Unix vs AS/400 - some objective comparison?
spark...@my-dejanews.com
experience and from Soltis' book I know that there are a number of key
architectural features would appear to give the AS/400 superior performance,
scalability etc. I am also interested in 'business-case' evidence, such as
lower administration overhead, lower TCO, superior support organization (this
would be a global deployment). However, I need hard facts, not anecdotes, not
religion, not marketing hype. Any pointers anyone?
Steve
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum
David Pace
I am a unix user who tried to get this kind of info. It is not easy to find.
I never did find precise information.
AS400 fans do not have hard facts. I could not even find benchmark
comparisons of speed. I think that a lot depends on the programming
tools that you select. Forget about the hardware and operating system
differences; I think they are less relevant than the programming tools.
At the time that I made my decision the cost comparison was 10:1 ratio
of AS400 to SUN (6 years ago) and I figured that the $400,000 that I saved
up front would go a long way to pay for long term support differences,
if any. I do not have current cost comparisons.
David Pace
Rocky Marquiss
AIX (IBM's Unix) and multiple AS/400's. I was in charge of making all
the computers to communicate (also ES9000, S/36's, PC's). As in all
things, there is a right tool for the right job. The RS/6000's were
wonderful tools for plant automation. I wouldn't put an AS/400 in
there to save my soul (ok, ok, I might under those circumstances). On
the same note, by the time you installed a database engine on the
RS/6000 that was almost as robust as the AS/400's you spent as much if
not more money in accomplishing that task as just getting the AS/400
to begin with. What I'm saying is, for business applications/database
systems, the AS/400 is a better fit than an RS/6000.
All to often we get caught up in - "this is the best" when in reality
the best depends on the job function. We tend to want to hit in nails
with a pair of pliers rather than a hammer. AS/400 is a fantastic
machine for database and data manipulation activities - none surpass
it as far as I'm concerned. However, Unix type machines are much more
adept at communications - especially at the device level like
scanners, scales, etc..
Dan Hopkins
David Pace wrote:
>I am a unix user who tried to get this kind of info. It is not easy to
find.
>I never did find precise information.
>
>AS400 fans do not have hard facts. I could not even find benchmark
>comparisons of speed. I think that a lot depends on the programming
>tools that you select. Forget about the hardware and operating system
>differences; I think they are less relevant than the programming tools.
>
>At the time that I made my decision the cost comparison was 10:1 ratio
>of AS400 to SUN (6 years ago) and I figured that the $400,000 that I saved
>up front would go a long way to pay for long term support differences,
>if any. I do not have current cost comparisons.
David,
Sounds like perhaps the AS/400 was oversized for its purpose if there was a
10:1 cost differential. Also as Rocky has said, it makes a big difference
as to the intended function(s) and application(s) to be supported.
There are studies available at http://www.as400.ibm.com in the Consultant
Reports section.
Regards,
Dan
Darryl Johns
wrote:
>spark...@my-dejanews.com wrote:
>
>> I need to make a careful, objective comparison of AS/400 vs unix. From my own
>> experience and from Soltis' book I know that there are a number of key
>> architectural features would appear to give the AS/400 superior performance,
>> scalability etc. I am also interested in 'business-case' evidence, such as
>> lower administration overhead, lower TCO, superior support organization (this
>> would be a global deployment). However, I need hard facts, not anecdotes, not
>> religion, not marketing hype. Any pointers anyone?
>>
>> Steve
>>
>> -----== Posted via Deja News, The Leader in Internet Discussion ==-----
>> http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum
>
>
>I am a unix user who tried to get this kind of info. It is not easy to find.
>I never did find precise information.
>
>AS400 fans do not have hard facts. I could not even find benchmark
>comparisons of speed. I think that a lot depends on the programming
>tools that you select. Forget about the hardware and operating system
>differences; I think they are less relevant than the programming tools.
>
>At the time that I made my decision the cost comparison was 10:1 ratio
>of AS400 to SUN (6 years ago) and I figured that the $400,000 that I saved
>up front would go a long way to pay for long term support differences,
>if any. I do not have current cost comparisons.
>
>David Pace
>
>
David,
What a difference 6 years makes. There are numerous consultants
reports comparing these environments. IDC, for instance has rated the
AS/400 to have the least Total Cost of Ownership for the last 4 or 5
years in a row. the AS/400 was compared to UNIX (HP w/Oracle),
NetWare, and NT.
The Client Server Labs site; often sited by, and employed by
ComputerWorld has a number of Client/Server benchmarks and rates
various platforms. The AS/400 is number one in the RPMark test...and
has been for over a year. We consistently beat the UNIX variants here.
( see: www.clinc.com)
Did you see Computerworld's database survey from last year? DB2/400
beat out Oracle, Sybase, SQL Server, and Informix in the majority of
categories surveyed.
(see:
http://www.computerworld.com/home/online9697.nsf/all/970519dbms2
http://www.computerworld.com/home/online9697.nsf/all/970421guide2
With the announcements of last week, the AS/400 is the most scaleable
2-tiered SAP R/3 platform...beating ALL other vendors and our
Mainframe brethren.
With Domino for the AS/400 we scale to 10,400 audited Notesbench users
on a single machine. The #2 platform was SUN at 6,400. That was in
March. And as of last week we will probably scale to over 20,000.
There ARE objective comparisons to UNIX in the marketplace.
Regards,
Darryl Johns
IBM AS/400 e-business Specialist
(opinions expressed are mine, not necessarily those of IBM)
Frank Healy
There are some Consultant reports listed at the AS/400 site that might
be useful. The site is : http://www.as400.ibm.com/ .
Frank
danh...@millcomm.com
>I need to make a careful, objective comparison of AS/400 vs unix. From my own
>experience and from Soltis' book I know that there are a number of key
>architectural features would appear to give the AS/400 superior performance,
>scalability etc. I am also interested in 'business-case' evidence, such as
>lower administration overhead, lower TCO, superior support organization (this
>would be a global deployment). However, I need hard facts, not anecdotes, not
>religion, not marketing hype. Any pointers anyone?
There are a couple of the big consulting companies that study this sort of
stuff and create reports you can buy. I can't think of any names right
now, though.
Dan Hicks
Hey!! My advice is free -- take it for what it's worth!
http://www.millcomm.com/~danhicks
David Pace
> On Wed, 09 Sep 1998 14:15:12 -0400, David Pace <dp...@echo-on.net>
> wrote:
>
> >spark...@my-dejanews.com wrote:
> >
> >> I need to make a careful, objective comparison of AS/400 vs unix. From my own
> >> experience and from Soltis' book I know that there are a number of key
> >> architectural features would appear to give the AS/400 superior performance,
> >> scalability etc. I am also interested in 'business-case' evidence, such as
> >> lower administration overhead, lower TCO, superior support organization (this
> >> would be a global deployment). However, I need hard facts, not anecdotes, not
> >> religion, not marketing hype. Any pointers anyone?
> >>
> >> Steve
> >>
> >> -----== Posted via Deja News, The Leader in Internet Discussion ==-----
> >> http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum
> >
> >
Darryl:
I read these surveys (just now) and I am challenging their validity.
I have been reading these kinds of surveys that magazines present
for many years and I always wondered how they are conducted.
Here is how I think they are conducted:
They ask a bunch of users to answer a bunch of simple
questions that attach a numeric value to a low to high set of
subjective levels of satisfaction or perceived cost and then they
tally the results. The users tend to be using one or another
system ( not all systems ).
So, these surveys have a very big flaw. They assume that on
average (if the sample size is large enough) that a rating of
2,3 or 4 from a UNIX person means the same thing as 2,3, or 4
from an AS400 person. Why is this assumption flawed ?
Well, because there is a certain kind of personality difference
in a person who would want the security of big blue versus
the risk taker who would embrace something new (and cheap)
like WindowsNT on a PC. So, by the very nature of this
difference in personality, a WindowsNT person would say
that WindowsNT is EXPENSIVE (because he is always looking
for cheap-cheap-cheap). An AS400 person would say
that AS400 is total-cost-of-ownership-INEXPENSIVE
because he feels SAFE and secure with big blue and he has a
long history of being told that total cost of ownership is cheap
with the AS400 line (I always hear this from AS400 people
even though they have never used UNIX and they know
nothing about it. It is like communist party propaganda. Some
people believe it because it is said often enough).
So, you see, the numbers for AS400 vs WindowsNT vs
UNIX will be biased depending on the personality and marketing
hype and culture that each user grows up with. WindowsNT
could get an EXPENSIVE rating only because cheapskates buy
it (no offense, I am exaggerating to make a point).
In fact, the only valid survey might be of users that have
experience with ALL systems compared in the survey.
Of course, no such persons exist. So, no such survey
could be done. In fact, even such a survey would be
converting subjective answers to numeric values.
How many people really have hard-core facts in front
of them when they answer surveys? We just skip though
and answer these surveys from memory and spend five
minutes at most.
And even more important: Have you
ever found yourself answering a survey with a value that
you want to be true rather than IS true. For example:
Say I just invested a ton of money in a bunch
of WindowsNT computers and they crash once a week
(because of an operating system bug). Oh, well,
that is not so bad. I don't want to admit to myself that
I might have made a mistake. Maybe it is not the
WindowsNT. It must be my application. (It is called
cognitive dissonance). I will give WindowsNT
a high 5 rating for reliability. I don't want AS400 people
to win the debate in this survey.
You see, the survey becomes more of a horse race
where we want our horse to win so we skew the results
in our favour.
So, I have three reasons to not believe these surveys:
1) personality/cultural differences between the different
user groups.
2) surveys pull subjective answers from us (not objective
facts).
3) the horse race phenomenon.
By the way, I mean no offense to AS400 users. I am
a unix guy who is about to embrass AS400 to help
a client. (The server/connectivity strategy that IBM
used is helping their product line).
But, as with all people, I find that what I learned
first is, of course, the easiest (for me).
These surveys are completely useless.
David Pace
Charles M. Wilt
resist.
>
> Here is how I think they are conducted:
>
I think that is your most important point...Since you don't know, or have a
degree in statistical sampling, the rest is just BS.
I don't read a survey results with the idea results are written in stone,
but I do think they can be a source of good information. BTW, the rest of
the world apparently thinks so to.
--
Charles Wilt
Miami Luken, Inc.
Springboro, OH. 45066
e-mail: charle...@worldnet.no.spam.att.net
--remove the .no.spam
Dan Hopkins
Interesting that you perceive that the results of these "surveys" come from
users marking a survey form (like JD Powers car dealer satisfaction
surveys). To the contrary, these are real-life comparisons in processing
like transaction loads on each platform and seeing which one can handle the
most transactions per hour and the most users to get the scalability numbers
Darryl cited. If you read his post carefully you'll see he used the word
"benchmark" instead of "survey".
The cost of ownership figures are also drawn from real-life budget numbers
for companies running the various platforms.
Go to the links that Darryl has provided and see for yourself.
Dan Hopkins
President - Hopkins Development, Inc.
Bob
>
>
> I read these surveys (just now) and I am challenging their validity.
>
> I have been reading these kinds of surveys that magazines present
> for many years and I always wondered how they are conducted.
>
> Here is how I think they are conducted:
>
David,
Excellent thought process on the subject. Often, we're so convinced
that it's hard to be objective.
But, the IDC studies have not been "how do you feel about your computer
today?" phone surveys. They used actual IT budget expenditures to reach
their conclusions.
________________ _______________
\_____________ Robert Bittner <bit...@vnet.ibm.com> _____________/
\__________ IBM Rochester Server Group-AS/400 __________/
\________ AS/400 Partners In Development ________/
G Moore
>I need to make a careful, objective comparison of AS/400 vs unix. From my own
>experience and from Soltis' book I know that there are a number of key
>architectural features would appear to give the AS/400 superior performance,
>scalability etc.
hmm. i would not worry so much which system you use, as which
programming language, and how portable it is. scalability is a
software specific issue. performance is a software location issue.
when i refer to location, i mean how many chips it has to go through
to get the information.
gvwm...@ix.spam.netcom.com to reply remove the spam
David Abramowitz
consists of.
For example:
OS/400 has a built in RDBMS, with UNIX you have to buy an extra
product.
OS/400 has built in object level security, with UNIX you have to
program extra security routines.
OS/400 has built in help screens which list all values of all
parameters, and guides you through entering the command. UNIX has
"MAN" pages which has a buglist for each command.
Under OS/400 you can check on the status of every job in the
system. Under UNIX you would have to purchase an additional
package to accomplish this.
Under UNIX you can use any character for a file name. You are
limited to alphanumeric characters with OS/400. Of course OS/400
will never confuse a filename with an arguement -- UNIX will do
this.
You can add more to these examples.
--
David Abramowitz
Paul Nicolay
On UNIX you have to do and buy everything yourself (keep this is mind for
price comparisons). This is a strategy that has plusses and minusses,
however in a business evironment, this causes a lot of work and maintenance
of issues that don't have to do anything at all with actually running your
business... hence this is lost effort and money.
The freedom of all these tools that you need to buy has another
disadvantage. If I write a tool for AS/400... I know it will run on every
AS/400, as they all use the same database, the same OS flavor, the same
features, ... On Unix on the other hand, you don't have any garantuees at
all.
Anyway, this is a nice one
http://www.accountmate.com/products/as400_misconceptions.htm
Kind regards,
Paul
____________________
David Abramowitz <10544...@CompuServe.COM> wrote in article
<OfIY56S...@ntdwwaaw.compuserve.com>...
The contents of this message express only the sender's opinion.
This message does not necessarily reflect the policy or views of
my employer, Merck & Co., Inc. All responsibility for the statements
made in this Usenet posting resides solely and completely with the
sender.
Kenneth Tan
>On UNIX you have to do and buy everything yourself (keep this is mind for
>price comparisons). This is a strategy that has plusses and minusses,
>however in a business evironment, this causes a lot of work and maintenance
>of issues that don't have to do anything at all with actually running your
>business... hence this is lost effort and money.
Here, I'd say that the view is biased.
There is more "you have to do and buy everything" in AS/400 than there is
in (generic) Unix. There has been more unix support than there is for AS400,
and you'll find much more (useful) utilities and freebies (e.g., GNU)
in unix than even DOS/Windows, let alone AS/400. There aren't unfortunately
that many AS400 freeware/shareware on the 'net, not to mention that the
AS/400 has been a rather 'new entry' to the internet. However, really how
"useful" or applicable they are to your business is another matter.
Depending on how well one manages to setup and maintain the system, both
the unix and as400 are just as capable of 'self-maintainence' without
administrator intervention (in a similarly stable environment for fair
comparison)
>The freedom of all these tools that you need to buy has another
>disadvantage. If I write a tool for AS/400... I know it will run on every
>AS/400, as they all use the same database, the same OS flavor, the same
>features, ... On Unix on the other hand, you don't have any garantuees at
>all.
Again, it depends on how one writes the program.
If you write a script for AS/400, you know it will run on (maybe) every
AS/400, but it will ONLY run on AS/400. Not much help in a truly
hetrogenous environment.
In Unix, a program/scrip can be made to run on all platforms, be it Sun,
HP, DEC, etc., it doesn't have to be an IBM.
I've written lots of programs & scripts like that. And anyone who programs
with the strict discipline of keeping them to generic/ANSI standard C, will
achieve the highest portability, even to the DOS environment (I hate GUI/
windoz, so nifty little programs are so much more portable).
Ofcourse, for the purist, go for BASIC/JAVA!
Automan
Michael Mundy
Kenneth Tan wrote:
> There has been more unix support than there is for AS400,
> and you'll find much more (useful) utilities and freebies (e.g., GNU)
> in unix than even DOS/Windows, let alone AS/400.
There is no reason why the GNU utilities cannot be ported to AS/400. It just
hasn't been done yet.
> Again, it depends on how one writes the program.
> If you write a script for AS/400, you know it will run on (maybe) every
> AS/400, but it will ONLY run on AS/400. Not much help in a truly
> hetrogenous environment.
> In Unix, a program/scrip can be made to run on all platforms, be it Sun,
> HP, DEC, etc., it doesn't have to be an IBM.
AS/400 does have a shell interpreter and Perl so you can write scripts that are
portable to other platforms.
--
Michael Mundy
AS/400 Open Systems Enablement
mmu...@vnet.ibm.com
These opinions are mine and not necessarily those of IBM.
Alex
David Abramowitz wrote:
> One form of comparison might be what the operating system
> consists of.
>
> For example:
>
> OS/400 has a built in RDBMS, with UNIX you have to buy an extra
> product.
On UNIX you can also buy another RDBMS if you don't like the vendor,its
product or support.
>
>
> OS/400 has built in object level security, with UNIX you have to
> program extra security routines.
> OS/400 has built in help screens which list all values of all
> parameters, and guides you through entering the command. UNIX has
> "MAN" pages which has a buglist for each command.
On UNIX they where that honest to admit their errors...
>
>
> Under OS/400 you can check on the status of every job in the
> system. Under UNIX you would have to purchase an additional
> package to accomplish this.
On UNIX you type ps -l | more
>
>
> Under UNIX you can use any character for a file name. You are
> limited to alphanumeric characters with OS/400. Of course OS/400
> will never confuse a filename with an arguement -- UNIX will do
>
If you can't appreciate the flexibility & openness of UNIX; yes OS/400
isa better one in your case. PS: rm -rf ./-mystupidmistake removes files
with hyphens; the information is above their errors in the man page ;-).
Nicholas Bridge
Unless the parameter is defined as *SNAME (simple name), you can
specify quoted names:
"bugbear" (lowercase letters are preserved)
"(new)"
and
"!@#$%^&"
are all valid object names.
Also, you can specify quoted generics (ie for WRKOBJ etc):
"(n*"
"!@*"
Are valid generics (yes the last " is required and comes after the
genercic asterisk)
Restrictions:
No ' * or spaces allowed (except in generic, where a * before the " is
allowed).
Not very useful, but there it is!
Regards,
Nick B.
David Abramowitz wrote:
>
> One form of comparison might be what the operating system
> consists of.
>
> For example:
>
> OS/400 has a built in RDBMS, with UNIX you have to buy an extra
> product.
>
> OS/400 has built in object level security, with UNIX you have to
> program extra security routines.
>
> OS/400 has built in help screens which list all values of all
> parameters, and guides you through entering the command. UNIX has
> "MAN" pages which has a buglist for each command.
>
> Under OS/400 you can check on the status of every job in the
> system. Under UNIX you would have to purchase an additional
> package to accomplish this.
>
> Under UNIX you can use any character for a file name. You are
> limited to alphanumeric characters with OS/400. Of course OS/400
> will never confuse a filename with an arguement -- UNIX will do
G Moore
On Wed, 09 Sep 1998 23:44:22 GMT, djo...@us.ibm.com (Darryl Johns)
wrote:
>The Client Server Labs site; often sited by, and employed by
>ComputerWorld has a number of Client/Server benchmarks and rates
>various platforms. The AS/400 is number one in the RPMark test...and
>has been for over a year. We consistently beat the UNIX variants here.
hmm. whats the rpm test do?
>Did you see Computerworld's database survey from last year? DB2/400
>beat out Oracle, Sybase, SQL Server, and Informix in the majority of
>categories surveyed.
hmm. so computerworld tests beat database servers. big deal. the
problem is, the db2 manual has 11 megs of documentation
(specification).
the question to ask is:
given a dynamic database command (sql statement or other type),
with indexed or non-indexed,
3 types of cache (no, index is cached, index+somedata is cached),
defragmented/ non-defragmented
the same amount of money to
allocated to hard drive seektime, turntime ratings,
which gives the fastest read rating?
this is the standard OLTP transaction most databases cannot schedule
and so would also be the most effective method of determining to use.
>With the announcements of last week, the AS/400 is the most scaleable
>2-tiered SAP R/3 platform...beating ALL other vendors and our
>Mainframe brethren.
hmm. i can't disagree since i don't know what r/3 means, but i do know
what symetric asymetric processing, and 2 tiers are.
>With Domino for the AS/400 we scale to 10,400 audited Notesbench users
>on a single machine. The #2 platform was SUN at 6,400. That was in
>March. And as of last week we will probably scale to over 20,000.
yes, but how many corporations would scale to 10,400 or even 6,400
licenses of lotus notes?
if i were a company, and used notes, i would simply connect 2 suns,
or multiple lower level machines along company division lines.
the one thing about notes, though, is that it has some text/database
features, so that level of scalability might be useful, if something
at the top had to reach every1.
but i am sorry, that just ain't the case. ceos rarely have any
*specific* advice which directly affects each individual manager or
employee. they may have a message of
the day or some vision statement (increase sales 10%), but thats
not really helpful.
>There ARE objective comparisons to UNIX in the marketplace.
yes, its just that magazines won't publish those comparison test
methods, for fear of customers.
the real catch about micros is: they have 30 years of custom software
programs. that isn't going to be easy to port. as400 will be around
for a long time. its even less easy to convince MIS depts to allow
you to rewrite the code, much less pay people for it.
G Moore
wrote:
>If you write a script for AS/400, you know it will run on (maybe) every
>AS/400, but it will ONLY run on AS/400. Not much help in a truly
>hetrogenous environment.
>In Unix, a program/scrip can be made to run on all platforms, be it Sun,
>HP, DEC, etc., it doesn't have to be an IBM.
hmm. porting across machines is a bitch, whether u go unix or as400.
again, a software specific issue. writing portable code is possible
in JAVA or cobol. the catch is, unless your system is totally
seperated from the world, u r going to setup applications, and
find and install devices and applications.
as for unix or dos having more utilities, i would agree, BUT...
most of the utilities do the same thing. its too bad, too, cause
if half of those shareware writers were paid by big-companies,
Daniel Côté
>
> Paul Nicolay wrote:
> >On UNIX you have to do and buy everything yourself (keep this is mind for
> >price comparisons). This is a strategy that has plusses and minusses,
> >however in a business evironment, this causes a lot of work and maintenance
> >of issues that don't have to do anything at all with actually running your
> >business... hence this is lost effort and money.
>
> Here, I'd say that the view is biased.
> There is more "you have to do and buy everything" in AS/400 than there is
> and you'll find much more (useful) utilities and freebies (e.g., GNU)
> that many AS400 freeware/shareware on the 'net, not to mention that the
> AS/400 has been a rather 'new entry' to the internet. However, really how
> "useful" or applicable they are to your business is another matter.
>
> Depending on how well one manages to setup and maintain the system, both
> the unix and as400 are just as capable of 'self-maintainence' without
> administrator intervention (in a similarly stable environment for fair
> comparison)
>
> >The freedom of all these tools that you need to buy has another
> >disadvantage. If I write a tool for AS/400... I know it will run on every
> >AS/400, as they all use the same database, the same OS flavor, the same
> >features, ... On Unix on the other hand, you don't have any garantuees at
> >all.
>
> AS/400, but it will ONLY run on AS/400. Not much help in a truly
> hetrogenous environment.
> In Unix, a program/scrip can be made to run on all platforms, be it Sun,
> HP, DEC, etc., it doesn't have to be an IBM.
> with the strict discipline of keeping them to generic/ANSI standard C, will
> achieve the highest portability, even to the DOS environment (I hate GUI/
> windoz, so nifty little programs are so much more portable).
> Ofcourse, for the purist, go for BASIC/JAVA!
>
> Automan
Dear Automan, I guess by your writing that you still have to learn some
stuff
about the AS/400. Paul is right when he mention that anything written on
one
AS/400 is portable across the whole park. The only hick are things like
fitting
the programs in a menuing or in a customized environment (JOBD, SBSD,
etc.). This
situation in my opinion is rare since most shops (really hope so) uses
default
settings in their machine.
On the UNIX side, I agree that if you take all precautions to write a
program, set
aside the script, it might be portable. You will be able to compile as
long you
don't use includes that pertain to the technology you a working on. But
even if
you compile, will it run ? How will you fit it in every possible
arrangement a UNIX
box can have as far as security, database, menuing, user interface
(Green screen,
XWindows, C/S, etc.) ?
I don't think you can argue to the fact that within all installed AS/400
on the globe,
the environment are much more homogenous than all HP9000, or DEC ULTRIX,
or RS6000 AIX,
or PCs SCO, or LINUX, or ??? loaded with thousand of utilities and work
management tools
and secutiry and database and... I can see this in our own company where
we have a dozen
of AS/400 and about 30 UNIX boxes that can be supported mainly by the
supplier of the
application we run on them. How can we support 8 database managers, 20
different security
and menuing utilities, performance monitors, network config, Xwindows,
green scree and
Cient/Server things in a cost effective way ? We are bound to the
suppliers... And I don't
see how I can write an application on all of those machines without
spending a great deal
of work to see if it can work, fit etc. before ever writing a line of
code that would have
to interface to all of these.
See what I'm sayin ?
Daniel.
P.S.- Another thig the AS/400 allows: I bet that I can fix any AS/400
around the world
with my present knowledge (As long as I can get to an english
console). Can you
do the same for a UNIX box ? If you can, you are much more than a
genius!
David Pace
I read the links and I did not see details about how the survey is conducted.
The survey was more than just a single benchmark (although Darryl's point was
maybe just regarding the benchmark test). Bob says that these surveys are from
actual IT budget figures. That is better than what I though. I stand
corrected.
But, I can't imagine how one IT budget could be compared to another.
A lot depends on how the results are assembled. Does everyone just trust
that IDC does a good job?
David Pace
PS: (I know that Charles will use my line that says 'I can't imagine' to insult
me
again. I put it in so that he can amuse himself).
David Pace
> David,
>
> Interesting that you perceive that the results of these "surveys" come from
> users marking a survey form (like JD Powers car dealer satisfaction
> surveys). To the contrary, these are real-life comparisons in processing
> like transaction loads on each platform and seeing which one can handle the
> most transactions per hour and the most users to get the scalability numbers
> Darryl cited. If you read his post carefully you'll see he used the word
> "benchmark" instead of "survey".
>
> The cost of ownership figures are also drawn from real-life budget numbers
> for companies running the various platforms.
>
> Go to the links that Darryl has provided and see for yourself.
>
> Dan Hopkins
> President - Hopkins Development, Inc.
Dan:
I checked the original Darryl message and you are not correct.
The computer world article was not a benchmark. It was a survey.
You were polite about it. Charles M. Wilt is nasty and ignorant
of the point of my message. He said that I wrote BS.
Here is my point again:
I was referring to Darryl's mail message which
I fetched from dejanews just now:
Darryl wrote this:
<top snipped>
Did you see Computerworld's database survey from last year? DB2/400
beat out Oracle, Sybase, SQL Server, and Informix in the majority of
categories surveyed.
(see:
http://www.computerworld.com/home/online9697.nsf/all/970519dbms2
http://www.computerworld.com/home/online9697.nsf/all/970421guide2
With the announcements of last week, the AS/400 is the most scaleable
2-tiered SAP R/3 platform...beating ALL other vendors and our
Mainframe brethren.
With Domino for the AS/400 we scale to 10,400 audited Notesbench users
on a single machine. The #2 platform was SUN at 6,400. That was in
March. And as of last week we will probably scale to over 20,000.
There ARE objective comparisons to UNIX in the marketplace.
Regards,
Darryl Johns
IBM AS/400 e-business Specialist
(opinions expressed are mine, not necessarily those of IBM)
-------------------------------------------
So, I wrote a critique of the www.computerworld.com
surveys (not benchmarks!).
Here is a clip from one of the surveys:
Computerworld surveyed 1,205 distributed database users to measure
satisfaction levels with their database and their DBMS vendor. The
results, drawn from a quarterly customer satisfaction survey [CW, April
21], focus on the top five DBMS vendors: IBM, Informix Software, Inc.,
Microsoft Corp., Oracle Corp. and Sybase, Inc. The mail survey asked
respondents to grade their vendor in areas such as software quality,
service and support, reliability, comfort with vendor and overall
satisfaction.
-------------------------------------------
Once again, I say that this kind of survey is useless
for the detailed reasons that I listed in a previous message.
The work that IDC and Gartner Group do, might be useful.
And Darryl's mention of a benchmark from http://www.clinc.com
is a valid point. I am challenging the construction of most other
statistical surveys.
I await an apology from Charles.
David Pace
David Pace
I wasn't going to bother to post a response to this, but I just couldn't
resist.
I think that is your most important point...Since you don't know, or have a
degree in statistical sampling, the rest is just BS.
I do have a degree in statistical sampling.I guess you did not read the specific survey that this thread discusses.
So, take back the BS insult.
-- David Pace
Charles M. Wilt
If you want an apology from me, you'll need to have some proof that the
surveys you are talking about are indeed useless. Rather than your own
uneducated, unless you have some degree in statistical sampling, opinion.
So David, enlighten me and the rest of the world who have the mistaken
impression that these types of survey do indeed have some type of useful
information.
BTW, would you find a survey that showed UNIX at the top of the pile as
useless as you've found the ones that show the AS/400 at the top?
--
Charles Wilt
Miami Luken, Inc.
Springboro, OH. 45066
e-mail: charle...@worldnet.no.spam.att.net
--remove the .no.spam
David Pace <dp...@echo-on.net> wrote in article
<35FDAAE6...@echo-on.net>...