[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [dist-obj] B-A-I-T :) (was: Re: [dist-obj] [Longish] ... )
> I see a lot of (C), people concerned with nitty gritty performance
> issues when some bigger picture is swamping everything else.
... or its generalization into architecture:
a lot of people concerned with nitty gritty architectural
decisions when some bigger picture is swamping everything else.
> I also see a lot of (D), people enamored with technology they don't
> really need in all cases. Even distributed systems sometimes.
Or add an even more drastic twist:
... specially distributed systems.
> > Even worse, one can grow systems no one understands at all.
> > Folks might be surprised to realize how many systems in
> > the world today operate with no one at all grasping
> > how they are effective.
> Like Internet bubbles? 8*)
Most of large enterprise system architecture in the Global 2000
are "unmanaged". They exist, in most cases because they
existed yesterday, and there is an incredible amount of
inconsistencies and embarrassing duplications.
In contrast, suppose you enter a house, and you saw 3
distinct stairs to the upside floors. However, one of
the stairs, abruptly stopped into a wall, with a tiny
little hole where you could see one of the rooms upstairs.
(Of course, the real purpose of this stair, is that it
provides a "budget" and "direct reports" to someone
that has played good politics in the organization.) As
you turn, then you encounter 7 kitchens, and one of them
inside the washroom -- just in case you ever wanted
to cook something there. And to top it all, you have
5 electrical systems, with one that covers the whole
house sparingly, and 4 other more localized but
with overlaps, and 3 plumbing systems, one which is
never used anymore except for the fact that is has
a heating system attached to it.
That's how a typical *enterprise architecture* would
look like if we could only see it. In a weird kind of
way, it is good that software is _invisible_, otherwise
many of us in the IT world would be either categorized
as idiots or as thieves (or both).
I can't disclose the name of this international
multi-billion client of mine, but here is more or less
how their current architecture looks like
(please sit down):
1) Every mission-critical system in the company has
at least one duplicate system. The worst case scenario
is four (4 major distinct mission-critical systems
that do the same thing!!!), with several million
lines of duplicated and inconsistent code (mostly
COBOL, Assembler, C, C++, Smalltalk, Java and VB)
2) over a dozen IBM mainframes (9000 series),
hundreds of UNIX servers and thousands of
NT servers distributed worldwide
3) A growing family of _distributed_ applications
that work against the current mainframe systems
that are furtherly duplicating and making their
architecture even more inconsistent. On top
of that, they are adding several flavors to
this architecture (on a system by system basis):
Java, CORBA, EJB, DNA(MTS), XML all over the place,
VB, ASP, PHP, JSP, etc.; that bridge to Oracle,
DB2, IMS; over Tuxedo, IMS/TM and CICS .....
Of course you are already anticipating the punch line:
Who is going to ever fix their systems and their
Culturally I would say that it would be impossible...
Technically, I would say that it would be hard, very hard...
Philosophically, I would say someone must do it soon or
else the company eventually will not be able to compete with
a better equipped competitor under the assumption that
in the future:
Your system is your business and your
business is your system
Well, I am glad we are in the consulting business -- we will
never run out of work with customers like these....
But it is sad to see so much waste, ignorance, arrogance
and bad politics at work.... in an "industry leader",
To manage your subscription, mailto:firstname.lastname@example.org
Archives, FAQ, etc. http://www.distributedcoalition.org/mailing_lists/