ADSM-L

Re: [ADSM-L] TSM 6.1 Installation Problems

2009-06-04 08:18:29
Subject: Re: [ADSM-L] TSM 6.1 Installation Problems
From: Richard Sims <rbs AT BU DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Thu, 4 Jun 2009 08:18:02 -0400
On Jun 4, 2009, at 7:21 AM, Hans Christian Riksheim wrote:

...
Frankly, I am a bit worried. I look at the bloat at Passport
Advantage(4GB download for reporting? Seriuos?) and it seems that
every
piece of s**t software that IBM has produced now is forced upon us.
Already TSM in itself has a steep learning curve with a lot of
different
"strange" concepts, but IBM obviously doesn't seem to think this is
enough in their ongoing effort to scare off potential customers. An
installation of 6.1 with reporting and administration on a small site
with for example a 2 drive, 40 slot library will require how many
servers? 3? One for TSM, one for admin and one for reporting?
...

Indeed.  I've been with the product since ADSM v.2, and have seen it
grow enormously.  Even having been part of that evolution, the growing
amount of complexity can be overwhelming.  We see customers writing in
having difficulty with even basic functionality in the product: Having
to additionally cope with LAN-Free, Library Manager/Client, encryption
key management, Fibre Channel fabric and other technologies to make
things work has to be daunting.  (And don't overlook inscrutable
licensing regimens.)  This certainly creates opportunities for
competitors offering streamlined, straightforward solutions to data
assurance needs.  Realization of this may be why IBM acquired the
targeted B/R product Fastback from FilesX last year.

The TSM product is obviously trying to accommodate all the latest
technologies out there so as to meet all needs.  The difficulty in
trying to do that is that the result can be a huge monolith of a
product with such intertwined development requirements that
implementing seemingly simple new features can entail an inordinate
amount of time and coordination.  The danger in that approach is in
ending up with a massive composite like Microsoft did with Windows,
with its Longhorn/Vista development debacle.  I'd instead go for a
more modular approach, where customers can acquire and plug in what
they really need, toward more efficient, focused solutions to their
needs.  This would make for more streamlined, timely development, and
new releases that customers eagerly embrace rather than recoil from in
fear.

    Richard Sims