One of the major issues I continue to hound IBM about on their download
sites is there is no way to subscribe to a change page. So, you have no
idea when code is made available.
Now to the subject of local versus IBM client software. Every site needs to
control the introduction of the client levels to their environments. You
are probably blasting your users with a notification to go to the IBM site
to download the latest client and what level to use. But, consider this,
your internet access is probably overwhelmed every time you do this with
hundreds of users downloading this large client.
In our case we are very controlled on client levels and use SMS on Windows,
so we do use a site repository.
What I really think we want is a way for a user to subscribe to the TSM
server for the latest client and the TSM server provides it with the ability
to reinstall the previous for the customers that are not doing SMS or other
central software control. We also need the ability to push from the TSM
server to the client the next time it connects.
This all said you have very valid concerns. IBM has not thought through how
their site was being used before. They do need to get the software
distribution under some kind of control. That is just good business, but
not at the expense of disrupting our operations. Yes, we may have to make
some adjustments to evolve to a controlled software distribution model, but
at least we would have input to the design and solution and an idea what is
coming.
Paul D. Seay, Jr.
Technical Specialist
Naptheon Inc.
757-688-8180
-----Original Message-----
From: Roger Deschner [mailto:rogerd AT UIC DOT EDU]
Sent: Wednesday, October 30, 2002 12:51 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Does IBM want us to make local client libraries?
In response to one of my many, many complaints about the support web page
train wreck, IBM responded that they now want all client download requests
to go through IBM Registration ID login, since it is licensed software.
I wonder if they contemplated the effects of that?
One option is for each end-user at our site to get and use an IBM
Registration ID. Some already have, but they have complained to me about the
non-navigable site there. While some of the worst malfunctions are being
addressed, it is still not anything I would inflict on end-users. Is IBM
prepared for each of our thousands of ITSM client end-users to register for
one of these new IDs? I'm not quite sure that's what they had in mind with
this change.
The only other option is for each of us to build a local library of client
code. This has all sorts of problems. DISK SPACE AND BANDWIDTH IS NOT THE
ISSUE! The issue is maintaining this library - I now have a new maintenance
headache. And how will I know when a client is updated? I am sure not going
to go check each day or week, just in case a new client is available. This
raises a serious support issue at the installation level.
It also makes it less likely that client code in actual use by end-users
will be up-to-date, which will complicate support calls from end-users to
us, and from us to IBM. When an IBM/Tivoli support rep requests that we use
the latest client code on the end-user's machine, that will now become a
LESS reasonable request to make. Updating it will be a multi-step process,
involving effort on the administrator's part as well as the end-user's part.
Previously, the end-user could comply with such a request in one step.
What we might end up needing is an automated "push" process to send new
client code down to our local client code library. Or just remove the IBM
Registration ID login from the client code download page.
Roger Deschner University of Illinois at Chicago rogerd AT uic DOT edu
========== Only adults have difficulty with childproof caps. ===========
|
