On 10/24/2022 9:10 PM, Dave Froble wrote:
> On 10/24/2022 7:30 PM, Arne Vajhøj wrote:
>> On 10/24/2022 6:56 PM, Bill Gunshannon wrote:
>>> On 10/24/22 16:42, Arne Vajhøj wrote:
>>>> It may make sense to pay for a custom solution tailored for
>>>> specific needs.
>>>
>>> Sites like the University already had this.  Developed over many
>>> years and lots of man hours.
>>>
>>>> It may make sense to go for a standard solution and use it
>>>> the standard way.
>>>
>>> Why would that make sense if you already had (and had paid for)
>>> the working solution your going to replace?
>>
>> The model where you pay X to develop an application
>> and you use that for decades for zero annual cost
>> is not a common model.
>
> I'd agree with that.  Things change.  The apps must reflect the changes.
>
>> A way more common model is to pay X to develop an application
>> and then every year pay Y to maintain and enhance the application
>> to meet new requirements.

>>>> That sort of defies the point of the standard solution.
>>>
>>> The big question is why even consider a "standard solution" if you
>>> already had a perfectly functioning system specific to your
environment?
>>
>> A perfectly functioning system for yesterdays requirements
>> not for tomorrows.
>
> NO, a perfectly functioning system for the past, today, and easily
> modified for tomorrow, if that becomes necessary. Nothing stays the
> same, one advances or regresses.

Yes.

Maintenance stops the day the system is shut down for good.

>> With Y being 10-30% of X. Over the life cycle of the application
>> the maintenance cost are way higher than the initial development.
>
> Now which side of your ass did you pull those numbers?  10-30% is
> ridiculous.

Pick your choice from:

1) Picking a random number 1..50, add current wind speed and deduct
the date.

2) Barry Boehm in 1983 created a software maintenance model
and the example he used was 700 person month development and
140 person month annual maintenance. 20% widened to 10-30%
to cover variations seems reasonable.

3) The most estimated metric is percentage of software total
life cycle cost attributed to maintenance. Numbers vary quite a
bit but 50%, 67%, 75%, 80% and 90% are common. If we assume an
average lifespan of 20 years we get 5%, 10%, 15%, 20% and 45%.
Which one could narrow to 10-30% by assuming that the most
extreme values are rare.

>> Reducing that maintenance cost can make sense.
>
> In any reasonable situation, that never makes any sense.  Any total
> replacement will cost way more than anything else, and probably cause
> all kinds of business problems.

Businesses are always interested in reducing cost.

Not all initiatives promising to do so actually deliver
on the promise.

Arne

On 10/25/2022 6:55 PM, Arne Vajhøj wrote:
> On 10/25/2022 10:48 AM, Dave Froble wrote:
>> My message was, sometimes being restricted to what's in a "standard" app can
>> be a real big problem for users.
>>
>> Software should fit the business, not expecting the business to fit the software.
>
> In the end it is about money.
>
> if increase_in_revenue + reduction_in_cost > cost_customization then
> customize
> else
> do_standard
> end if

One thing you can count on. The vendors of software packages tend to attempt to
make them as good and versatile as possible. (Then there is SAP.) So many
times a well designed system will be better than what it's replacing. Of
course, some existing apps may be doing a great job, better than a replacement.
There is no "one size fits all".

A good replacement should never make things harder.

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

VMS has been kind to me, it's allowed me to move across lots of different industries over the years

Sadly, VMS has exited most of those places I have worked or has been relegated to some corner on it's own, deemed too difficult to replace...yet

Some of them were/are...

- Pathology (integrated system, tracking customers, lab equipment interfacing, financials etc)
Replaced due to merger/acquisition

- Municipality (Billing, rates, equipment tracking etc - they ran their own gas and electricity divisions until government forced their exit from those industries)
Replaced by cheaper more flexible and up to date standardized offerings

- Government (Central healthcare across the board spanning a number of hospitals (lab interfacing/control, billing, patient care/hospital administration, financials, doctor/nurse registrations etc)
Replaced by cheaper more flexible and up to date standardized offerings

- Education (Central system maintaining student information, financials, course registration and course scheduling)
Replaced by cheaper more flexible and up to date standardized offerings

- Logistics (Logistics, intermediary warehousing, financials, billing etc)
Not replaced but in time will be.
All the highly specialized functions around least path analysis, bulk discounting, special order processing (shipping military equipment, from plane parts to rocket launchers) will be scrapped and new more simplified offering set up in a new system. Companies often don't even bother porting functionality over anymore. Simplification is cheaper to manage than keeping extensive offerings going

- Telco (a few of them). (Billing, rating, order management, switch interfacing etc)
Most have gone, Replaced by cheaper more flexible and up to date standardized offerings.
One still remains but it only handles internal customers (external branded companies).
Despite it running a web server and having a more direct customer interface, I heard through the grapevine it's going to get to axed in the next few years
Telco used to be a VMS stronghold :-(

- Financial (banking). They have VMS systems that handle specific products. The systems are in a specialized group outside of standardized offerings.
It's very difficult to replace a system that handle 30 year old mortgages for example due to regulations protecting those customers. In Telco it was easy, you just offered them some sweet deal like a ton of free calls and customers would happily move to a completely new product and you could replace the supporting VMS system, not so easy in banking however.
I don't know the numbers but I suspect the customers left are not many because there is chatter about the VMS systems being retired in the 'next few years'

So many VMS systems get shown the door because they lack integration with the rest of the organisation. Things like Applications dynamics, Splunk (yes, there's VMSSPI), CyberARK, etc.
Then there's things like MongoDB and other NoSQL flavored stuff that people/products want to use. Yes, they might all be one off's in their own right but have a few of these missing on your platform and suddenly your seen as difficult to work with/non supportive of 'mainstream' technologies.

Risk drives everything and replacing VMS is seen as risk mitigation. I'm on a mainframe project that is trying to replace an application that runs perfectly fine and has done so for many many many decades but it effectively an in-house DBMS and the skills (not programming, but in the DBMS itself) are extremely difficult to find now, so replacing the system wholesale is the only option left because no young person is interested in training up in such a one-off system.

In terms of current deployments for VMS, I have not heard of VMS gaining any new business other than upgrades to an existing operation but I don't frequent VMS circles anymore

On 10/29/2022 2:08 PM, IanD wrote:
> Replaced by cheaper more flexible and up to date standardized offerings

> Replaced by cheaper more flexible and up to date standardized offerings

> Replaced by cheaper more flexible and up to date standardized offerings

That is seen.

:-(

> So many VMS systems get shown the door because they lack integration
> with the rest of the organisation. Things like Applications dynamics,
> Splunk (yes, there's VMSSPI), CyberARK, etc. Then there's things
> like MongoDB and other NoSQL flavored stuff that people/products want
> to use. Yes, they might all be one off's in their own right but have
> a few of these missing on your platform and suddenly your seen as
> difficult to work with/non supportive of 'mainstream' technologies.

Available software is essential for an OS.

But I suspect database will look better soon:

RDBMS - today we have SQLite and old version of MySQL/MariaDB, but if
things work out then we will soon have SQLite, new MySQL/MariaDB and
PostgreSQL

NoSQL KVS - VMS index-sequential files are OK in most cases

NoSQL DS (MongoDB market) - newer versions of MySQL/MariaDB and
PostgreSQL can fill this role via their JSON column type so we kjust
have to wait

NoSQL CS (HBase/Cassandra market) - nobody will want to use VMS for this

> In terms of current deployments for VMS, I have not heard of VMS
> gaining any new business other than upgrades to an existing operation
> but I don't frequent VMS circles anymore

There has probably not been many new VMS customers the last 10 years.

And a steady attrition of user base.

But still I think VMS future looks better now than in the past.

* there is a company dedicated to VMS (VSI)
* VMS runs on on standard hardware (x86-64) with a future
* VMS can run in common enterprise environments like
VMWare ESXI and AWS/Azure/GCP

VMS now needs software!

Arne

On Sunday, October 30, 2022 at 3:05:53 PM UTC-5, Arne Vajhøj wrote:
> On 10/29/2022 2:08 PM, IanD wrote:
> > Replaced by cheaper more flexible and up to date standardized offerings
>
> > Replaced by cheaper more flexible and up to date standardized offerings
>
> > Replaced by cheaper more flexible and up to date standardized offerings
> That is seen.
>
> :-(
> > So many VMS systems get shown the door because they lack integration
> > with the rest of the organisation. Things like Applications dynamics,
> > Splunk (yes, there's VMSSPI), CyberARK, etc. Then there's things
> > like MongoDB and other NoSQL flavored stuff that people/products want
> > to use. Yes, they might all be one off's in their own right but have
> > a few of these missing on your platform and suddenly your seen as
> > difficult to work with/non supportive of 'mainstream' technologies.
> Available software is essential for an OS.
>
> But I suspect database will look better soon:
>
> RDBMS - today we have SQLite and old version of MySQL/MariaDB, but if
> things work out then we will soon have SQLite, new MySQL/MariaDB and
> PostgreSQL
>
> NoSQL KVS - VMS index-sequential files are OK in most cases
>
> NoSQL DS (MongoDB market) - newer versions of MySQL/MariaDB and
> PostgreSQL can fill this role via their JSON column type so we kjust
> have to wait
>
> NoSQL CS (HBase/Cassandra market) - nobody will want to use VMS for this
> > In terms of current deployments for VMS, I have not heard of VMS
> > gaining any new business other than upgrades to an existing operation
> > but I don't frequent VMS circles anymore
> There has probably not been many new VMS customers the last 10 years.
>
> And a steady attrition of user base.
>
> But still I think VMS future looks better now than in the past.
>
> * there is a company dedicated to VMS (VSI)
> * VMS runs on on standard hardware (x86-64) with a future
> * VMS can run in common enterprise environments like
> VMWare ESXI and AWS/Azure/GCP
>
> VMS now needs software!
>
> Arne

One thing that stands out about OpenVMS is its record-oriented filesystem (Files-11), along with its extensive batch facilities. The Linux and Microsoft Windows environments don't really seem to have those kinds of tools as part of their baseline. It seems like at least some organizations where there is a significant culture separating duties, they might choose OpenVMS as an alternative to z/OS and friends.

On Thursday, November 3, 2022 at 5:45:54 AM UTC+11, John Dallman wrote:
> In article <278cc014-2429-4664...@googlegroups.com>,
> r...@rbcarleton.com (Robert Carleton) wrote:
>
> > One thing that stands out about OpenVMS is its record-oriented
> > filesystem (Files-11), along with its extensive batch facilities.
> > The Linux and Microsoft Windows environments don't really seem to
> > have those kinds of tools as part of their baseline.
> They don't. They follow the classical UNIX model of files as nothing but
> sequences of bytes. Windows' NTFS provides Alternate Data Streams, which
> were created to support classic Mac OS resource forks, but they're very
> little used. There are third-party batch systems for both Linux and
> Windows, but I haven't looked at them this century.
>
> Record-orientated filesystems are really quite alien to people used to
> the UNIX model, and caused me significant FUD when I first started
> working on VMS. This wasn't helped by record-orientation not actually
> being useful for the work that I was doing, supporting programmers
> working on code that had to work on many platforms.
> > It seems like at least some organizations where there is a significant
> > culture separating duties, they might choose OpenVMS as an alternative
> > to z/OS and friends.
> Well, VMS is doubtless a /lot/ cheaper to run than an IBM mainframe, but
> has less certainty of being around in the long term. Going after some of
> the other legacy mainframe markets might make more sense, but they're all
> rather different from each other.
>
> John

The IBM Mainframes are seeing expansion where I work

Security is seen as paramount and being able to wrap up whole applications under a single framework in terms of security is seen as a huge advantage

We are seeing certain applications being moved off other platforms and back onto the Mainframes, especially anything pertaining to data transfers and external interface exposure

I can't see VMS competing in this space, IBM really has done a good sales job on the security side even down to enabling security aspects on their CPU's that will offload certain operations to hardware for increased protections in flight (good sales pitch no doubt)

I would have thought VMS could leverage it's historical reputation in security to give it an advantage against Linux at least, but I'm not convinced it has done enough to ensure it's up to date in the modern security landscape and it really needs to make sure it has it's ducks all in a row and then some because any failure in the security arena could/would end VMS chances of making a comeback

On 11/2/2022 1:32 PM, Robert Carleton wrote:
> One thing that stands out about OpenVMS is its record-oriented
> filesystem (Files-11), along with its extensive batch facilities. The
> Linux and Microsoft Windows environments don't really seem to have
> those kinds of tools as part of their baseline.

Records for sequential files is mostly a hassle from application
perspective.

But index-sequential files is a "builtin NoSQL key value store".
It is not a problem getting one of those on other platforms
(RocksDB or whatever), but VMS got it.

And batch are useful as well. There are other options - from
expensive enterprise schedulers to Quartz for the J world. But
again VMS got it.

None of this is the killer feature that will make customers
stand in line at VSI's front door, but it is still something
that VMS has.

And I don't believe in the unique killer feature
idea at all - such get copied very quickly.

Arne

On 11/2/2022 4:06 PM, IanD wrote:
> The IBM Mainframes are seeing expansion where I work

IBM mainframe sales has been falling for decades. Not dramatically but
consistently. Or to be more correct: a consistent downwards trend
with lots of cycles correlating to model releases.

But different companies chose different strategies. Some companies
move off their mainframe(s). Some reduce usage of their mainframe(s).
Some keep usage of their mainframe(s). It does not surprise me that
some increase usage of their mainframe(s).

> Security is seen as paramount and being able to wrap up whole
> applications under a single framework in terms of security is seen as
> a huge advantage
>
> We are seeing certain applications being moved off other platforms
> and back onto the Mainframes, especially anything pertaining to data
> transfers and external interface exposure
>
> I can't see VMS competing in this space, IBM really has done a good
> sales job on the security side even down to enabling security aspects
> on their CPU's that will offload certain operations to hardware for
> increased protections in flight (good sales pitch no doubt)

IBM has done a way better job with z/OS than HP/HPE did with VMS.

> I would have thought VMS could leverage it's historical reputation in
> security to give it an advantage against Linux at least, but I'm not
> convinced it has done enough to ensure it's up to date in the modern
> security landscape and it really needs to make sure it has it's ducks
> all in a row and then some because any failure in the security arena
> could/would end VMS chances of making a comeback

VMS definitely has some catching up to do.

Arne

> -----Original Message-----
> From: Info-vax <info-vax-bounces@rbnsn.com> On Behalf Of Arne Vajhøj
> via Info-vax
> Sent: Wednesday, November 02, 2022 9:13 PM
> To: info-vax@rbnsn.com
> Cc: Arne Vajhøj <arne@vajhoej.dk>
> Subject: Re: [Info-vax] What does VMS get used for, these days?
>
> On 11/2/2022 4:06 PM, IanD wrote:
> > The IBM Mainframes are seeing expansion where I work
>
> IBM mainframe sales has been falling for decades. Not dramatically but
> consistently. Or to be more correct: a consistent downwards trend with
lots
> of cycles correlating to model releases.
>
> But different companies chose different strategies. Some companies move
> off their mainframe(s). Some reduce usage of their mainframe(s).
> Some keep usage of their mainframe(s). It does not surprise me that some
> increase usage of their mainframe(s).
>
> > Security is seen as paramount and being able to wrap up whole
> > applications under a single framework in terms of security is seen as
> > a huge advantage
> >
> > We are seeing certain applications being moved off other platforms and
> > back onto the Mainframes, especially anything pertaining to data
> > transfers and external interface exposure
> >
> > I can't see VMS competing in this space, IBM really has done a good
> > sales job on the security side even down to enabling security aspects
> > on their CPU's that will offload certain operations to hardware for
> > increased protections in flight (good sales pitch no doubt)
>
> IBM has done a way better job with z/OS than HP/HPE did with VMS.
>
> > I would have thought VMS could leverage it's historical reputation in
> > security to give it an advantage against Linux at least, but I'm not
> > convinced it has done enough to ensure it's up to date in the modern
> > security landscape and it really needs to make sure it has it's ducks
> > all in a row and then some because any failure in the security arena
> > could/would end VMS chances of making a comeback
>
> VMS definitely has some catching up to do.
>
> Arne
>

Oct 24/22 analysis of IBM, mainframes and revenue streams:
<https://www.nextplatform.com/2022/10/24/the-ever-reddening-revenue-streams-
of-big-blue/>

Interesting extract:
"But don’t get the wrong idea. For IBM, hybrid is the next platform, not Red
Hat. There will come a day, perhaps, when the Red Hat platform contributes
as much revenue – and maybe even the same profit – as the two enterprise
system families that have defined IBM for most of its history. But rest
assured: When we are all dead and buried, there will still be mainframes and
Power Systems running mission critical, back-end systems deep in the bowels
of tens of thousands of companies worldwide. As it is, there are about
165,000 of them today, and they spend a premium price for a premium product
that they know how to make sit up and bark as well as any hyperscaler knows
how to make a cluster hum. At the prices IBM charges for its systems –
particularly mainframes – they have to run them at 99.5 percent utilization.
And with mainframes, they do, which is an amazing thing you never hear about
RISC/Unix servers or clusters."

Regards,

Kerry Main
Kerry dot main at starkgaming dot com

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com

On 11/2/22 20:12, Arne Vajhøj wrote:
> On 11/2/2022 4:06 PM, IanD wrote:
>> The IBM Mainframes are seeing expansion where I work
>
> IBM mainframe sales has been falling for decades. Not dramatically but
> consistently. Or to be more correct: a consistent downwards trend
> with lots of cycles correlating to model releases.
>
> But different companies chose different strategies. Some companies
> move off their mainframe(s). Some reduce usage of their mainframe(s).
> Some keep usage of their mainframe(s). It does not surprise me that
> some increase usage of their mainframe(s).
>
>> Security is seen as paramount and being able to wrap up whole
>> applications under a single framework in terms of security is seen as
>> a huge advantage
>>
>> We are seeing certain applications being moved off other platforms
>> and back onto the Mainframes, especially anything pertaining to data
>> transfers and external interface exposure
>>
>> I can't see VMS competing in this space, IBM really has done a good
>> sales job on the security side even down to enabling security aspects
>> on their CPU's that will offload certain operations to hardware for
>> increased protections in flight (good sales pitch no doubt)
>
> IBM has done a way better job with z/OS than HP/HPE did with VMS.
>
>> I would have thought VMS could leverage it's historical reputation in
>> security to give it an advantage against Linux at least, but I'm not
>> convinced it has done enough to ensure it's up to date in the modern
>> security landscape and it really needs to make sure it has it's ducks
>> all in a row and then some because any failure in the security arena
>> could/would end VMS chances of making a comeback
>
> VMS definitely has some catching up to do.
>

Personally, I would really like to see VMS catch up in the security
sphere and go after the government market (specifically DOD who have
a very strong need for secure systems). A conversation with DISA
would go a long way in accomplishing that. But I have no expectation
that it would ever happen.

bill

On Thursday, November 3, 2022 at 7:20:10 AM UTC-5, Bill Gunshannon wrote:
> On 11/2/22 20:12, Arne Vajhøj wrote:
> > On 11/2/2022 4:06 PM, IanD wrote:
> >> The IBM Mainframes are seeing expansion where I work
> >
> > IBM mainframe sales has been falling for decades. Not dramatically but
> > consistently. Or to be more correct: a consistent downwards trend
> > with lots of cycles correlating to model releases.
> >
> > But different companies chose different strategies. Some companies
> > move off their mainframe(s). Some reduce usage of their mainframe(s).
> > Some keep usage of their mainframe(s). It does not surprise me that
> > some increase usage of their mainframe(s).
> >
> >> Security is seen as paramount and being able to wrap up whole
> >> applications under a single framework in terms of security is seen as
> >> a huge advantage
> >>
> >> We are seeing certain applications being moved off other platforms
> >> and back onto the Mainframes, especially anything pertaining to data
> >> transfers and external interface exposure
> >>
> >> I can't see VMS competing in this space, IBM really has done a good
> >> sales job on the security side even down to enabling security aspects
> >> on their CPU's that will offload certain operations to hardware for
> >> increased protections in flight (good sales pitch no doubt)
> >
> > IBM has done a way better job with z/OS than HP/HPE did with VMS.
> >
> >> I would have thought VMS could leverage it's historical reputation in
> >> security to give it an advantage against Linux at least, but I'm not
> >> convinced it has done enough to ensure it's up to date in the modern
> >> security landscape and it really needs to make sure it has it's ducks
> >> all in a row and then some because any failure in the security arena
> >> could/would end VMS chances of making a comeback
> >
> > VMS definitely has some catching up to do.
> >
> Personally, I would really like to see VMS catch up in the security
> sphere and go after the government market (specifically DOD who have
> a very strong need for secure systems). A conversation with DISA
> would go a long way in accomplishing that. But I have no expectation
> that it would ever happen.
>
> bill

Part of that would be there being SRG/STIGs being available for OpenVMS. I wasn't able to find any. That will complicate the authorization of new OpenVMS systems in DoD circles.

On 2022-11-02, Arne Vajhøj <arne@vajhoej.dk> wrote:
>
> And I don't believe in the unique killer feature
> idea at all - such get copied very quickly.
>

It took a quarter of a century for clustering on other operating systems
to catch up to VMS clustering.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On 03/11/2022 13:26, Simon Clubley wrote:
> On 2022-11-02, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>
>> And I don't believe in the unique killer feature
>> idea at all - such get copied very quickly.
>>
>
> It took a quarter of a century for clustering on other operating systems
> to catch up to VMS clustering.
>
> Simon.
>

Has it actually caught up?

--
Chris

In article <tk0r9u$1f47k$1@dont-email.me>,
Chris Townley <news@cct-net.co.uk> wrote:
>On 03/11/2022 13:26, Simon Clubley wrote:
>> On 2022-11-02, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>>
>>> And I don't believe in the unique killer feature
>>> idea at all - such get copied very quickly.
>>>
>>
>> It took a quarter of a century for clustering on other operating systems
>> to catch up to VMS clustering.
>
>Has it actually caught up?

This is a definitional question. Other systems
took a different tack, where "clustering" happens
at a higher level (e.g., an RDBMS or something).
In that sense, not only have they caught up, but
they've surpassed what VMS does.

- Dan C.

On Thu, 2022-11-03 at 13:26 +0000, Simon Clubley wrote:
> > And I don't believe in the unique killer feature
> > idea at all - such get copied very quickly.
> >
>
> It took a quarter of a century for clustering on other operating
> systems to catch up to VMS clustering.

That's not my understanding. VMS clustering is unique and still is.
Linux clustering doesn't have a global uptime.
--
Tactical Nuclear Kittens

On 11/3/2022 12:46 PM, Chris Townley wrote:
> On 03/11/2022 13:26, Simon Clubley wrote:
>> On 2022-11-02, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>> And I don't believe in the unique killer feature
>>> idea at all - such get copied very quickly.
>>
>> It took a quarter of a century for clustering on other operating systems
>> to catch up to VMS clustering.
>
> Has it actually caught up?

IBM came up with HACMP for AIX in the early 90's.

Linux got DLM and cluster file system (GFS) in the early 00's.

In the mean time the industry moved into another direction.
Application clusters instead of OS clusters.

So instead of having:
- applications X and Y on OS A using foo cluster software
- applications X and Y on OS B using bar cluster software
it was:
- application X using foo cluster software on OS A and B
- application Y using bar cluster software on OS A and B

Arne

On 11/3/2022 2:32 PM, Dan Cross wrote:
> In article <tk0r9u$1f47k$1@dont-email.me>,
> Chris Townley <news@cct-net.co.uk> wrote:
>> On 03/11/2022 13:26, Simon Clubley wrote:
>>> On 2022-11-02, Arne Vajhøj <arne@vajhoej.dk> wrote:
>>>>
>>>> And I don't believe in the unique killer feature
>>>> idea at all - such get copied very quickly.
>>>>
>>>
>>> It took a quarter of a century for clustering on other operating systems
>>> to catch up to VMS clustering.
>>
>> Has it actually caught up?
>
> This is a definitional question. Other systems
> took a different tack, where "clustering" happens
> at a higher level (e.g., an RDBMS or something).
> In that sense, not only have they caught up, but
> they've surpassed what VMS does.

Yes.

Application clusters over OS clusters.

Shared everything:

Databases - Oracle RAC, DB2 Purescale, ASE Cluster Edition,
Cache servers
Application servers

Shared nothing:

Databases - MySQL cluster, DB2 Warehouse Edition, Cassandra, HBase,
MongoDB etc.
Web servers

Arne

Arne Vajhøj wrote:

> In the mean time the industry moved into another direction.
> Application clusters instead of OS clusters.
>
> So instead of having:
> - applications X and Y on OS A using foo cluster software
> - applications X and Y on OS B using bar cluster software
> it was:
> - application X using foo cluster software on OS A and B
> - application Y using bar cluster software on OS A and B

Then shove a hypervisor underneath the lot, so you can push and pull the servers
around, largely independent of the hardware.

> -----Original Message-----
> From: Info-vax <info-vax-bounces@rbnsn.com> On Behalf Of Chris Townley
> via Info-vax
> Sent: Thursday, November 03, 2022 1:47 PM
> To: info-vax@rbnsn.com
> Cc: Chris Townley <news@cct-net.co.uk>
> Subject: Re: [Info-vax] What does VMS get used for, these days?
>
> On 03/11/2022 13:26, Simon Clubley wrote:
> > On 2022-11-02, Arne Vajhøj <arne@vajhoej.dk> wrote:
> >>
> >> And I don't believe in the unique killer feature idea at all - such
> >> get copied very quickly.
> >>
> >
> > It took a quarter of a century for clustering on other operating
> > systems to catch up to VMS clustering.
> >
> > Simon.
> >
>
> Has it actually caught up?
>
> --
> Chris
>

The difference between OpenVMS clustering and other platforms is really a difference in technical strategies between shared disk (OpenVMS, z/OS, Linux GFS) vs. shared nothing (Windows, most Linux, UNIX's).

A good whitepaper which leaves out the OS religion and marketing and that discusses the pro's and con's of each cluster strategy.

Shared Nothing v.s. Shared Disk Architectures: An Independent View
<http://www.benstopford.com/2009/11/24/understanding-the-shared-nothing-architecture/>

A good extract:
" Shared Disk Architectures are write-limited where multiple writer nodes must coordinate their locks around the cluster. Shared Nothing Architectures are write limited where writes span multiple partitions necessitating a distributed two phase commit."

Regards,

Kerry Main
Kerry dot main at starkgaming dot com

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com

On Friday, October 14, 2022 at 8:35:06 PM UTC-5, Scott Dorsey wrote:
> John Dallman <j...@cix.co.uk> wrote:
> >
> >My employers used it as a software development system, producing
> >mathematical modelling code for VMS, plus a wide range of other platforms.
> >Demand for the code on VMS shrank in the 1990s, and it became expensive
> >compared to doing development on Windows. We had dropped it by the year
> >2000. We'd resume support if there was significant demand for it on
> >x86-64, which is why I joined this newsgroup.
> In the eighties we did a lot of this, and we did this in spite of VMS being
> badly-suited for scientific computing. We did it because the DEC hardware
> was the fastest for the dollar and the DEC fortran compiler was the best and
> most advanced. But we spent a lot of time fighting features in VMS that were
> advantageous in a commercial data processing environment.
>
> When Sun machines started having better floating point performance for the
> buck, we dropped VMS quickly as a scientific programming environment in
> favor of SunOS which was sort of unreliable and worse in a lot of ways
> but nobody cared because the machines were fast.
> >What do you use VMS for in the 2020s?
> Large scale data processing applications that benefit from the heavyweight
> filesystems but aren't totally transaction-based.
> --scott
>
> --
> "C'est un Nagra. C'est suisse, et tres, tres precis."

I work in an environment where there's lots of scientific computing on the Linux platform. One of the things about it, is that getting compute time is very competitive, and our users/coders game the batch systems to gain an advantage in getting their jobs to run. We can't use the stock Linux batch systems (at, batch, atd, cron, and friends) for that work, though the system administrators probably use those for some of what they do. We have to use add-on batch systems for controlling those jobs.

I'm not familiar with the VMS batch facilities yet (I'm a Linux/BSD jockey), but I've heard that they are pretty advanced. Perhaps that would provide an advantage, at least when there is a lot of competition for compute resources.

On 11/5/2022 8:25 PM, Robert Carleton wrote:
> I work in an environment where there's lots of scientific computing
> on the Linux platform. One of the things about it, is that getting
> compute time is very competitive, and our users/coders game the batch
> systems to gain an advantage in getting their jobs to run. We can't
> use the stock Linux batch systems (at, batch, atd, cron, and friends)
> for that work, though the system administrators probably use those
> for some of what they do. We have to use add-on batch systems for
> controlling those jobs.
>
> I'm not familiar with the VMS batch facilities yet (I'm a Linux/BSD
> jockey), but I've heard that they are pretty advanced. Perhaps that
> would provide an advantage, at least when there is a lot of
> competition for compute resources.

VMS batch queues are pretty good. You setup the queues with
priority, task limit, CPU limit, memory limit etc.. Good cluster
support. The only thing I don't like is the defaults for
the log file - and that is a minor inconvenience.

Arne

On 11/5/2022 8:42 PM, Arne Vajhøj wrote:
> On 11/5/2022 8:25 PM, Robert Carleton wrote:
>> I work in an environment where there's lots of scientific computing
>> on the Linux platform. One of the things about it, is that getting
>> compute time is very competitive, and our users/coders game the batch
>> systems to gain an advantage in getting their jobs to run. We can't
>> use the stock Linux batch systems (at, batch, atd, cron, and friends)
>> for that work, though the system administrators probably use those
>> for some of what they do. We have to use add-on batch systems for
>> controlling those jobs.
>>
>> I'm not familiar with the VMS batch facilities yet (I'm a Linux/BSD
>> jockey), but I've heard that they are pretty advanced. Perhaps that
>> would provide an advantage, at least when there is a lot of
>> competition for compute resources.
>
> VMS batch queues are pretty good. You setup the queues with
> priority, task limit, CPU limit, memory limit etc.. Good cluster
> support. The only thing I don't like is the defaults for
> the log file - and that is a minor inconvenience.
>
> Arne

Not sure what you dislike about batch log files ...

SUBMIT

/LOG_FILE

/LOG_FILE[=filespec]
/NOLOG_FILE

Names the log file. The asterisk (*) and the percent sign (%)
wildcard characters are not allowed in the file specification.

When you use the /LOG_FILE qualifier, the system writes the batch
job's output to the file you specify. If you use the /NOLOG_FILE
qualifier, no log file is created. By default, a log file is
created, is written to the directory defined by the logical name
SYS$LOGIN in the UAF, and is given the batch job's name as its
file name with a file type of LOG. By default, a log file also is
given the batch job's name as its file name with a file type of
LOG.

Seems to cover a lot.

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

On 06/11/2022 04:23, Dave Froble wrote:
> On 11/5/2022 8:42 PM, Arne Vajhøj wrote:
>> On 11/5/2022 8:25 PM, Robert Carleton wrote:
>>> I work in an environment where there's lots of scientific computing
>>> on the Linux platform. One of the things about it, is that getting
>>> compute time is very competitive, and our users/coders game the batch
>>> systems to gain an advantage in getting their jobs to run. We can't
>>> use the stock Linux batch systems (at, batch, atd, cron, and friends)
>>> for that work, though the system administrators probably use those
>>> for some of what they do. We have to use add-on batch systems for
>>> controlling those jobs.
>>>
>>> I'm not familiar with the VMS batch facilities yet (I'm a Linux/BSD
>>> jockey), but I've heard that they are pretty advanced. Perhaps that
>>> would provide an advantage, at least when there is a lot of
>>> competition for compute resources.
>>
>> VMS batch queues are pretty good. You setup the queues with
>> priority, task limit, CPU limit, memory limit etc.. Good cluster
>> support. The only thing I don't like is the defaults for
>> the log file - and that is a minor inconvenience.
>>
>> Arne
>
> Not sure what you dislike about batch log files ...
>
> SUBMIT
>
>   /LOG_FILE
>
>         /LOG_FILE[=filespec]
>         /NOLOG_FILE
>
>      Names the log file. The asterisk (*)  and the percent sign (%)
>      wildcard characters are not allowed in the file specification.
>
>      When you use the /LOG_FILE qualifier, the system writes the batch
>      job's output to the file you specify. If you use the /NOLOG_FILE
>      qualifier, no log file is created. By default, a log file is
>      created, is written to the directory defined by the logical name
>      SYS$LOGIN in the UAF, and is given the batch job's name as its
>      file name with a file type of LOG. By default, a log file also is
>      given the batch job's name as its file name with a file type of
>      LOG.
>
> Seems to cover a lot.
>

Possibly the VFC format?

--
Chris

Robert Carleton <rbc@rbcarleton.com> wrote:
>I work in an environment where there's lots of scientific computing on the =
>Linux platform. One of the things about it, is that getting compute time is=
> very competitive, and our users/coders game the batch systems to gain an a=
>dvantage in getting their jobs to run. We can't use the stock Linux batch s=
>ystems (at, batch, atd, cron, and friends) for that work, though the system=
> administrators probably use those for some of what they do. We have to use=
> add-on batch systems for controlling those jobs.

These days it's also very common to have multiple synchronized jobs on
different nodes, so Linux has add-on tools like PBS for dealing with that.
And you get the added benefit of resource management of course.

>I'm not familiar with the VMS batch facilities yet (I'm a Linux/BSD jockey)=
>, but I've heard that they are pretty advanced. Perhaps that would provide =
>an advantage, at least when there is a lot of competition for compute resou=
>rces.

First thing: VMS has heavyweight processes. There's a lot of stuff in the
process, so spawning off new processes takes a good while, and you don't do
it very often. Conceptually different than Unix and Unixalikes where the
processes are lightweight and the overhead of a fork is minimal so you fork
off a new process for nearly everything.

Whereas the concept of "resources" in Linux is fairly simple, VMS has a lot
of different resources which are managed statically by the operating system.
Some of that resource management goes into making the processes more
heavyweight. This can be a powerful tool to keep multiple users from
interfering with one another on a system with limited resources. In a
scientific computing environment it can also be a pain in the neck because
people will run their job for three days and then hit a working set limit
and need to figure out what the limit really should be.

But yes, some of the "big computer" batch features that you get with PBS and
OS/360 are present by default in VMS, and that's a nice thing. Using VMS on
a machine acting as a front-end to a high-speed computer was great.

Mind you, the best batch management system in the world won't keep researchers
from paying the second shift operators under the table to move their jobs to
the front of the queue.
--scott

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

> -----Original Message-----
> From: Info-vax <info-vax-bounces@rbnsn.com> On Behalf Of Scott Dorsey
> via Info-vax
> Sent: Sunday, November 06, 2022 9:55 AM
> To: info-vax@rbnsn.com
> Cc: Scott Dorsey <kludge@panix.com>
> Subject: Re: [Info-vax] What does VMS get used for, these days?
>
> Robert Carleton <rbc@rbcarleton.com> wrote:
> >I work in an environment where there's lots of scientific computing on
> >the = Linux platform. One of the things about it, is that getting
> >compute time is= very competitive, and our users/coders game the batch
> >systems to gain an a= dvantage in getting their jobs to run. We can't
> >use the stock Linux batch s= ystems (at, batch, atd, cron, and friends)
> >for that work, though the system= administrators probably use those
> >for some of what they do. We have to use= add-on batch systems for
> controlling those jobs.
>
> These days it's also very common to have multiple synchronized jobs on
> different nodes, so Linux has add-on tools like PBS for dealing with that.
> And you get the added benefit of resource management of course.
>
> >I'm not familiar with the VMS batch facilities yet (I'm a Linux/BSD
> >jockey)= , but I've heard that they are pretty advanced. Perhaps that
> >would provide = an advantage, at least when there is a lot of
> >competition for compute resou= rces.
>
> First thing: VMS has heavyweight processes. There's a lot of stuff in the
> process, so spawning off new processes takes a good while, and you don't
do
> it very often. Conceptually different than Unix and Unixalikes where the
> processes are lightweight and the overhead of a fork is minimal so you
fork
> off a new process for nearly everything.
>
> Whereas the concept of "resources" in Linux is fairly simple, VMS has a
lot of
> different resources which are managed statically by the operating system.
> Some of that resource management goes into making the processes more
> heavyweight. This can be a powerful tool to keep multiple users from
> interfering with one another on a system with limited resources. In a
> scientific computing environment it can also be a pain in the neck because
> people will run their job for three days and then hit a working set limit
and
> need to figure out what the limit really should be.
>
> But yes, some of the "big computer" batch features that you get with PBS
> and
> OS/360 are present by default in VMS, and that's a nice thing. Using VMS
on
> a machine acting as a front-end to a high-speed computer was great.
>
> Mind you, the best batch management system in the world won't keep
> researchers from paying the second shift operators under the table to move
> their jobs to the front of the queue.
> --scott
>

One feature in OpenVMS that is not used as much for server resource
management as it might (should?) be is the native class scheduler.

SYSMAN> help class (for more info - can implement by primedays, UIC,
username, account etc.)

Extract from SYSMAN> Help class add
"The class scheduler provides the ability to limit the amount of CPU time
that a system's users receive by placing users in scheduling classes. Each
class is assigned a percentage of the overall system CPU time. As the system
runs, the combined set of users in a class is limited to the percentage of
CPU execution time allocated to their class.

Users might get some additional CPU time if the qualifier /WINDFALL is
enabled for their scheduling class. Enabling the qualifier /WINDFALL allows
the system to give a small amount of CPU time to a scheduling class when the
scheduling class's allotted time has been depleted, but a free CPU is
available."

Regards,

Kerry Main
Kerry dot main at starkgaming dot com

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com