Prev: Procobol error.. please help
Next: Variable Length Input File
From: P. Raulerson on 22 Dec 2006 19:54
Your guess would be partly right - I support Mainframe (z/OS, z/VM, zLinux),
AIX, Linux, iSeries (OS/400, i5OS, Linux) and several varieties of Windows,
including that blasted 64bit variant of Server 2003 that gives me heartburn.
In fact, our entire enterprise (about 350 people geographically scattered
over 25 sites) runs on Citrix served up from servers here in Austin. Citrix
would be a bit expensive for what this chap wants, but it saves us a ton of
money ,and an incredible number of support hours.

That's a corollary duty by the way, along with phone systems, network
design, security, and second assistant cook and bottle washer too I suppose.
I am best at being your everyday Sr. Software Engineer. That's in COBOL, C,
Java, Ada, Assembler, and assembly language for a few other processors. I
dabble in Visual Basic, RPG, and a half dozen other languages just because
they interest me.

(Yep: I work for good people and I have a great job, when it does not wear
me down to a nub... :)

To be fair though, I did sound arrogant, and that was not intended. I
apologize for that. It is just that I tend to solve these kinds of issues
several times per quarter, and, if this case is what I think it is, using a
centralized VM to deal with this is probably a very good idea.

On the other hand, to me, you sound like a PC person who has little
experience on other platforms. Bet you would argue that running Microsoft
Office on standalone machines is faster than running it under Citrix.
(Hint: it is much faster under Citrix. :) Where are you coming from, and am
I anywhere close to right? :)

-Paul

"Richard" <riplin(a)Azonic.co.nz> wrote in message
news:1166755729.112914.302850(a)i12g2000cwa.googlegroups.com...
>
> P. Raulerson wrote:
>
>> Tsk tsk... not so. There is such a thing as a virtual machine image that
>> allows remote access. See for example:
>> http://www.vmware.com/solutions/desktop/vdi.html. It's even cost
>> effective.
>
> That may be so, but that would be single-user access at any one time
> for each virtual machine. If the application is installed and used
> widely then it is unlikely to be useful to get the users to queue for
> access to it.
>
> I am guessing that you are a mainframer and do not have much experience
> with interactive desktop or PC applications.
>


From: Richard on 22 Dec 2006 20:30

P. Raulerson wrote:

> To be fair though, I did sound arrogant, and that was not intended. I
> apologize for that. It is just that I tend to solve these kinds of issues
> several times per quarter, and, if this case is what I think it is, using a
> centralized VM to deal with this is probably a very good idea.

Given that they probably have Windows PCs everywhere they need them,
then taking the programs off those machines and loading them onto
sufficient new central machines with VMs may be just additional cost.

I think that it would be a very good idea to actually find out what the
application does, how it works, how many use it and where, before
thinking up pre-packaged 'solutions' that may be totally inappropriate.

> On the other hand, to me, you sound like a PC person who has little
> experience on other platforms.

I have done a decade with mainframes and nearly three decades with
departmental (Unix and other) servers, and with networked small
machines, and (overlapping) a decade of web server based systems.

> Bet you would argue that running Microsoft
> Office on standalone machines is faster than running it under Citrix.
> (Hint: it is much faster under Citrix. :) Where are you coming from, and am
> I anywhere close to right? :)

I wouldn't argue anything at all about Windows stuff, I try to avoid it
as much as possible having started developing for Windows 2 and getting
more disenchanted with each new version.

Whether MS Office is faster on a standalone or on a central server a)
depends entirely on the level or resources of each and on such things
as the speed of the connection and utilization and b) is not
partcularly interesting.

From: P. Raulerson on 23 Dec 2006 01:30

"Richard" <riplin(a)Azonic.co.nz> wrote in message
news:1166837413.635751.150290(a)h40g2000cwb.googlegroups.com...
>
> P. Raulerson wrote:
>
>> To be fair though, I did sound arrogant, and that was not intended. I
>> apologize for that. It is just that I tend to solve these kinds of
>> issues
>> several times per quarter, and, if this case is what I think it is, using
>> a
>> centralized VM to deal with this is probably a very good idea.
>
> Given that they probably have Windows PCs everywhere they need them,
> then taking the programs off those machines and loading them onto
> sufficient new central machines with VMs may be just additional cost.
>

I think you are not aware of how incompatible the different versions of
Windows are. Just because there is a 'Windows' machine sitting on someone's
desk, it does not by any strech of the imagination guarantee that an old
program like this is described to be will run.

Running the thing on VM does.


> I think that it would be a very good idea to actually find out what the
> application does, how it works, how many use it and where, before
> thinking up pre-packaged 'solutions' that may be totally inappropriate.
>

Well - you might think that - but then again, you might also realize the
application apparently already works in the environment it is in, and
changing the environment is not a snap your fingers just make a change
thing.

>> On the other hand, to me, you sound like a PC person who has little
>> experience on other platforms.
>
> I have done a decade with mainframes and nearly three decades with
> departmental (Unix and other) servers, and with networked small
> machines, and (overlapping) a decade of web server based systems.
>

Then we are about equal in time, though we obviously have experiences. I
tend to find solutions for clients that are right and work the first time,
and make economic sense. I'm not really wedded to any particular technology.


>> Bet you would argue that running Microsoft
>> Office on standalone machines is faster than running it under Citrix.
>> (Hint: it is much faster under Citrix. :) Where are you coming from, and
>> am
>> I anywhere close to right? :)
>
> I wouldn't argue anything at all about Windows stuff, I try to avoid it
> as much as possible having started developing for Windows 2 and getting
> more disenchanted with each new version.
>

Pre 1990 Windows development (i.e. Windows 3.0 or above) was not a lot of
fun. I used Mac's back then for clients in the publishing industry who
needed graphics. They just worked, and in fact, I have one client who still
runs Mac II's today for some software he is unwilling to pay to upgrade. I
keep a couple of them in the garage just in case one of his breaks. And a
copy of the old AWP loaded on them JIC.

Still - developing on Windows today is nothing at all like that, and it is a
miracle anytime older software will run. Windows XP SP2 especially seemed to
break the world.

> Whether MS Office is faster on a standalone or on a central server a)
> depends entirely on the level or resources of each and on such things
> as the speed of the connection and utilization and b) is not
> partcularly interesting.
>

Well - not to be intentionally insulting, but you are simply wrong here. A
slow DSL connection connected to a Citrix Server will usually outperform a
fast machine loading off a local disk or a network share. The network share,
will almost always be faster. And your statement that it depends entirely on
the relative resources shows that you do not have a sufficient or working
knowledge of the subject.

As to interesting or not - well your whole point seems to be that you think
VM would be a bad choice, and you don't like having your opinion challenged.

As I said, depending on the details, it might not be a viable choice.

But it is probably a good choice and one that will save money, effort, time,
and please the client. Those things that tend to keep
consultants employed and solvent.

That usually is interesting to me.

YMMV, etc. etc. etc.






From: Richard on 23 Dec 2006 16:13

P. Raulerson wrote:

> I think you are not aware of how incompatible the different versions of
> Windows are. Just because there is a 'Windows' machine sitting on someone's
> desk, it does not by any strech of the imagination guarantee that an old
> program like this is described to be will run.

I am quite aware of how new versions of Windows breaks old stuff. Some
may alledge deliberately.

> Well - not to be intentionally insulting, but you are simply wrong here. ...
> ... And your statement that it depends entirely on
> the relative resources shows that you do not have a sufficient or working
> knowledge of the subject.

So, you are saying that Citrix would be faster even if the P128 was the
Citrix server and the quad P3.2G was the desktop ?

> As to interesting or not - well your whole point seems to be that you think
> VM would be a bad choice,

Not at all, what I noticed is that the original said there was no
budget yet you continually want to solve it with your solution
regardless of what the cost would be.

> and you don't like having your opinion challenged.

I am not sure that refers to _me_ in this discussion, after all I have
not put forward any particular opinion other than it would be useful to
find out the scope of the problem before throwing possibly expensive
solutions at it.

> As I said, depending on the details, it might not be a viable choice.

I failed to notice any such qualification, you claimed it as 'best'
without even knowing anything about the nature of the problem.

> But it is probably a good choice and one that will save money, effort, time,
> and please the client. Those things that tend to keep
> consultants employed and solvent.

To a hammer all problems look like nails.

From: gary drummond on 23 Dec 2006 22:45
P. Raulerson wrote:
> Yow- tough situation.
>
> Forgetting all the good reasons to move to something new, you don't have a
> lot of options. I suppose they have reasons not to move the platform right
> away.
>
> The best answer to this situation is simple - do not move it. Just 'rehost'
> it on VM. You can put it on a very fast high quality machine, use the same
> OS that is working now, but have the ability to quickly move it to another
> physical machine or VM. By quickly, we are talking about within five to ten
> minutes.
>
> The benefits? 10 to 20 times the processing power, insulation from hardware
> issues, absolute compatibility with the existing system, safe secure
> backups, rapid deployment for disaster recovery, and on top of that, you can
> (disregarding license issues) easily have a test/development system and a
> production system. The disadvantages - still dependent upon outdated
> unsupported software and non-supported OS. But this is not a problem because
> they are going to replace that software with some new whizbang stuff, right?
>
> This is something you can implement in about a day. Do it, do the changes
> and testing, impress you client, and then engage them for immediate
> follow-on work.
>
> Roughly the cost you are looking at to this:
>
> New PC Hardware: $4K - $11K. Depending upon the needs. $11K = new dual-core
> dual processor IBM LS20 blade with dual fibre connections & eight gigs of
> RAM on board.
>
> VM software (Free to about $4K).
>
> NT Software, Compiler, etc. Already in place.
>
> So cost (in the U.S. you are on your own in the U.K.) is anywhere from $4K
> to $15K, a very reasonable cost to extend the life of a critical system. And
> yes, you could host it on a Windows Server or Windows Vista machine without
> any trouble.
>
> Time to implement: Time for you to make the changes and test them plus one
> to two days.
>
> -Paul
>
>
>
> "UK-Contractor42" <lawrence.foster(a)uk.zurich.com> wrote in message
> news:1166615483.662939.45100(a)48g2000cwx.googlegroups.com...
>> One of my current contract tasks is to put a old, vulnerable
>> application based on 16bit code on a firmer footing by researching,
>> documenting and testing changes to it.
>> Build architecture is Windows NT4 SP4, Microfocus COBOL 3.2.43, Gupta
>> SQL Pre-Compiler 2.5.2, SLQ*Base9.
>> Prod. platform is NT4 SP6, XP SP1. Will be XP SP2 in 2007, and likely
>> to be Vista before the system is decommissioned.
>>
>> Yes I realise this is stone age stuff but there is no budget to
>> re-compile as 32bit COBOL or rewrite using Visual Studio.
>>
>> Have had some success understanding the components that need to be in
>> place to amend, pro-compile, compile and link and built a second
>> working build PC.
>> Will be trawling though the help files next (there are no manuals).
>>
>> It would be great to get in touch with anyone who is using a 16bit
>> version of Microfocus COBOL, or who has used it relatively recently.
>>
>> The Client, an Global Insurer, appreciate the situation is risky -
>> that's why I'm onsite! If anyone knows of any imminent dangers
>> presented by staying with 3.2.43 I would especially like to hear from
>> them.
>>
>
>
You could try running the environment under WINE on Solaris or Linux
too. Bringing up a simple system is not too bad, at least compared to
VM, and if the products work under it, they would no doubt run faster
because the OS would be faster. Running in a zone or container would
allow development and/or testing to be done in a separate
zone/container, even with a different compiler if you wish.

Gary
First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4
Prev: Procobol error.. please help
Next: Variable Length Input File