Microsofts dream of Windows 2000 being the de facto
platform standard for the enterprise is still a long way from fulfillment. This is not
because Win2K is not an enterprise class platform (we will really need to wait for the
Data Center Server before judgement can start to be made on that issue), but because
common sense tells us that organisations worldwide are not going to throw away their
current infrastructure and associated applications just so they can have a bright shiny
new toy to play with.
The Microsoft integration strategy
Fortunately, Microsoft has realised this fact albeit, some might say, reluctantly
and it has gone a long way to address the issues of interoperability between and
integration with Unix platforms. Given that there are some two million Unix servers in
operation, it would have been an exceedingly dumb move to ignore these issues. Microsoft
may be many things, but dumb it isnt. Win2K will need to go to market as a platform
that can fit into the heterogeneous systems that exist in all but the smallest market
The Microsoft strategy for integration and interoperability has been multi-focused, but in
the main has been concerned with building bridges and gateways to other platforms,
enabling third-parties and partners to build in interoperability, providing support for a
set of standards to give interoperability.
This strategy has been given form with Microsofts Interoperability Framework, which
is divided into four layers dealing with integration of Network, Data, Applications and
Management. The first three layers are handled by support for standards and the
introduction of the SNA Server. The issue of management has needed a more complex
solution, which is the domain of Active Directory which for a long time appeared to
be the bÍte noire of Win2K. The fact that Microsoft became so possessive of Active
Directory and its functionality gives a clear indication of how seriously it was taking
this layer of the Interoperability Framework. This is hardly surprising given the emphasis
that Microsoft places on TCO and the amount of resource that can be expended on management
of heterogeneous systems.
Application layer and databases
While it is clear that Microsoft has made moves towards integration of Win2K and other
platforms, there are still some niggles as to how far this extends and how strong its
commitment is; especially when the application layer and databases are concerned. The
former is the one that needs to come under the closest scrutiny. One of the stated
standards that Microsoft is supporting is COM+ (which is hardly surprising); however, in
the world of distributed applications that follow the component model, Unix vendors are
fixating on Java, and especially the framework provided by the Enterprise Java Bean (EJB)
standard. Java and Microsoft go together like particularly belligerent cats and dogs, and
while COM and DCOM certainly have more working implementations in operation, EJBs are set
to catch up rapidly now that the tools market is reaching maturity.
As far as the data layer is concerned, the implementation of Microsoft SQL Server as
NT/Win2K specific would also raise concerns. To be fair, given the length of time that
RDBMSs have been around and that we are now into the third version of the ASCII SQL
standard, the fact that we are still relying on ODBC is something that the whole industry
should be ashamed of. It is also an indication that standards are not always the answer.
Wake up Unix
One other aspect of integration between Win2K and Unix needs to be considered. The
emphasis always seems to be on Microsoft and how it will integrate with other platforms.
From an historical perspective this can be considered as a correct way of looking at
things: Unix was here first, therefore the new upstarts should fit in. But history is a
continuous process, and we have reached a point where it needs to be accepted that Windows
is not going to go away, people like it too much for that. Unix vendors should also start
to put their own integration house in order and see how they can improve their platforms
to make them more interoperable with NT/Win2K instead of always trying to find the Unix
flavour that will finish off Microsoft. This is not going to happen; even the mighty Linux
cannot make this happen, no matter how much its supporters would like to believe it.
Integration is always a two-way street, and businesses should start to support those
vendors who offer true interoperability. At least it would demonstrate that they had put
aside their own petty differences and desire for platform power, and were interested in
supplying solutions that would be of real benefit to business as a whole.
Mike Thompson, Director of Research, Butler Group
Check out Butler Groups newest website at www.ITXplorer.com