On Wednesday 16 December 2009 16:25:43 Dale wrote:
> Alan McKinnon wrote:
> > On Wednesday 16 December 2009 01:34:33 Dale wrote:
> >>> A real world scenario would be a bank server doing transactions. Those
> >>> big irons do never ever get shut down.
> >>> (But they also don't ever get really updated ;)
> >>>
Not true.* :-)
> >>> Did you know, that they still use cobol-code from decades ago. The code
> >>> has to interact with newer systems, but the existing code is not
> >>> allowed to be altered, they just run it inside hugh java application
> >>> servers on their main frames :D
> >>>
Somewhat true, but inaccurate. :-)
> >>> Bye,
> >>> Daniel
> >>
> >> Well, I wish someone would tell my bank that. They are down pretty
> >> regular "upgrading" something. I use the term upgrading lightly here.
> >> It usually makes things worse but anyway. They run windoze on their rig
> >> so they most likely can't help that. ;-)
> >
> > They upgrade the *front*ends*, not the real stuff at the back.
> >
> > Switching a mainframe off is not a supported activity :-)
> >
Again, not true. But sometimes they run so long between IPLs the
operations
staff have to look up the procedures for doing it. :-)
> > Along those lines I could tell you some funny stories about monumental
> > cockups banks do to their front ends (my S.O. does banking data
> > warehousing), but I'm not actually supposed to know some of that stuff so
> > I won't :-)
>
> I'm not sure about back end or front end but they sure make a mess of it
> at times.
>
Of course, not all banks use the same technology, nor do they all have
the
same level of competence. :-)
> >> Hearing they use old code is not to surprising actually. Look at air
> >> traffic control. Every time they try to upgrade, it crashes. I guess
> >> the cheapest bidder is not always the best. o_O
> >
> > Every such crash after an upgrade I know of is trying to run the thing on
> > Windows...
>
> Yep, I read the same thing. Why not use a real OS? I'm thinking BSD or
> something. Linux would be good but I think BSD is even better suited
> for basically 100% uptime.
>
> Dale
>
* This is an interesting discussion, but I feel obligated to point out
that
much of it is fantasy. :-)
As a 30-year veteran of the IBM mainframe programming environment, I
can say
with authority that most of the enterprises that use them for
mission-critical business applications (banking, stock-brokerage, etc.) are
running systems that are updated frequently (sometimes daily) and are fully
capable of being shut down and restarted (on purpose :-D ). Yes, some of
them are front-ended with Linux servers; mainframe systems are not well
designed for managing dynamic web traffic, although systems that do not have
to support very high-volume workflows can do it themselves. The last system
that I worked on was only shut down and restarted twice per year, because 90%
of maintenance could be done while it was running (just like Linux), and
because it was not a business-critical system, it was only required to be
available 99.95% of the time. :-)
The banking and brokerage systems that I first referred to use a more
robust
configuration than we did, which is capable of providing services 100% of the
time, much like a Linux cluster system does. IBM calls the
configuration "Parallel Sysplex." Here's an excerpt of their technical
description, from
<http://www-03.ibm.com/systems/z/advantages/pso/sysover.html>:
'This "shared data" (as opposed to "shared nothing") approach enables
workloads to be dynamically balanced across all servers in the Parallel
Sysplex cluster. This approach allows critical business applications to take
advantage of the aggregate capacity of multiple servers to help ensure
maximum system throughput and performance during peak processing periods. In
the event of a hardware or software outage, either planned or unplanned,
workloads can be dynamically redirected to available servers thus providing
near continuous application availability.
Another significant and unique advantage of using Parallel Sysplex
technology
is the ability to perform hardware and software maintenance and installations
in a nondisruptive manner. Through data sharing and dynamic workload
management, servers can be dynamically removed from or added to the cluster
allowing installation and maintenance activities to be performed while the
remaining systems continue to process work. Furthermore, by adhering to IBM's
software and hardware coexistence policy, software and/or hardware upgrades
can be introduced one system at a time. This capability allows customers to
roll changes through systems at a pace that makes sense for their business.
The ability to perform rolling hardware and software maintenance in a
nondisruptive manner allows business to implement critical business function
and react to rapid growth without affecting customer availability.'
Respectfully,
Leslie
