On Tue, Mar 23, 2004 at 10:31:29AM -0600, McKown, John wrote:

> Given this, I would gather that the decision to not do NJE on an IFL is a
> "political" decision within IBM, not a technical one. That is, it will work,
> but IBM is actively discouraging its use.

I don't think it's a case of actively discouraging it, but more of a
"show me how it's useful and we'll talk about it". Show IBM a way to
make money, and they're happy to help -- after all, at the end of the
day, they gotta keep us shareholders happy.

> My question is why is z/VM NJE needed on an IFL? I don't think that z/VM on
> an IFL is meant to be used by CMS users for any kind of development. And
> even if it were, I personally would use ftp to send jobs to remote sites,
> not NJE.

FTP is a lot harder to automate. Compare:

ftp mvs.foo.com
userid
password
put foo.jcl
quit

to

SENDFILE foo text to SYSTEM at MVSHOST

You have to deal with a number of possible errors at each step in the
FTP, and recover appropriately -- NJE "just works", and handles all
the retries and errors and queuing due to limited capacity, etc. NJE
still has a place; it's the SNA and CTC connectivity parts that are
the pain. If MVS ever got TCPNJE support, it's be a lot less of a
pain.

IBM is putting a lot of emphasis on Linux development for IFLs, but
consider this: if you could move a big chunk of your editing and
interactive work off TSO and ISPF but still provide easy job
submission and output management that is z/OS-friendly in terms of
device management, how much CPU on the z/OS side would you save? How
long would that allow you to put off a capacity upgrade to your
standard engine LPAR (and the corresponding financial beating from CA, etc for
your 3rd party software)??

If you can combine interactive workloads (for CMS and Linux) on the
IFLs and keep the production bits on zOS, suddenly the resource
requirements vs cost models for z/OS aren't quite so ugly. That's a
big piece of why something like RSCS on IFLs makes sense.

> What am I missing? Remote printing? Use "lpd" on the remote end. Or is this
> for VSE support of some kind? I know nothing of VSE (z/VM, z/Linux, and z/OS
> only at present).

VSE won't run on IFLs, so it's more likely to be the above interactive
processing scenario. Consider also that lpr/lpd lose a lot of
functionality compared to NJE (forms, fwd/backspacing, etc
 -- there's a reason why IPP was invented), and that
not all lpr/lpd implementations are compatible.

It's probably not what IBM intended for IFLs, but since when has the
VM community worried much about that?...8-) Overall, creative uses for
resources have helped the environment to survive. This would be
another creative use for resources that makes the overall environment
better.

-- db

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to