Cloud Computing and SOA

[dustin....@gmail.com]

I'd like to here some ideas on the relationship of cloud computing and
SOA. To me, the relationship is quite interesting.

For one, I think the explosion of services hosted in the upper-most
layer of the cloud (SaaS) can be attributed partly to SOA standards
and principles. To be sure, there are a fair amount of web consumers
of these services (i.e. those using GMail), but there are also
enterprise consumers who utilize these services as part of larger
composite applications. SOA paves the way to enable service
composition acorss multiple domains, and this helps to increase use
and demand of services in the cloud.

Secondly, I believe SOA is to some degree an enabler of effective
virtualization. It's one thing to virtualize, but quite another to
offer intelligent, managed virtualization that delivers real gains.
SOA has helped to define how services are discovered, obtained, and
governed, and it seems that virtualization management techniques could
piggyback on these SOA protocols to effectively govern virtualized
services.

In the enabling layers of the cloud, application and physical
infrastructure, I think SOA is a pioneer into helping us understand
how such infrastructure can be consumed as a set of services. This may
lead to something like a Services Oriented Infrastructure, and I think
SOI could benefit quite a bit from the lessons already learned in SOA
efforts.

[Ross Mason]

Hi,

SOA really paved the way for Cloud. SOA taught us how to design in
terms of
services, develop components not monoliths, think through how others
will consume
the services we develop. This evolution is more important than the
shift to OO development.
We're still going through the evolution but we have reached a point
where we can start thinking
about different deployment models; the cloud is driven by this
thinking. I blogged about this recently:
http://rossmason.blogspot.com/2009/02/how-soa-paved-way-for-cloud.html

Cheers,

Ross
http://twitter.com/rossmason

[Sundeep Singatwaria]

 
 
Main value proposition of SOA is that it allows to build composite apps from different entities or to enable legacy systems to effectively talk to newer apps without wholesale rewrites. Cloud computing is certainly not that.
 
Clouds can certainly use the some of the standards of Web services but the two technologies have different value propositions for enterprises.
 
Composite SOA enabled apps would be better served with a cloud computing infrastructure so that the partners that talk to my SOA enabled apps can get a consistent response time from my apps because of elasticity of my compute infrastruture.
 
SOA brings savings on OPEX side whereas Clouds on CAPEX.
 
 
> Date: Thu, 26 Feb 2009 05:40:35 -0800
> Subject: [ Cloud Computing ] Cloud Computing and SOA
> From: dustin....@gmail.com
> To: cloud-c...@googlegroups.com

[dustin....@gmail.com]

Right now clouds are focused on CAPEX, but hopefully cloud management
tools evolve, we will hear about clouds reducing OPEX as well.

I wouldn't argue that individually SOA (not web services) and cloud
computing have different value propositions. I do think however that
the most interesting value propostions are those that exploit the
capabilities of SOA within cloud environments.

On Feb 26, 2:13 pm, Sundeep Singatwaria

<sundeep_singatwa...@hotmail.com> wrote:
> Main value proposition of SOA is that it allows to build composite apps from different entities or to enable legacy systems to effectively talk to newer apps without wholesale rewrites. Cloud computing is certainly not that.
>
> Clouds can certainly use the some of the standards of Web services but the two technologies have different value propositions for enterprises.
>
> Composite SOA enabled apps would be better served with a cloud computing infrastructure so that the partners that talk to my SOA enabled apps can get a consistent response time from my apps because of elasticity of my compute infrastruture.
>
> SOA brings savings on OPEX side whereas Clouds on CAPEX.
>
>
>
>
>
> > Date: Thu, 26 Feb 2009 05:40:35 -0800
> > Subject: [ Cloud Computing ] Cloud Computing and SOA

> > From: dustin.amrh...@gmail.com

> > To: cloud-c...@googlegroups.com
>
> > I'd like to here some ideas on the relationship of cloud computing and
> > SOA. To me, the relationship is quite interesting.
>
> > For one, I think the explosion of services hosted in the upper-most
> > layer of the cloud (SaaS) can be attributed partly to SOA standards
> > and principles. To be sure, there are a fair amount of web consumers
> > of these services (i.e. those using GMail), but there are also
> > enterprise consumers who utilize these services as part of larger
> > composite applications. SOA paves the way to enable service
> > composition acorss multiple domains, and this helps to increase use
> > and demand of services in the cloud.
>
> > Secondly, I believe SOA is to some degree an enabler of effective
> > virtualization. It's one thing to virtualize, but quite another to
> > offer intelligent, managed virtualization that delivers real gains.
> > SOA has helped to define how services are discovered, obtained, and
> > governed, and it seems that virtualization management techniques could
> > piggyback on these SOA protocols to effectively govern virtualized
> > services.
>
> > In the enabling layers of the cloud, application and physical
> > infrastructure, I think SOA is a pioneer into helping us understand
> > how such infrastructure can be consumed as a set of services. This may
> > lead to something like a Services Oriented Infrastructure, and I think
> > SOI could benefit quite a bit from the lessons already learned in SOA
> > efforts.
>

> _________________________________________________________________
> Windows Live™ Hotmail®…more than just e-mail.http://windowslive.com/howitworks?ocid=TXT_TAGLM_WL_t2_hm_justgotbett...- Hide quoted text -
>
> - Show quoted text -

[C Wegrzyn]

I can second that.In the cloud storage fabric I am building to sell I
focus on reduced capex and opex. That seems to get great traction in the
marketplace.

Chuck Wegrzyn

[Tim M. Crawford]

This is great news. But I think you're leaving quite a bit on the table by only focusing on CapEx/ OpEx. There is another aspect to consider: As more vendors start to develop their cloud offerings (SaaS, PaaS and IaaS) and create increased competition, it will drive the costs down in an area that is already headed toward a commodity business (ie: infrastructure). This means that you need to volume to cover the reduced margins that competition will bring.

We need to get beyond cloud solutions just being a cost reduction. Cost reductions are definitely a positive thing. They still need to be part of the evaluation. However, they should be a check-box rather the key reason for moving. The problem is that CFO's and CIO's will see through the cost reductions and consider the risks too. The balance between just risk and cost reduction is equal to negative without any other considerations. There are additional benefits that will come with the change. This is where I believe the real value is for IT organizations. Cloud providers that look beyond the better-faster-cheaper argument as their key driver will be the ones that ultimately prosper.

That being said, have you considered what other value-adds the bring moves?

Tim

---

From: C Wegrzyn <eprpa...@gmail.com>
To: cloud-c...@googlegroups.com
Sent: Thursday, February 26, 2009 4:32:28 PM
Subject: [ Cloud Computing ] Re: Cloud Computing and SOA

[C Wegrzyn]

The other "hot button" issue that has come up repeatedly, and it
surprised me, was no vendor lock in! When I heard it I thought I must
have misunderstood. But they want a solution which gives them access to
the data outside of the normal software delivery. Now why they want that
I am not sure but I imagine it has to do with not feeling indebted to a
hardware vendor.

C.

Tim M. Crawford wrote:
> This is great news. But I think you're leaving quite a bit on the table
> by only focusing on CapEx/ OpEx. There is another aspect to consider: As
> more vendors start to develop their cloud offerings (SaaS, PaaS and
> IaaS) and create increased competition, it will drive the costs down in
> an area that is already headed toward a commodity business (ie:
> infrastructure). This means that you need to volume to cover the reduced
> margins that competition will bring.
>
> We need to get beyond cloud solutions just being a cost reduction. Cost
> reductions are definitely a positive thing. They still need to be part
> of the evaluation. However, they should be a check-box rather the key
> reason for moving. The problem is that CFO's and CIO's will see through
> the cost reductions and consider the risks too. The balance between just
> risk and cost reduction is equal to negative without any other
> considerations. There are additional benefits that will come with the
> change. This is where I believe the real value is for IT organizations.
> Cloud providers that look beyond the better-faster-cheaper argument as
> their key driver will be the ones that ultimately prosper.
>
> That being said, have you considered what other value-adds the bring moves?
>
> Tim
>

> ------------------------------------------------------------------------
> *From:* C Wegrzyn <eprpa...@gmail.com>
> *To:* cloud-c...@googlegroups.com
> *Sent:* Thursday, February 26, 2009 4:32:28 PM
> *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA

>
>
> I can second that.In the cloud storage fabric I am building to sell I
> focus on reduced capex and opex. That seems to get great traction in the
> marketplace.
>
> Chuck Wegrzyn
>

> dustin....@gmail.com <mailto:dustin....@gmail.com> wrote:
>> Right now clouds are focused on CAPEX, but hopefully cloud management
>> tools evolve, we will hear about clouds reducing OPEX as well.
>>
>> I wouldn't argue that individually SOA (not web services) and cloud
>> computing have different value propositions. I do think however that
>> the most interesting value propostions are those that exploit the
>> capabilities of SOA within cloud environments.
>>
>> On Feb 26, 2:13 pm, Sundeep Singatwaria
>> <sundeep_singatwa...@hotmail.com

> <mailto:sundeep_singatwa...@hotmail.com>> wrote:
>>> Main value proposition of SOA is that it allows to build composite
> apps from different entities or to enable legacy systems to effectively
> talk to newer apps without wholesale rewrites. Cloud computing is
> certainly not that.
>>>
>>> Clouds can certainly use the some of the standards of Web services
> but the two technologies have different value propositions for enterprises.
>>>
>>> Composite SOA enabled apps would be better served with a cloud
> computing infrastructure so that the partners that talk to my SOA
> enabled apps can get a consistent response time from my apps because of
> elasticity of my compute infrastruture.
>>>
>>> SOA brings savings on OPEX side whereas Clouds on CAPEX.
>>>
>>>
>>>
>>>
>>>
>>>> Date: Thu, 26 Feb 2009 05:40:35 -0800
>>>> Subject: [ Cloud Computing ] Cloud Computing and SOA

>>>> From: dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com>
>>>> To: cloud-c...@googlegroups.com
> <mailto:cloud-c...@googlegroups.com>

[dustin....@gmail.com]

By move, I would assume we are considering both the case of the
private cloud and the public cloud? I think there are definitely
propositions beyond cost savings. The one that always comes to mind
for me is that companies can focus human resource toward delivering
technology that provides real business value, and less time can be
invested enabling technology that brings business value. What are some
other values of cloud computing?

On Feb 26, 6:49 pm, "Tim M. Crawford" <timmcrawf...@yahoo.com> wrote:
> This is great news. But I think you're leaving quite a bit on the table by only focusing on CapEx/ OpEx. There is another aspect to consider: As more vendors start to develop their cloud offerings (SaaS, PaaS and IaaS) and create increased competition, it will drive the costs down in an area that is already headed toward a commodity business (ie: infrastructure). This means that you need to volume to cover the reduced margins that competition will bring.
>
> We need to get beyond cloud solutions just being a cost reduction. Cost reductions are definitely a positive thing. They still need to be part of the evaluation. However, they should be a check-box rather the key reason for moving. The problem is that CFO's and CIO's will see through the cost reductions and consider the risks too. The balance between just risk and cost reduction is equal to negative without any other considerations. There are additional benefits that will come with the change. This is where I believe the real value is for IT organizations. Cloud providers that look beyond the better-faster-cheaper argument as their key driver will be the ones that ultimately prosper.
>
> That being said, have you considered what other value-adds the bring moves?
>
> Tim
>
> ________________________________

> From: C Wegrzyn <eprparad...@gmail.com>

> >> Windows Live™ Hotmail®…more than just e-mail.http://windowslive.com/howitworks?ocid=TXT_TAGLM_WL_t2_hm_justgotbett...Hide quoted text -
>
> >> - Show quoted text -- Hide quoted text -

[John R Savageau]

Tim

 

There is definitely the green argument.  Less physical data centers, better power and cooling efficiency, less cost in building out and maintaining data center infrastructure, potential for “green” credits (such as in Australia), and other benefits - not the least being public relations.

 

John

 

John Savageau

CTO, CRG West

One Wilshire Building

Los Angeles

[dustin....@gmail.com]

This gives them the freedom to move across vendors with their data and
applications as well. Look at the recent announcement of Coghead going
out of business, and you can see the tight spot some companies are in.

> > *From:* C Wegrzyn <eprparad...@gmail.com>

> > *To:* cloud-c...@googlegroups.com
> > *Sent:* Thursday, February 26, 2009 4:32:28 PM
> > *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA
>
> > I can second that.In the cloud storage fabric I am building to sell I
> > focus on reduced capex and opex. That seems to get great traction in the
> > marketplace.
>
> > Chuck Wegrzyn
>

> > Hide quoted text -
>
> >>> - Show quoted text -- Hide quoted text -

[Tim M. Crawford]

Thanks Dustin for clarifying. You're right that this is a critical issue for IT organizations. IT orgs need to have the flexibility to move services (and data) from one vendor to another. And it need to be relatively easy to do. Vendor Lock-In is a real issue that needs a solution.

One possibility is for a new market of vendors to develop ways to integrate disparet sub-components. This would be somewhat similar to what Rightscale has done for the infrastructure level. Another possibility is that interoperability becomes a real possibility between vendors. But the market will most likely need to mature a bit before that can happen.

Tim

---

From: "dustin....@gmail.com" <dustin....@gmail.com>
To: Cloud Computing <cloud-c...@googlegroups.com>
Sent: Thursday, February 26, 2009 7:18:09 PM
Subject: [ Cloud Computing ] Re: Cloud Computing and SOA

[Tim M. Crawford]

I agree. There are a number of metrics that support your point of re-focusing IT organization on higher-value functions that tie directly to the core business. We (as an industry and profession) need to look at way to increase our value to the business...not just look for ways to reduce costs.

Self Promotion: I'm presenting at the SaaScon and Cloud Computing Expo East conferences. The presentation will outline a framework for Cloud Computing (SaaS, PaaS, Iaas), opportunities for IT organizations and challenges to work through. It will cover game-changing strategies for organizations to leverage.

 

______________________________________
Tim M. Crawford
650.804.1300

timmcr...@yahoo.com
http://www.linkedin.com/in/timcrawford

---

From: "dustin....@gmail.com" <dustin....@gmail.com>
To: Cloud Computing <cloud-c...@googlegroups.com>
Sent: Thursday, February 26, 2009 6:51:26 PM

[Tim M. Crawford]

John,
You bring up good points. But we need better alignment between the industry (ie: customer) and those that are providing the incentives. The Data Center Pulse industry group (http://www.datacenterpulse.com/) represents data center owner-operators to create a common voice for the customer. At the conference last week, one of the points discussed is that there is quite a bit of money on the table for organizations to leverage and take advantage of (ie: rebates). However, most poeple don't know how to go about doing it. We, as an industry, need to make this process easier, more streamlined and more visible! BTW, we're talking about Billions (not just millions) available...so it is significant!

Tim

 

______________________________________
Tim M. Crawford
650.804.1300

timmcr...@yahoo.com
http://www.linkedin.com/in/timcrawford

---

From: John R Savageau <sava...@pacific-tier.com>
To: cloud-c...@googlegroups.com
Sent: Thursday, February 26, 2009 7:17:22 PM

[C Wegrzyn]

As I said the number one bullet point that came up (on internal or
external cloud data storage fabric) was no vendor lock in. This was told
to me time and again by the very largest users of data storage.

Chuck Wegrzyn

[C Wegrzyn]

John,

Those are "opex costs".

Chuck Wegrzyn

John R Savageau wrote:
> Tim
>
>
>
> There is definitely the green argument. Less physical data centers,
> better power and cooling efficiency, less cost in building out and
> maintaining data center infrastructure, potential for “green” credits
> (such as in Australia), and other benefits - not the least being public
> relations.
>
>
>
> John
>
>
>
> John Savageau
>
> CTO, CRG West
>
> One Wilshire Building
>
> Los Angeles
>
>
>

> *From:* cloud-c...@googlegroups.com
> [mailto:cloud-c...@googlegroups.com] *On Behalf Of *Tim M. Crawford
> *Sent:* Thursday, February 26, 2009 4:49 PM
> *To:* cloud-c...@googlegroups.com
> *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA

>
>
>
> This is great news. But I think you're leaving quite a bit on the table
> by only focusing on CapEx/ OpEx. There is another aspect to consider: As
> more vendors start to develop their cloud offerings (SaaS, PaaS and
> IaaS) and create increased competition, it will drive the costs down in
> an area that is already headed toward a commodity business (ie:
> infrastructure). This means that you need to volume to cover the reduced
> margins that competition will bring.
>
> We need to get beyond cloud solutions just being a cost reduction. Cost
> reductions are definitely a positive thing. They still need to be part
> of the evaluation. However, they should be a check-box rather the key
> reason for moving. The problem is that CFO's and CIO's will see through
> the cost reductions and consider the risks too. The balance between just
> risk and cost reduction is equal to negative without any other
> considerations. There are additional benefits that will come with the
> change. This is where I believe the real value is for IT organizations.
> Cloud providers that look beyond the better-faster-cheaper argument as
> their key driver will be the ones that ultimately prosper.
>
> That being said, have you considered what other value-adds the bring moves?
>
> Tim
>
>
>

> ------------------------------------------------------------------------
>
> *From:* C Wegrzyn <eprpa...@gmail.com>
> *To:* cloud-c...@googlegroups.com
> *Sent:* Thursday, February 26, 2009 4:32:28 PM
> *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA

>
>
> I can second that.In the cloud storage fabric I am building to sell I
> focus on reduced capex and opex. That seems to get great traction in the
> marketplace.
>
> Chuck Wegrzyn
>

> dustin....@gmail.com <mailto:dustin....@gmail.com> wrote:
>> Right now clouds are focused on CAPEX, but hopefully cloud management
>> tools evolve, we will hear about clouds reducing OPEX as well.
>>
>> I wouldn't argue that individually SOA (not web services) and cloud
>> computing have different value propositions. I do think however that
>> the most interesting value propostions are those that exploit the
>> capabilities of SOA within cloud environments.
>>
>> On Feb 26, 2:13 pm, Sundeep Singatwaria
>> <sundeep_singatwa...@hotmail.com

> <mailto:sundeep_singatwa...@hotmail.com>> wrote:
>>> Main value proposition of SOA is that it allows to build composite
> apps from different entities or to enable legacy systems to effectively
> talk to newer apps without wholesale rewrites. Cloud computing is
> certainly not that.
>>>
>>> Clouds can certainly use the some of the standards of Web services but
> the two technologies have different value propositions for enterprises.
>>>
>>> Composite SOA enabled apps would be better served with a cloud
> computing infrastructure so that the partners that talk to my SOA
> enabled apps can get a consistent response time from my apps because of
> elasticity of my compute infrastruture.
>>>
>>> SOA brings savings on OPEX side whereas Clouds on CAPEX.
>>>
>>>
>>>
>>>
>>>
>>>> Date: Thu, 26 Feb 2009 05:40:35 -0800
>>>> Subject: [ Cloud Computing ] Cloud Computing and SOA

>>>> From: dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com>
>>>> To: cloud-c...@googlegroups.com

> <mailto:cloud-c...@googlegroups.com>

[C Wegrzyn]

This is the "lower capex cost" view. The idea is to take legacy systems
and to be able to incorporate them into the storage fabric. I'm finding
the companies just don't want to throw things away now a days or buy any
new hardware!

Chuck Wegrzyn

Tim M. Crawford wrote:
> Thanks Dustin for clarifying. You're right that this is a critical issue
> for IT organizations. IT orgs need to have the flexibility to move
> services (and data) from one vendor to another. And it need to be
> relatively easy to do. Vendor Lock-In is a real issue that needs a solution.
>
> One possibility is for a new market of vendors to develop ways to
> integrate disparet sub-components. This would be somewhat similar to
> what Rightscale has done for the infrastructure level. Another
> possibility is that interoperability becomes a real possibility between
> vendors. But the market will most likely need to mature a bit before
> that can happen.
>
> Tim
>

> ------------------------------------------------------------------------
> *From:* "dustin....@gmail.com" <dustin....@gmail.com>
> *To:* Cloud Computing <cloud-c...@googlegroups.com>
> *Sent:* Thursday, February 26, 2009 7:18:09 PM

> *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA
>
>

> This gives them the freedom to move across vendors with their data and
> applications as well. Look at the recent announcement of Coghead going
> out of business, and you can see the tight spot some companies are in.
>
> On Feb 26, 7:06 pm, C Wegrzyn <eprparad...@gmail.com

>> > *From:* C Wegrzyn <eprparad...@gmail.com <mailto:eprparad...@gmail.com>>
>> > *To:* cloud-c...@googlegroups.com
> <mailto:cloud-c...@googlegroups.com>

>> > *Sent:* Thursday, February 26, 2009 4:32:28 PM
>> > *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA
>>
>> > I can second that.In the cloud storage fabric I am building to sell I
>> > focus on reduced capex and opex. That seems to get great traction in the
>> > marketplace.
>>
>> > Chuck Wegrzyn
>>
>> > dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com>

> <mailto:dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com>> wrote:
>> >> Right now clouds are focused on CAPEX, but hopefully cloud management
>> >> tools evolve, we will hear about clouds reducing OPEX as well.
>>
>> >> I wouldn't argue that individually SOA (not web services) and cloud
>> >> computing have different value propositions. I do think however that
>> >> the most interesting value propostions are those that exploit the
>> >> capabilities of SOA within cloud environments.
>>
>> >> On Feb 26, 2:13 pm, Sundeep Singatwaria
>> >> <sundeep_singatwa...@hotmail.com
> <mailto:sundeep_singatwa...@hotmail.com>

>> > <mailto:sundeep_singatwa...@hotmail.com

> <mailto:sundeep_singatwa...@hotmail.com>>> wrote:
>> >>> Main value proposition of SOA is that it allows to build composite
>> > apps from different entities or to enable legacy systems to effectively
>> > talk to newer apps without wholesale rewrites. Cloud computing is
>> > certainly not that.
>>
>> >>> Clouds can certainly use the some of the standards of Web services
>> > but the two technologies have different value propositions for
> enterprises.
>>
>> >>> Composite SOA enabled apps would be better served with a cloud
>> > computing infrastructure so that the partners that talk to my SOA
>> > enabled apps can get a consistent response time from my apps because of
>> > elasticity of my compute infrastruture.
>>
>> >>> SOA brings savings on OPEX side whereas Clouds on CAPEX.
>>
>> >>>> Date: Thu, 26 Feb 2009 05:40:35 -0800
>> >>>> Subject: [ Cloud Computing ] Cloud Computing and SOA
>> >>>> From: dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com>

> <mailto:dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com>>
>> >>>> To: cloud-c...@googlegroups.com
> <mailto:cloud-c...@googlegroups.com>
>> > <mailto:cloud-c...@googlegroups.com

[Peglar, Robert]

It is difficult to place "legacy systems" and "incorporate them into a cloud" in the same sentence, unless the word "not" is used :-)

Interestingly enough, it's not so much that companies don't _want_ to buy new hardware or throw the old junk out - it's that they can't. They are locked in. This, of course, is one of the macro business drivers behind cloud compute and cloud storage in the first place - avoid lock-in within their own datacenter.

OTOH, many CIOs have validated the old cliché of "be careful what you ask for, you may get it." Many CIOs still buy hardware as if it were 1985. It is a never-ending source of puzzlement for me to hear CIOs universally pan vendor-lock-in, and then turn around internally and buy the same old hardware (perhaps in new clothes) from the same old vendor and thus perpetuate the lock-in cycle. They are indeed getting what they asked for.

Rob

No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 8.0.237 / Virus Database: 270.11.3/1975 - Release Date: 02/27/09 07:05:00

[C Wegrzyn]

All I can tell you is my experience in talking to customers about
storage fabric (I hate cloud and grid and prefer the more generic term
fabric). I hear time and again (for the last 2+ years) that they want
solutions that will incorporate, as much as possible, legacy storage
systems they have deployed (by this I mean SAN, NAS, iSCSI, Centera,
etc). They also don't want to buy racks and racks of new servers,
instead wanting to use/reuse already deployed banks of servers. Part of
this is to keep staffing costs and training lower and to know what their
deployment costs are going to be.

It is a whole different world...

Chuck Wegrzyn

[Peglar, Robert]

It is a different world, indeed. FYI, the use of the phrase 'storage fabric' leads many to immediately jump to a Fibre Channel fabric...just a heads-up to be careful with terminology, as is the case with much in clouds these days.

Again, what they want is contrary to their buying/operating behavior, though. If they really wanted to incorporate storage systems into a cloud architecture, and keep their FTE costs to a minimum, they wouldn't have bought what they did in the first place. You can't have it both ways. But, as in a good 12-step program, the first step is to recognize your own problem, and I will give the CIOs credit here - they (now) realize the trap they got themselves into.

Another problem - particularly acute in data storage in traditional data centers - is that using (or continuing to use) legacy arrays is very inefficient, since the gestalt effect of exponential data growth, SSD and HDD pricing curves, concerns over inputs (amps, BTUs, rack U) is a very strong counter-force. It is often optimal to indeed junk the old and bring in the new. YMMW, of course, but how long does one continue to drive the '59 Edsel? If the answer is "because I can", or "because I think I have to, since I am only trained to drive Edsels" then we have a different problem.

It is a fallacy, however, to think that training on or operating a new (non-legacy) storage system is difficult. There are many examples to the contrary.

Rob




---

Robert Peglar
Vice President, Technology, Storage Systems Group

Email: Robert...@xiotech.com
Office: 952 983 2287
Mobile: 314 308 6983
Fax: 636 532 0828

Xiotech Corporation
1606 Highland Valley Circle
Wildwood, MO 63005

www.xiotech.com : www.xiotech.com/demo : Toll-Free 866 472 6764

[mij...@sbcglobal.net]

"(I hate cloud and grid and prefer the more generic term
fabric)."

Nice statemenb. It's like visiting US and saying: "I hate Americans" "I like the more generic term of people " .

But not all people are American and the word fabric is just old fashioned buzz., meaning absolutely nothing out of context. I guessed you mean "cloud's fabric"

This makes sense. Because this is something the outside user should not know about it, in order to have a cloud. If they have to learn that fabric, we are back to the legacy nightmarish enterprise apps that are hard to use and deploy and manage

If you dislike also "cloud fabric", maybe you are posting in the wrong group

M



Sent from my Verizon Wireless BlackBerry

-----Original Message-----
From: C Wegrzyn <eprpa...@gmail.com>

Date: Fri, 27 Feb 2009 09:49:57

To: <cloud-c...@googlegroups.com>
Subject: [ Cloud Computing ] Re: Cloud Computing and SOA

All I can tell you is my experience in talking to customers about
storage fabric (I hate cloud and grid and prefer the more generic term
fabric). I hear time and again (for the last 2+ years) that they want
solutions that will incorporate, as much as possible, legacy storage
systems they have deployed (by this I mean SAN, NAS, iSCSI, Centera,
etc). They also don't want to buy racks and racks of new servers,
instead wanting to use/reuse already deployed banks of servers. Part of
this is to keep staffing costs and training lower and to know what their
deployment costs are going to be.

It is a whole different world...

Chuck Wegrzyn

Peglar, Robert wrote:
> It is difficult to place "legacy systems" and "incorporate them into a cloud" in the same sentence, unless the word "not" is used :-)
>

> Interestingly enough, it's not so much that companies don't_want_ to buy new hardware or throw the old junk out - it's that they can't. They are locked in. This, of course, is one of the macro business drivers behind cloud compute and cloud storage in the first place - avoid lock-in within their own datacenter.

[C Wegrzyn]

Rob,

Thanks for the point on fibre channel. I'll keep that in mind though it
hasn't manifest itself until your comment.

I don't think what they want is contrary to their buying behavior. I
can tell you they don't know any other way is possible. Once they get
educated on the alternatives they start to look differently at their
storage and how to deal with it.

They are receptive to new approaches having seen Google leverage a
simpler storage solution and not face all the problems.

[C Wegrzyn]

Funny you should say Americans vs. People. When I talk about people
living in the U.S and I want to limit them to a geo. area I will talk
about Americans. But when I talk about immigration and the world, I see
People. Borders are artificial, caused more problems then good and are
foolish.

That said I still like "storage fabric". If you don't, please don't use
it. I think of the cloud as composed of two different things: computing
devices that provide computing services, and computing devices that
provide storage. Yes I know they can do both I just don't think of it
that way.

Chuck Wegrzyn

[Jan Klincewicz]

      As someone who made a living off hardware vendor lock-in for 10 years, I have to concur.  Old habits die hard, and server manufacturers work very hard to avoid "commoditization" of their wares.  Whether it comes from a custom BIOS, remote access devices, Management software, or a whole proprietary architecture (think blades !!), it is always in a server/storage/networking vendor's best interests to get customers "addicted" to their products such that barriers to change  are a difficult as possible.

CIOs, though, rarely make the buying decisions at an X86 level (though they tend to be the ones keeping AS/400's etc. alive.)  It is usually the Directors who set standards for "Wintel" servers (and that title ironically still applies to Linux running on AMD !)

I never had a moral problem selling what I did, as it was very good quality hardware, and (at least at the SE level) we never intentionally sold customers anything they didn't need.

The reason I got out of the hardware biz was that very obviously, Virtualization was cannibalizing the "one-app-one-server" paradigm, and that seemed to be a more sensible place to be, at least in the long term ..  I addition to ridiculous consolidation ratios, Virtual servers tend to de-emphasize the necessity for all the proprietary goodies that add margin to servers.

I still think Tier-1 vendors make better servers than white-box counterparts (though in very high volumes, a Google may not care) and SOMEONE still needs to monitor and manage hardware.

Of course, this all matters to Cloud PROVIDERS and not CONSUMERS (who in the end benefit most from not having to worry about such matters anymore EVER !!)

--
Cheers,
Jan

[Jan Klincewicz]

"Storage Fabric" at one time had a very specific meaning ..to distinguish SWITCHED FC vs. Arbitrated Loop (the earlier implementation, which was analogous to Token Ring.)

--
Cheers,
Jan

[Peglar, Robert]

Well said Chuck...they don't know any other way is possible. Still, one has to wonder about their (lack of) due diligence on the subject. Enough said.

As for Google, all the problems surrounding proper care and feeding of storage are still present...SATA drives velcro'd to baking pans...but, to a CIO, since it's not in his/her datacenter, it's "someone else's" problem.

[Peglar, Robert]

Still has that meaning…hasn’t gone away.

 

As for FC-AL being analogous to Token Ring, well, it’s OK to think of it that way.  But AL is not really a token-passing scheme, on the inside.  It was, truth be told, a solution to an economic problem in the late ‘90s – the cost of an FC port.   Made little sense to plug in a $500 disk into a $5,000 FC port, so AL was invented.

 

AL today is a primary reason (one of several) why legacy storage systems don’t behave well at large scale.  Forward-thinking designers are using fabric exclusively, with various interconnects (IB in particular, FC as well) for the storage elements.  No more loops.

 

We now return you to your regularly scheduled (cloud) program J

 

Rob

[C Wegrzyn]

Rob,

I agree. I find that once I spend time talking to them and showing them
a solution that isn't another "stove pipe" they start to see the
possibilities.. Then we get into what they really want in storage not
what they have been fed! That is why I hold out so much hope for cloud
storage and cloud computing - it separates out the artifacts of poor
design/implementation/biased solutions to just the essence of what
storage out to be, or computing.

Chuck

[Sundeep Singatwaria]

 
Do you know how much OPEX there is for management tools of current data centers and cloud management tools.
 
What is SOA and web services. Could you point me to something where SOA is being used outside of Web Services.
 
Web services is certainly SOA and may be not all SOA is web service but I have yet to come across something meaningful that describes a SOA and isn't web service.
 
 
> Date: Thu, 26 Feb 2009 15:12:48 -0800
> Subject: [ Cloud Computing ] Re: Cloud Computing and SOA
> From: dustin....@gmail.com
> To: cloud-c...@googlegroups.com

[Ben Bloch]

I agree that is great.  But, if not about cost and cost savings, then what? 

 

Remote storage, handling spikes, scaling in general – they are all about cost.

 

You ** may ** be able to make the case that a massively scalable app like Google would just not be possible – at any price - without a new architecture – but you can make a stronger case IMV that Google would not be possible/viable unless its costs were controlled – to scale at low cost. 

 

I believe the 3 main drivers to cloud computing will be cost, cost, cost.   Given the evaporating IT budgets of nearly all companies, that is not a bad place to be.

 

- ben

 

 

---

[Greg Pfister]

On Feb 27, 10:51 am, "Peglar, Robert" <Robert_Peg...@xiotech.com>
wrote:
> Still has that meaning...hasn't gone away.
>
> As for FC-AL being analogous to Token Ring, well, it's OK to think of it that way.  But AL is not really a token-passing scheme, on the inside.  It was, truth be told, a solution to an economic problem in the late '90s - the cost of an FC port.   Made little sense to plug in a $500 disk into a $5,000 FC port, so AL was invented.

Agreed.

> AL today is a primary reason (one of several) why legacy storage systems don't behave well at large scale.  Forward-thinking designers are using fabric exclusively, with various interconnects (IB in particular, FC as well) for the storage elements.  No more loops.

To quote someone's teenage daughter's text messages:

OMG!

There are actually people installing FC-AL, today!?!?!

While on another thread they're discussing CEE... Seldom before has
the distance between reality and the leading edge seemed to me to
include so many light-years.

I'm kinda happy to hear about IB, though, since I was very heavily
involved at one time.

Greg Pfister
http://perilsofparallel.blogspot.com/

> We now return you to your regularly scheduled (cloud) program J
>
> Rob
>
> From: cloud-c...@googlegroups.com [mailto:cloud-c...@googlegroups.com] On Behalf Of Jan Klincewicz
> Sent: Friday, February 27, 2009 10:22 AM
> To: cloud-c...@googlegroups.com
> Subject: [ Cloud Computing ] Re: Cloud Computing and SOA
>
> "Storage Fabric" at one time had a very specific meaning ..to distinguish SWITCHED FC vs. Arbitrated Loop (the earlier implementation, which was analogous to Token Ring.)
>

> >>> *From:* "dustin.amrh...@gmail.com" <dustin.amrh...@gmail.com>

> ...
>
> read more »

[John D. Mitchell]

On Friday 2009.02.27, at 06:35 , Peglar, Robert wrote:
[...]

> It is difficult to place "legacy systems" and "incorporate them into
> a cloud" in the same sentence, unless the word "not" is used :-)

Since nobody else has jumped in with this yet...

Of course such legacy systems *can* be fit in via a hybrid datacenter
+ cloud approach. In fact, the whole "SOA" thing is an enabler to
make that, generally speaking, a lot easier in a lot more cases.

> Interestingly enough, it's not so much that companies don't _want_
> to buy new hardware or throw the old junk out - it's that they
> can't. They are locked in. This, of course, is one of the macro
> business drivers behind cloud compute and cloud storage in the first
> place - avoid lock-in within their own datacenter.

FWIW, I don't believe it's really a "can't" issue so much as a
"won't" (even that that's because they are too chickensh!t :-). Yes,
there are contracts, etc. but it's still fundamentally a question of
choice and the will to make decisions and live with the consequences.

Take care,
John

[Greg Pfister]

On Feb 27, 8:13 am, C Wegrzyn <eprparad...@gmail.com> wrote:
> As I said the number one bullet point that came up (on internal or
> external cloud data storage fabric) was no vendor lock in. This was told
> to me time and again by the very largest users of data storage.

I'm not at all surprised that the largest users would say that. They
have enough focus on it to train to or acquire more long-term thoughts
- and be willing to pay (some) overhead for non-lock-in, or at least
appear to be willing before they start serious haggling.

I know great big fat juicy accounts attract more attention than 10K
ants, but can you say if smaller users were as concerned?

Greg Pfister
http://perilsofparallel.blogspot.com/

>
> Chuck Wegrzyn

> >>>> Windows Live™ Hotmail®…more than just e-mail.http://windowslive.com/howitworks?ocid=TXT_TAGLM_WL_t2_hm_justgotbett...quoted text -

[John D. Mitchell]

On Friday 2009.02.27, at 08:30 , Peglar, Robert wrote:
[...]

> As for Google, all the problems surrounding proper care and feeding
> of storage are still present...SATA drives velcro'd to baking
> pans...but, to a CIO, since it's not in his/her datacenter, it's
> "someone else's" problem.

Well, yes, and no. As in the thread on privacy, etc. issues, the
corporation (and it's officers) are on the hook for the information
and so the storage, etc. of said information *is* their problem
regardless of where it's located.

IMHO, if anything, the laws will get much more strict on defining
liability so that corporations cannot sidestep their responsibilities
on these issues just because they don't control the "cloud" in which
the information is stored, manipulated, etc.

John

[Peglar, Robert]

Greg, to your OMG point, many storage arrays on the market today are FC-AL in the 'back-end'...so yes, people are installing FC-AL these days, whether they realize it or not. Some arrays use (or are moving towards) serial SCSI (SAS) but most of the 'industrial storage iron' is still FC-AL under the hood. As I said, best-of-breed are using FC-SW (fabric) for the backend and not loop. But loop is cheap, so most array vendors still design it in.

[Peglar, Robert]

Well said John.

-----Original Message-----
From: cloud-c...@googlegroups.com
[mailto:cloud-c...@googlegroups.com] On Behalf Of John D. Mitchell
Sent: Friday, February 27, 2009 3:19 PM
To: cloud-c...@googlegroups.com

[C Wegrzyn]

Let me add something very affirmative - I know it can be done since
Twisted Storage does it and has for the past 2+ years.

Chuck Wegrzyn

[ben]

I agree that is great. But, if not about cost and cost savings, then
what?

Remote storage, handling spikes, scaling in general – they are all
about cost.

You ** may ** be able to make the case that a massively scalable app
like Google would just not be possible – at any price - without a new
architecture – but you can make a stronger case IMV that Google would
not be possible/viable unless its costs were controlled – to scale at
low cost.

I believe the 3 main drivers to cloud computing will be cost, cost,
cost. Given the evaporating IT budgets of nearly all companies, that
is not a bad place to be.

- ben

On Feb 26, 7:49 pm, "Tim M. Crawford" <timmcrawf...@yahoo.com> wrote:
> This is great news. But I think you're leaving quite a bit on the table by only focusing on CapEx/ OpEx. There is another aspect to consider: As more vendors start to develop their cloud offerings (SaaS, PaaS and IaaS) and create increased competition, it will drive the costs down in an area that is already headed toward a commodity business (ie: infrastructure). This means that you need to volume to cover the reduced margins that competition will bring.
>
> We need to get beyond cloud solutions just being a cost reduction. Cost reductions are definitely a positive thing. They still need to be part of the evaluation. However, they should be a check-box rather the key reason for moving. The problem is that CFO's and CIO's will see through the cost reductions and consider the risks too. The balance between just risk and cost reduction is equal to negative without any other considerations. There are additional benefits that will come with the change. This is where I believe the real value is for IT organizations. Cloud providers that look beyond the better-faster-cheaper argument as their key driver will be the ones that ultimately prosper.
>
> That being said, have you considered what other value-adds the bring moves?
>
> Tim
>

> ________________________________
> From: C Wegrzyn <eprparad...@gmail.com>
> To: cloud-c...@googlegroups.com

> Sent: Thursday, February 26, 2009 4:32:28 PM

> Subject: [ Cloud Computing ] Re: Cloud Computing and SOA
>

> I can second that.In the cloud storage fabric I am building to sell I
> focus on reduced capex and opex. That seems to get great traction in the
> marketplace.
>
> Chuck Wegrzyn
>

> dustin.amrh...@gmail.com wrote:
> > Right now clouds are focused on CAPEX, but hopefully cloud management
> > tools evolve, we will hear about clouds reducing OPEX as well.
>
> > I wouldn't argue that individually SOA (not web services) and cloud
> > computing have different value propositions. I do think however that
> > the most interesting value propostions are those that exploit the
> > capabilities of SOA within cloud environments.
>
> > On Feb 26, 2:13 pm, Sundeep Singatwaria

> > <sundeep_singatwa...@hotmail.com> wrote:
> >> Main value proposition of SOA is that it allows to build composite apps from different entities or to enable legacy systems to effectively talk to newer apps without wholesale rewrites. Cloud computing is certainly not that.
>

> >> Windows Live™ Hotmail®…more than just e-mail.http://windowslive.com/howitworks?ocid=TXT_TAGLM_WL_t2_hm_justgotbett...Hide quoted text -

[C Wegrzyn]

I agree and it is why over the last 2+ years I've been talking about
cloud storage designs that focus on reduced capex and opex costs. Once
you decide that is a goal, it starts to affect the design of the system,
making you think of issues related to deployment, "green" features of
operation, even all the way to staff. It has stopped me from thinking
like a programmer and more like a customer!

Chuck Wegrzyn

[Tim M. Crawford]

I should have been clearer in my statement. You're right that many things ultimately boil down to costs. The point that I was trying to make is that I keep hearing providers talk in terms of cloud solutions being 'cheaper' than the alternatives. Here's what I was trying to point out:

The argument of better-faster-cheaper needs to be a check-box, not the focus of the argument. There needs to be significant value-add to the CFO/ CIO beyond just the cheaper argument alone. For example, with IaaS, cloud solutions can provide flexibility, disaster recover/ business continuity, geo-diversity, CapEx to OpEx movement, etc. Yes, all of those can ultimately be turned into a value argument. But as a provider, one needs to start with the business problem that is being solved and then explain how they're solving it compared with the alternative. There also needs to be a discussion around risk that the IT organization will assume with the changes. That's another point that is often glossed over. This partnership of clearly outlining the business problem being solved and the pros/ cons associated with the change will gain the interest of the CFO/ CIO.

PaaS and SaaS would have a similar list of items beyond simply stating better-faster-cheaper.

-t

---

From: C Wegrzyn <eprpa...@gmail.com>
To: cloud-c...@googlegroups.com

Sent: Saturday, February 28, 2009 8:57:54 AM

[Chas Ginger]

Jim,

I understood. I was saying that if you focus on costs, you get a
different design and feature set than if you just focus on functionality.

Chuck

Tim M. Crawford wrote:
> I should have been clearer in my statement. You're right that many
> things ultimately boil down to costs. The point that I was trying to
> make is that I keep hearing providers talk in terms of cloud solutions
> being 'cheaper' than the alternatives. Here's what I was trying to point
> out:
>
> The argument of better-faster-cheaper needs to be a check-box, not the
> focus of the argument. There needs to be significant value-add to the
> CFO/ CIO beyond just the cheaper argument alone. For example, with IaaS,
> cloud solutions can provide flexibility, disaster recover/ business
> continuity, geo-diversity, CapEx to OpEx movement, etc. Yes, all of
> those can ultimately be turned into a value argument. But as a provider,
> one needs to start with the business problem that is being solved and
> then explain how they're solving it compared with the alternative. There
> also needs to be a discussion around risk that the IT organization will
> assume with the changes. That's another point that is often glossed
> over. This partnership of clearly outlining the business problem being
> solved and the pros/ cons associated with the change will gain the
> interest of the CFO/ CIO.
>
> PaaS and SaaS would have a similar list of items beyond simply stating
> better-faster-cheaper.
>
> -t
>

> ------------------------------------------------------------------------

> *From:* C Wegrzyn <eprpa...@gmail.com>
> *To:* cloud-c...@googlegroups.com

> *Sent:* Saturday, February 28, 2009 8:57:54 AM
> *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA

>
>
> I agree and it is why over the last 2+ years I've been talking about
> cloud storage designs that focus on reduced capex and opex costs. Once
> you decide that is a goal, it starts to affect the design of the system,
> making you think of issues related to deployment, "green" features of
> operation, even all the way to staff. It has stopped me from thinking
> like a programmer and more like a customer!
>
> Chuck Wegrzyn
>
> ben wrote:
>> I agree that is great. But, if not about cost and cost savings, then
>> what?
>>
>> Remote storage, handling spikes, scaling in general – they are all
>> about cost.
>>
>> You ** may ** be able to make the case that a massively scalable app
>> like Google would just not be possible – at any price - without a new
>> architecture – but you can make a stronger case IMV that Google would
>> not be possible/viable unless its costs were controlled – to scale at
>> low cost.
>>
>> I believe the 3 main drivers to cloud computing will be cost, cost,
>> cost. Given the evaporating IT budgets of nearly all companies, that
>> is not a bad place to be.
>>
>> - ben
>>
>>
>> On Feb 26, 7:49 pm, "Tim M. Crawford" <timmcrawf...@yahoo.com

> <mailto:timmcrawf...@yahoo.com>> wrote:
>>> This is great news. But I think you're leaving quite a bit on the
> table by only focusing on CapEx/ OpEx. There is another aspect to
> consider: As more vendors start to develop their cloud offerings (SaaS,
> PaaS and IaaS) and create increased competition, it will drive the costs
> down in an area that is already headed toward a commodity business (ie:
> infrastructure). This means that you need to volume to cover the reduced
> margins that competition will bring.
>>>
>>> We need to get beyond cloud solutions just being a cost reduction.
> Cost reductions are definitely a positive thing. They still need to be
> part of the evaluation. However, they should be a check-box rather the
> key reason for moving. The problem is that CFO's and CIO's will see
> through the cost reductions and consider the risks too. The balance
> between just risk and cost reduction is equal to negative without any
> other considerations. There are additional benefits that will come with
> the change. This is where I believe the real value is for IT
> organizations. Cloud providers that look beyond the
> better-faster-cheaper argument as their key driver will be the ones that
> ultimately prosper.
>>>
>>> That being said, have you considered what other value-adds the bring
> moves?
>>>
>>> Tim
>>>
>>> ________________________________

>>> From: C Wegrzyn <eprparad...@gmail.com <mailto:eprparad...@gmail.com>>
>>> To: cloud-c...@googlegroups.com
> <mailto:cloud-c...@googlegroups.com>

>>> Sent: Thursday, February 26, 2009 4:32:28 PM
>>> Subject: [ Cloud Computing ] Re: Cloud Computing and SOA
>>>
>>> I can second that.In the cloud storage fabric I am building to sell I
>>> focus on reduced capex and opex. That seems to get great traction in the
>>> marketplace.
>>>
>>> Chuck Wegrzyn
>>>
>>> dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com> wrote:
>>>> Right now clouds are focused on CAPEX, but hopefully cloud management
>>>> tools evolve, we will hear about clouds reducing OPEX as well.
>>>> I wouldn't argue that individually SOA (not web services) and cloud
>>>> computing have different value propositions. I do think however that
>>>> the most interesting value propostions are those that exploit the
>>>> capabilities of SOA within cloud environments.
>>>> On Feb 26, 2:13 pm, Sundeep Singatwaria
>>>> <sundeep_singatwa...@hotmail.com

> <mailto:sundeep_singatwa...@hotmail.com>> wrote:
>>>>> Main value proposition of SOA is that it allows to build composite
> apps from different entities or to enable legacy systems to effectively
> talk to newer apps without wholesale rewrites. Cloud computing is
> certainly not that.
>>>>> Clouds can certainly use the some of the standards of Web services
> but the two technologies have different value propositions for enterprises.
>>>>> Composite SOA enabled apps would be better served with a cloud
> computing infrastructure so that the partners that talk to my SOA
> enabled apps can get a consistent response time from my apps because of
> elasticity of my compute infrastruture.
>>>>> SOA brings savings on OPEX side whereas Clouds on CAPEX.
>>>>>> Date: Thu, 26 Feb 2009 05:40:35 -0800
>>>>>> Subject: [ Cloud Computing ] Cloud Computing and SOA

>>>>>> From: dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com>
>>>>>> To: cloud-c...@googlegroups.com
> <mailto:cloud-c...@googlegroups.com>

[Tim M. Crawford]

Maybe we're talking about the same thing. I'm suggesting that you need to consider both.

Tim

---

From: Chas Ginger <eprpa...@gmail.com>
To: cloud-c...@googlegroups.com
Sent: Sunday, March 1, 2009 3:24:43 PM

[dustin....@gmail.com]

I thought this article was interesting and pertinent in light of all
the posts dealing with CAPEX and OPEX:

http://gevaperry.typepad.com/main/2009/01/accounting-for-clouds-stop-saying-capex-vs-opex.html

On Mar 1, 4:44 pm, "Tim M. Crawford" <timmcrawf...@yahoo.com> wrote:
> I should have been clearer in my statement. You're right that many things ultimately boil down to costs. The point that I was trying to make is that I keep hearing providers talk in terms of cloud solutions being 'cheaper' than the alternatives. Here's what I was trying to point out:
>
> The argument of better-faster-cheaper needs to be a check-box, not the focus of the argument. There needs to be significant value-add to the CFO/ CIO beyond just the cheaper argument alone. For example, with IaaS, cloud solutions can provide flexibility, disaster recover/ business continuity, geo-diversity, CapEx to OpEx movement, etc. Yes, all of those can ultimately be turned into a value argument. But as a provider, one needs to start with the business problem that is being solved and then explain how they're solving it compared with the alternative. There also needs to be a discussion around risk that the IT organization will assume with the changes. That's another point that is often glossed over. This partnership of clearly outlining the business problem being solved and the pros/ cons associated with the change will gain the interest of the CFO/ CIO.
>
> PaaS and SaaS would have a similar list of items beyond simply stating better-faster-cheaper.
>
> -t
>
> ________________________________

> From: C Wegrzyn <eprparad...@gmail.com>

> >>>> Windows Live™ Hotmail®…more than just e-mail.http://windowslive.com/howitworks?ocid=TXT_TAGLM_WL_t2_hm_justgotbett...quoted text -
> >>>> - Show quoted text -- Hide quoted text -

[Miha Ahronovitz]

Geva Perry's article recommended by Dustin is spot on. CAPEX or OPEX is an accounting conversation. When we say "we save money" , in the cloud context it is wrong. People think of a Data Center, where we have a cost center. If it cost less (CAPEX, OPEX or "SHMATEX" :-) ) be it. It is not dependent on how we calculate the lower cost.

But if we bill the cloud resources, we can always have a profit center, even in a private cloud for internal usage. A real cloud is a profit center, not a cost center.If we make more money than we spend money, the accountants havea pleasant job.

Miha

[Jan Klincewicz]

       Thanks for sharing that very scholarly post.  I'm sure many people were unclear about the distinction.

BTW, I am also an MBA and have regretfully found myself working for accountants for many years.  Although everything you say is true from an accounting perspective, I think what people are often referring to in their arguments are what budgetary funds are made available.

      When I used to buy IT gear, I used to select internal modems, as they were purchased from the Opex budget and not the Capex budget (which was often frozen.)  The fact that internal modems cost 2X what I could buy an external model, and the fact that each one required a reseller to crack the case and add ANOTHER $100 to the cost did not matter.  This is one example.

       What it came down to was getting stuff done, and IT folks who are  often more creative (or at least devious) than bean counters will find a way, regardless of its eventual affect on the "bottom line."

--
Cheers,
Jan

[Chas Ginger]

Thanks for the reference Dustin. I never believed in the idea of
shifting costs from capex to opex, that is a fool's errand. I believe
they are two independent functions and my idea is to minimize both, at
least as much as possible.

Chuck

[Chas Ginger]

I only mentioned CapEx and OpEx in the context of design issues forced
on you if you try to minimize them. It makes you think differently.

Chuck Wegrzyn

[Tim M. Crawford]

I too have an MBA, but we should not focus on that point.

It's an interesting article and points out some key items. Thanks for the link! It's the comments at the end that I think bring some things to light. It isn't about the differences between CapEx and OpEx from an accounting perspective. It's the "type" of funding that this represents (and Jan explains below). To many, it's just money and just another cost to be spent (one way or another). However, from a practical standpoint one often provides greater flexibility than the other.

Tim

---

From: Jan Klincewicz <jan.kli...@gmail.com>
To: cloud-c...@googlegroups.com
Sent: Monday, March 2, 2009 5:42:03 AM

[JimL]

I think in this current economic environment CAPEX V OPEX is more than
an accounting conversation in the context of Cloud.

I discuss this in two posts:

The Economics of Cloud: http://vehera.jsn-server7.com/LiddleBlog/?p=234
Will the credit crunch help cloud computing cross the chasm:
http://vehera.jsn-server7.com/LiddleBlog/?p=271

Jim Liddle
http://www.gigaspaces.com/cloud

On 2 Mar, 04:13, Miha Ahronovitz <mij...@sbcglobal.net> wrote:
> GevaPerry'sarticle recommended by Dustin is spot on. CAPEX or OPEX is an accounting conversation. When we say "we save money" , in the cloud context it is wrong. People think of a Data Center, where we have a cost center. If it cost less (CAPEX, OPEX or "SHMATEX" :-) ) be it. It is not dependent on how we calculate the lower cost.

>
> But if we bill the cloud resources, we can always have a  profit center, even in a private cloud for internal usage. A real cloud is a profit center, not a cost center.If we make more money than we spend money, the accountants havea pleasant job.
>
> Miha
>
> ----- Original Message ----
> From: "dustin.amrh...@gmail.com" <dustin.amrh...@gmail.com>
> To: Cloud Computing <cloud-c...@googlegroups.com>
> Sent: Sunday, March 1, 2009 7:26:19 PM
> Subject: [ Cloud Computing ] Re: Cloud Computing and SOA
>
> I thought this article was interesting and pertinent in light of all
> the posts dealing with CAPEX and OPEX:
>

> http://gevaperry.typepad.com/main/2009/01/accounting-for-clouds-stop-...

> > >>>> Windows Live™ Hotmail®…more than just e-mail.http://windowslive.com/howitworks?ocid=TXT_TAGLM_WL_t2_hm_justgotbett...text -

[Ben Bloch]

The question is how easy is this to do - to move applications and data or at
least to manage of the risk of vendor lockin and even failure - like
Coghead. Data portability is one thing - and a great point/case for cloud
storage provider(s) but application portability is a whole another matter,
particularly for proprietary next gen platforms like Force.com, WaveFront or
AppEngine. How do you move a Force.com app? Obviously need to get the data
- contact lists, opportunity-related info, etc. - but may also have to be
able to get enough of the application to be able to use it.

And managing/mitigating this risk has to factor into the cost models.

At least CogHead is making data available....
http://www.coghead.com/letter-from-chairman

This policy should perhaps be something all customers demand and all vendors
offer, if even in the fine print. Thoughts?

-----Original Message-----
From: cloud-c...@googlegroups.com
[mailto:cloud-c...@googlegroups.com] On Behalf Of

dustin....@gmail.com
Sent: Thursday, February 26, 2009 10:18 PM
To: Cloud Computing
Subject: [ Cloud Computing ] Re: Cloud Computing and SOA

This gives them the freedom to move across vendors with their data and
applications as well. Look at the recent announcement of Coghead going
out of business, and you can see the tight spot some companies are in.

On Feb 26, 7:06 pm, C Wegrzyn <eprparad...@gmail.com> wrote:
> The other "hot button" issue that has come up repeatedly, and it
> surprised me, was no vendor lock in! When I heard it I thought I must
> have misunderstood. But they want a solution which gives them access to
> the data outside of the normal software delivery. Now why they want that
> I am not sure but I imagine it has to do with not feeling indebted to a
> hardware vendor.
>
> C.
>
>
>

> Tim M. Crawford wrote:
> > This is great news. But I think you're leaving quite a bit on the table
> > by only focusing on CapEx/ OpEx. There is another aspect to consider: As
> > more vendors start to develop their cloud offerings (SaaS, PaaS and
> > IaaS) and create increased competition, it will drive the costs down in
> > an area that is already headed toward a commodity business (ie:
> > infrastructure). This means that you need to volume to cover the reduced
> > margins that competition will bring.
>
> > We need to get beyond cloud solutions just being a cost reduction. Cost
> > reductions are definitely a positive thing. They still need to be part
> > of the evaluation. However, they should be a check-box rather the key
> > reason for moving. The problem is that CFO's and CIO's will see through
> > the cost reductions and consider the risks too. The balance between just
> > risk and cost reduction is equal to negative without any other
> > considerations. There are additional benefits that will come with the
> > change. This is where I believe the real value is for IT organizations.
> > Cloud providers that look beyond the better-faster-cheaper argument as
> > their key driver will be the ones that ultimately prosper.
>
> > That being said, have you considered what other value-adds the bring
moves?
>
> > Tim
>

> > ------------------------------------------------------------------------
> > *From:* C Wegrzyn <eprparad...@gmail.com>
> > *To:* cloud-c...@googlegroups.com
> > *Sent:* Thursday, February 26, 2009 4:32:28 PM
> > *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA

>
> > I can second that.In the cloud storage fabric I am building to sell I
> > focus on reduced capex and opex. That seems to get great traction in the
> > marketplace.
>
> > Chuck Wegrzyn
>

> > dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com> wrote:
> >> Right now clouds are focused on CAPEX, but hopefully cloud management
> >> tools evolve, we will hear about clouds reducing OPEX as well.
>
> >> I wouldn't argue that individually SOA (not web services) and cloud
> >> computing have different value propositions. I do think however that
> >> the most interesting value propostions are those that exploit the
> >> capabilities of SOA within cloud environments.
>
> >> On Feb 26, 2:13 pm, Sundeep Singatwaria
> >> <sundeep_singatwa...@hotmail.com

> > <mailto:sundeep_singatwa...@hotmail.com>> wrote:
> >>> Main value proposition of SOA is that it allows to build composite
> > apps from different entities or to enable legacy systems to effectively
> > talk to newer apps without wholesale rewrites. Cloud computing is
> > certainly not that.
>
> >>> Clouds can certainly use the some of the standards of Web services
> > but the two technologies have different value propositions for
enterprises.
>
> >>> Composite SOA enabled apps would be better served with a cloud
> > computing infrastructure so that the partners that talk to my SOA
> > enabled apps can get a consistent response time from my apps because of
> > elasticity of my compute infrastruture.
>
> >>> SOA brings savings on OPEX side whereas Clouds on CAPEX.
>
> >>>> Date: Thu, 26 Feb 2009 05:40:35 -0800
> >>>> Subject: [ Cloud Computing ] Cloud Computing and SOA

> >>>> From: dustin.amrh...@gmail.com <mailto:dustin.amrh...@gmail.com>
> >>>> To: cloud-c...@googlegroups.com

> > <mailto:cloud-c...@googlegroups.com>

[JL Valente]

Personally I don't think that any of the larger PaaS vendors including Microsoft, SalesForce or Google have any interest in application portability. Why would they?

The level of capabilities and integration offered within their platform is what separate them from their competition and more importantly what makes those platforms so compelling to developers in the first place. They seem to have enough traction and support to continue down this path.
The good news is that they have never misled developers about their intentions. So if folks are looking for a portable platform then either they need to consider an IaaS vendor and take the necessary measures to limit lock-in or they need to work with a PaaS vendor that can provide higher degrees of portability like EngineYard or LongJump.

> >>> Windows Live(tm) Hotmail®...more than just

[Chris Marino]

+1

[Greg Pfister]

+1, also.

Providers never want portability.

Clients have to ram it down their throats by refusing to purchase non-
portable solutions.

(Grid was created by users, so standardization / portability / etc.
was built in from the beginning.)

Greg Pfister
http://perilsofparallel.blogspot.com/

On Mar 2, 3:33 pm, Chris Marino <christopher.c.mar...@gmail.com>
wrote:
> +1

[Andre Merzky]

Quoting [Greg Pfister] (Mar 03 2009):
>
> +1, also.
>
> Providers never want portability.
>
> Clients have to ram it down their throats by refusing to purchase non-
> portable solutions.
>
> (Grid was created by users, so standardization / portability / etc.
> was built in from the beginning.)

+1, although I would love to see standardization happening
for clouds, someday...

[Sundeep Singatwaria]

 
Now that the counter of support for exclusivity with cloud provider has gone up - and for all the members that didn't do +1 - I'm assuming that they latently support this. Now given that I would end up being locked in by a cloud provider - if I move my data center within the provider's cloud - I'm pretty much left at the price / performance of that provider. For that very pesky reason if I'm a large enterprise connecting with few 100 sites - I would want to build my internal cloud and not rely on an external cloud.
 
> Date: Tue, 3 Mar 2009 23:15:26 +0100
> From: andre...@gmail.com
> To: cloud-c...@googlegroups.com

> Subject: [ Cloud Computing ] Re: Cloud Computing and SOA
>
>

[JL Valente]

Everybody would certainly respect that. Freedom of choice.

 

However this looks like a steep mountain to climb.

For one  is “building an internal cloud” really part of that large enterprise core competency? I doubt many enterprises would agree.

Second can “wanting” turn easily into “having”? So called public clouds are built and ran by professionals who have focused all their energy, creativity and smart on building a cloud business. Not sure internal IT folks can match that. They are already burdened with so much to do.

Finally will the final product – i.e. an internal cloud – deliver any level of portability whatsoever? Most likely it will be built on hardware and software solutions that will steer this internal cloud into a vendor direction.

 

Sure there will be internal clouds but it will take time. How many firms have any appetite for taking it on right now beyond talk?

> </html

[eprpa...@gmail.com]

When I think of "cloud" I just imagine all those nice desktop computers
sitting around for 16 hours a day unused....

Chuck Wegrzyn

JL Valente wrote:
> Everybody would certainly respect that. Freedom of choice.
>
>
>
> However this looks like a steep mountain to climb.
>
> For one is “building an internal cloud” really part of that large
> enterprise core competency? I doubt many enterprises would agree.
>
> Second can “wanting” turn easily into “having”? So called public clouds
> are built and ran by professionals who have focused all their energy,
> creativity and smart on building a cloud business. Not sure internal IT
> folks can match that. They are already burdened with so much to do.
>
> Finally will the final product – i.e. an internal cloud – deliver any
> level of portability whatsoever? Most likely it will be built on
> hardware and software solutions that will steer this internal cloud into
> a vendor direction.
>
>
>
> Sure there will be internal clouds but it will take time. How many firms
> have any appetite for taking it on right now beyond talk?
>
>
>
>
>

> *From:* cloud-c...@googlegroups.com
> [mailto:cloud-c...@googlegroups.com] *On Behalf Of *Sundeep Singatwaria
> *Sent:* Wednesday, March 04, 2009 11:47 AM
> *To:* cloud-c...@googlegroups.com
> *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA

[Gordon Jackson]

Greetings All,

On Wed, Mar 4, 2009 at 3:45 PM, JL Valente <jlva...@cittio.com> wrote:

For one  is “building an internal cloud” really part of that large enterprise core competency? I doubt many enterprises would agree.

I'm not so sure about that... As we are talking about "Enterprises" - organizations that already have a significant investment in IT - I would argue that they are going to look into getting the most out of the capacity they have already paid for/invested in, and understand very clearly how they do that, before they look beyond the firewall for additional capacity. In this case, organizations have a lot to gain by applying "cloud technologies" internally first.

Second can “wanting” turn easily into “having”? So called public clouds are built and ran by professionals who have focused all their energy, creativity and smart on building a cloud business. Not sure internal IT folks can match that. They are already burdened with so much to do.

I'm wondering if those poor, downtrodden, overburdened, lackadaisical, unintelligent, uncreative, non-professional enterprise IT folks might take exception to the suggestion that the application of *any* technology might be beyond them... but I'll let those poor poor souls respond themselves ;)

Seriously, as the technologies that enable Cloud Computing (however you may choose to define it) become more prevalent and more mature, you can bet that they are going to be every bit as interesting to internal IT folks as they are to "public" folks.

Finally will the final product – i.e. an internal cloud – deliver any level of portability whatsoever? Most likely it will be built on hardware and software solutions that will steer this internal cloud into a vendor direction.

Maybe, maybe not, or maybe who cares?

"A good plan executed today is better than a perfect plan executed at some indefinite point in the future." (Patton)

Sure interoperability and portability and no vendor lock-in would be sweet! But if an organization can get what they want, save money, make money, be green, etc... and do that today, at the expense of a risk like being steered into "a vendor direction", I'll bet on that risk being weighed carefully against the potential outcome.

You only have to look at the Cloud's predecessor (or Daddy, or Uncle maybe) Grid, and the work that OGF has done around trying to get portability and interoperability into the mix to see that it's still an ideal. But the reality is that neither the enterprise users nor enterprise vendors are all that interested in it. Maybe the market will decide otherwise in this case - only time will tell...

Sure there will be internal clouds but it will take time. How many firms have any appetite for taking it on right now beyond talk?

It's happening right now and more and more are asking the questions - they are talking. And I'm happy to talk with them ;)

Cheers,

Gordon

-- 
Gordon Jackson
Technical Director, Data Center Automation
Univa UD

www.univaud.com/reliance

[Ben Bloch]

+1  - I agree that providers WANT lock in. That doesn’t mean that is what customers want or need to accept, Cloud Lockin is even more restrictive/risky than a technology stack (LAMP, J2EE, ,Net, etc. ) lockin since cannot even run the application – let alone port it – without access to the provider’s services.

 

/ben

 

 

 

 

---

From: cloud-c...@googlegroups.com [mailto:cloud-c...@googlegroups.com] On Behalf Of Chris Marino

Sent: Monday, March 02, 2009 4:34 PM
To: cloud-c...@googlegroups.com

Subject: [ Cloud Computing ] Re: Cloud Computing and SOA

 

+1

[Nati Shalom]

This is where pure middleware platform providers have an edge.

For us protability is key to our business i.e. the motivation is built into the business.

 

You can see below an reference to a recent post that I just posted that shows how you can deploy and scale JEE application on AWS withotu code changes.

http://natishalom.typepad.com/nati_shaloms_blog/2009/03/its-time-for-auto-scaling-avoid-peak-load-provisioning.html

 

Nati S.

www.gigaspaces.com/cloud

 

 

 

---

From: cloud-c...@googlegroups.com [mailto:cloud-c...@googlegroups.com] On Behalf Of Ben Bloch
Sent: Thursday, March 05, 2009 12:47 AM
To: cloud-c...@googlegroups.com
Subject: [ Cloud Computing ] Re: Cloud Computing and SOA

[Jim Starkey]

You can't have innovation and standardization at the same time. One has
to precede the other. Once a solution or architecture is recognized as
"good enough", market share will generally force convergence, if not
standardization. But not always. Windows, Unix, and Apple never
converged on a drawing package. SQL has been standardized forever, but
not to the degree of interoperability.

Cloud application platforms are in their infancy. Google and Microsoft
are still in the first approximation stage. The database side of the
picture hasn't even started yet.

Complaining about the absence of standardization isn't going to get you
anywhere.

Ben Bloch wrote:
>
> +1 - I agree that providers WANT lock in. That doesn’t mean that is
> what customers want or need to accept, Cloud Lockin is even more
> restrictive/risky than a technology stack (LAMP, J2EE, ,Net, etc. )
> lockin since cannot even run the application – let alone port it –
> without access to the provider’s services.
>
> /ben
>

> ------------------------------------------------------------------------
>
> *From:* cloud-c...@googlegroups.com
> [mailto:cloud-c...@googlegroups.com] *On Behalf Of *Chris Marino
> *Sent:* Monday, March 02, 2009 4:34 PM
> *To:* cloud-c...@googlegroups.com
> *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA

>
> +1
>
> On Mon, Mar 2, 2009 at 12:28 PM, JL Valente <jlva...@cittio.com
> <mailto:jlva...@cittio.com>> wrote:
>
>
> Personally I don't think that any of the larger PaaS vendors including
> Microsoft, SalesForce or Google have any interest in application
> portability. Why would they?
>
> The level of capabilities and integration offered within their
> platform is what separate them from their competition and more
> importantly what makes those platforms so compelling to developers in
> the first place. They seem to have enough traction and support to
> continue down this path.
> The good news is that they have never misled developers about their
> intentions. So if folks are looking for a portable platform then
> either they need to consider an IaaS vendor and take the necessary
> measures to limit lock-in or they need to work with a PaaS vendor that
> can provide higher degrees of portability like EngineYard or LongJump.
>
>
> -----Original Message-----
> From: cloud-c...@googlegroups.com
> <mailto:cloud-c...@googlegroups.com>
> [mailto:cloud-c...@googlegroups.com
> <mailto:cloud-c...@googlegroups.com>] On Behalf Of Ben Bloch
> Sent: Monday, March 02, 2009 11:12 AM
> To: cloud-c...@googlegroups.com

> <mailto:cloud-c...@googlegroups.com>
>
> Subject: [ Cloud Computing ]
>
>
>
>
> >

[Jan Klincewicz]

Standards can change pretty quickly in this business.  When I think of servers going from ISA to Micro Channel / EISA to PCI-X to PCI-E in the short span they did (to say nothing of the rapid evolution of SCSI, Ethernet, FC etc.  it is pretty clear innovation based on standards will occur (with minor outlyers, like Micro Channel and OS/2 <g>.)

It seems that innovation with consideration to evolutionary change is the norm (Apple is one notable exception.)  It was said that Andy Grove when he leaded Intel insisted that every new iteration of X86 be able to boot DOS 1.0, and I guess that held until they couldn't get any more 5 1/4 inch floppy drives <g>.

--
Cheers,
Jan

[Ben Bloch]

Jim, I am not complaining about lack of standardization or even arguing
about standards per se. (In fact, I never even mentioned standards although
they are part of it - some - like http - even boost innovation/others
admittedly are an obstacle.) And I certainly am not arguing against
innovation - far from it. I am pointing out the risk of cloud service
(particularly PaaS but also IaaS to some degree) lockin and how this needs
to be understood, managed and ideally mitigated. And how this risk can in
fact suppress innovation by suppressing adoption. Sorry for the
misunderstanding.

XaaS adds additional risk that on premise technology does not - the risk of
loss of access to the application.

Cloud Technology (whether Cloud DB's or anything else), on the other hand,
most likely will follow a well trod path, particularly when deployed
on-premise (e.g. on a private cloud) - innovation, market acceptance or
rejection, grandfathering or continued development, standardization (defacto
or otherwise), maturation, possible obsolescence, etc.

The important difference is when an on-premise technology slides to
obsolescence (or is never really accepted by enough of the market in the
first place) or even if the vendor/provider goes belly up, you can still run
the technology and related application(s) for a time - possibly even a long
time. (Just look in any data center and company's application portfolio!)
And if the technology is acquired, it can still be maintained even if
development stops. This is part of the long tail of software maintenance -
and is a well understood often profitable model with fairly predictable
costs. PreSQL 4GL's are a good example.

So what I am arguing for, then, is not to suppress innovation but for models
to handle this XaaS access risk so that innovation can be adopted and
continue to occur. I don't have all the answers - I am speaking primarily as
an application architect (currently) needing to make platform choices - but
I believe the answer - or at least AN answer - will include a combination of
such elements as support/escrow policies, paths to portability (such as
declarative application models), on premise/cloud hybrids, and yes, in some
instances, standards and common accepted languages.

I think also it behooves vendors to be aware of this need/concern (I have
been on both sides.)

Hope that is more clear.

/ben

-----Original Message-----
From: cloud-c...@googlegroups.com
[mailto:cloud-c...@googlegroups.com] On Behalf Of Jim Starkey
Sent: Wednesday, March 04, 2009 8:52 PM
To: cloud-c...@googlegroups.com
Subject: [ Cloud Computing ] Re: Cloud Computing and SOA


You can't have innovation and standardization at the same time. One has
to precede the other. Once a solution or architecture is recognized as
"good enough", market share will generally force convergence, if not
standardization. But not always. Windows, Unix, and Apple never
converged on a drawing package. SQL has been standardized forever, but
not to the degree of interoperability.

Cloud application platforms are in their infancy. Google and Microsoft
are still in the first approximation stage. The database side of the
picture hasn't even started yet.

Complaining about the absence of standardization isn't going to get you
anywhere.


Ben Bloch wrote:
>
> +1 - I agree that providers WANT lock in. That doesn't mean that is
> what customers want or need to accept, Cloud Lockin is even more
> restrictive/risky than a technology stack (LAMP, J2EE, ,Net, etc. )

> lockin since cannot even run the application - let alone port it -

[Jim Starkey]

You pretty much make my case.

Intel innovated without regard to a standard. The 8088/86, 286, 386,
486, and Pentium,et al, were all to their design and specification.
When AMD found market acceptance with the AMD64, Intel had no choice but
to follow.

EISA is an excellent case in pre-mature standardization. It was defined
by an industry consortium, the "gang of nine", as an alternative to
IBM's micro-channel architecture. Like all standards that precede the
technology, it was a horrible compromise. It was extensive to build, a
pain in the but to use (every device required a floppy to configure the
bus), and wasn't nearly fast enough for the aggravation induced. It
didn't solve plug and play (Microsoft and Intel did that), and it's
horrible architecture was never repeated.

Commercial Ethernet was defined by Xerox, Intel, and DEC. It became an
IEEE standard after it demonstrated technical success and market
acceptance. Ethernet should be the post-child for how standards are
set, which is to let the market fight it out and standardize on the
winner. If the winner won't license, pick the runner up (Sun News was
vastly better than X11, but Sun's track record on NFS made them a
non-candidate for a standard).

If you want to slow down technical evolution, declare a standard before
its time.

> > [mailto:cloud-c...@googlegroups.com

> <mailto:cloud-c...@googlegroups.com>] *On Behalf Of *Chris Marino
> > *Sent:* Monday, March 02, 2009 4:34 PM
> > *To:* cloud-c...@googlegroups.com

> <mailto:cloud-c...@googlegroups.com>

> > *Subject:* [ Cloud Computing ] Re: Cloud Computing and SOA
> >
> > +1
> >
> > On Mon, Mar 2, 2009 at 12:28 PM, JL Valente
> <jlva...@cittio.com <mailto:jlva...@cittio.com>

> > <mailto:cloud-c...@googlegroups.com

> <mailto:cloud-c...@googlegroups.com>>
> >
> > Subject: [ Cloud Computing ]
> >
> >
> >
> >
> > >
>
>
>
>
>
>
> --
> Cheers,
> Jan
>
> >

--
Jim Starkey
President, NimbusDB, Inc.
978 526-1376

[Greg Pfister]

On Mar 4, 1:47 pm, Sundeep Singatwaria

<sundeep_singatwa...@hotmail.com> wrote:
> Now that the counter of support for exclusivity with cloud provider has gone up - and for all the members that didn't do +1 - I'm assuming that they latently support this. Now given that I would end up being locked in by a cloud provider - if I move my data center within the provider's cloud - I'm pretty much left at the price / performance of that provider.

This made me think of a perfect analogy: Telephone number portability.

Remember when you couldn't change cell phone or landline providers
without changing your telephone number? You were stuck with the
price / performance of the provider you had or had to risk losing
contact.

The providers never did change that voluntarily. It took regulations
crammed down their throats, moaning and squealing all the way.

Greg Pfister
http://perilsofparallel.blogspot.com/

>For that very pesky reason if I'm a large enterprise connecting with few 100 sites - I would want to build my internal cloud and not rely on an external cloud.
>
>
>
> > Date: Tue, 3 Mar 2009 23:15:26 +0100

> > From: andremer...@gmail.com

> _________________________________________________________________
> Windows Live™ Groups: Create an online spot for your favorite groups to meet.http://windowslive.com/online/groups?ocid=TXT_TAGLM_WL_groups_032009

[Jim Houghton]

In the spirit of +1 permit me to add to Gordon’s spot-on post…

 

I worked with UD (and DataSynapse and Platform and …) when I was at IBM in the early 2000’s doing exactly that – of course nobody called it an Enterprise Cloud then but it sure had many of the characteristics that we generally agree comprise a cloud service (please, please, let’s not have another round of what is/is not a cloud).  I left IBM when I was given the opportunity to go build this as a client – along with a number of my of colleagues working in the same space who were excited to do this end to end.  Over the course of a few years we built a comprehensive internal utility (enterprise cloud) comprised of composable applications (SOA) running on composable infrastructure services (SOI).  Was it proprietary?  Of course it was, and candidly we had to build a bunch of glueware and work closely with a cadre of vendors to fill capability gaps to operate at scale.  Did the enterprise we work for care?  Let’s see … over 40 independent awards for use of technology (helped to continue to attract top talent) … demonstrated over $500M in IT cost savings/avoidance … improved business performance (new capabilities, time to market, etc) that the business guys estimated to be over $2B in new revenues.  This was over 2 years ago, and yes competitors noticed…

 

Our team has since left that enterprise (not all that fun to work in an investment bank these days J) and started our own firm to help clients adopt these technologies leveraging our proven, templated approach.  And yes, there are a lot of people interested in doing this internally and will gladly pay small services/IP licensing fees to get the kind of benefits I mention above.  And it won’t take all that long for the internal guys to figure out how to do this on their own – a successful proof point, a cookbook on how to it, some mentoring through the first few steps and they will be off to the races.

 

Enterprise clouds are the stepping stone to external clouds, at least for any large organization.  This phase will actually help, not hurt external cloud adoption.

 

Jim

_________________________________

Jim Houghton                       CTO and Founder

O:  (704) 654-3303               M: (845) 494-9419

jim.ho...@adaptivity.com   www.adaptivity.com

 

Transforming Enterprise IT Requires ADAPTIVITY^SM

[Sundeep Singatwaria]

 
right on. The business model what you guys have would make perfect sense for folks using internal cloud. Do what you can and for what you don't know make use of the "cloudy consultants". I'm glad that someone on the list came forward to explain their cloudy business model.
 
There is a lot more money to be made serving internal cloud than there ever would be being an external cloud provider. 
 

---

From: jrhou...@yahoo.com

To: cloud-c...@googlegroups.com
Subject: [ Cloud Computing ] Re: Cloud Computing and SOA

Date: Thu, 5 Mar 2009 12:37:15 -0500

[Ben Bloch]

Thread has gone off topic but … +1 for:

-          Doing what it takes to build out cloud (standard/non standard, custom/off-the-shelf glue ….)

-          Private cloud as stepping stone to public cloud – or hybrid model. 

 

 

 

 

---

[Lynne VanArsdale]

"Service-oriented" is all about creating services that meet business needs in a flexible, standard way.  This approach is in contrast to decades of computing that force business to conform to the way we can provide computing. It is about creating service components that can be used in multiple contexts, flexibly to meet the needs of business.  No matter how we implement or what kind of returns we measure, that is what services are all about. Cloud is the logical deployment architecture for service-orientation, as it promises the flexible means to deliver to the business without delay, constraint or prohibiting expense/resources.

On Thu, Feb 26, 2009 at 6:40 AM, <dustin....@gmail.com> wrote:

I'd like to here some ideas on the relationship of cloud computing and
SOA. To me, the relationship is quite interesting.

For one, I think the explosion of services hosted in the upper-most
layer of the cloud (SaaS) can be attributed partly to SOA standards
and principles. To be sure, there are a fair amount of web consumers
of these services (i.e. those using GMail), but there are also

enterprise consumers who utilize these services as part of larger