed back any comment.

The HTTP/2 protocol is reasonably complex and there are undoubtedly
corner-cases that remain undiscovered at this stage. The final
release schedule will largely depend on what continuing field tests
reveal (and require remediation) but I'm hoping it will be in the
not-too-distant future.

As version 11.0 will be considered a classic point-zero release I
anticipate continuing to provide 10.4.n bugfix releases until v11.1.

Cheers, Mark.
-------------------------------------------------------------------

On the above page there is a test where the same fetches of small
icons can be compared between HTTP/1.n and HTTP/2. For 200 icons
I got the following times:

HTTP/1.n: approx 16.5 seconds
HTTP/2: approx 1.75 secs.


Jan-Erik.