Excellent post Ray... A couple of notes/corrections:

a)  The DW pieces you mention (Quality Manager, PRISM) were NOT acquired via
the UniData acquisition.  Rather, Ardent acquired Prism shortly after VMark
acquired UniData and became Ardent. UniData basically came with the UniData
product and the O2 database, which UniData had recently purchased.  It was
one of the leaders in the world of Object-Oriented Databases.  It's main
point of interest is that it's one of the few pieces of the UniData
acquisition that still remain...  Not the database, but many of the
engineers stayed with Ascential to work on a short-lived product called
Axielle, which was largely a portal product.  They are now the key people
working on our Web Services and Service Oriented Architecture products.  

b)  Regarding /.dshome - actually, we've removed this requirement all
together in the latest releases.  It's all geared at the environment
variables DSHOME.  Full support for rootless installs requires this.
Additionally, this allows for true multi-install support of the engine.
So, you can now install multiple copies/versions of the DSEngine on
a given machine.  An additional but related change was to allow for
Dynamic Shared Attach points.  Given that the UV world is heavily 
based on shared memory for things like concurrency, etc... this
was important.  In UV, the attach points and keys are fixed values,
determined during porting time.  This causes two potential issues.  First,
with conflicts in keys/attach points when hooking in to other products.
Second is potential memory resource limitations.  This is because the attach
point decided also determines the upper bound of addressable memory in UV.
In many cases this significantly limits the amount of memory.  For most UV
applications, this is largely a non-factor, but with DataStage it's a big
deal.  Now, in DS, those values are configurable.  All of these changes are
geared at the different customer that DS faces versus UV applications.
These enterprise customers have different needs, and these changes were
targetted at addressing those issues.

c)  Removals - actually, most of the changes involving component removal
weren't performance related, but rather, maintenance related.  Meaning, many
changes that were going to be required within the product would require also
changing these exterior components that weren't deemed strategic, so they
were removed largely to avoid the overhead of having to update them as well.


d)  As you elude to, it is still interesting to note that any customer who
was familiar with UV up until UV 9.6 would find the DS Engine environment
perfectly normal.  Except for some things that were removed or disabled
(Spooler, UV/Net, Replication), the general environment is exactly the same.
Some things were added to DS after that, and some things were added to UV
after that, but otherwise they are the same.

e)  As to the IBM/Ascential collaboration - there is none.  There was a good
bit during the UV 10.0 creation, since the IBM folks were inexperienced with
UV and asked for some consultation with certain aspects of things, but now
they are doing things purely on their own (as should be the case).  So, I
don't expect any collaboration on either side going forward. 

Dave
-----Original Message-----
From: Ray Wurlod
To: U2 Users Discussion List
Sent: 2/1/2004 4:29 AM
Subject: Re: [ot] Peoplesoft migrates to Ascential


It is to be hoped that the DataStage engineers (some of whom are ex
UniVerse) and the IBM U2 engineers continue to exchange ideas.
Ultimately, however, it is not engineers who decide product directions.
Recall the Golgafrinchan "B" Ark.

_______________________________________________
u2-users mailing list
[EMAIL PROTECTED]
http://www.oliver.com/mailman/listinfo/u2-users

Reply via email to