15
DECEMBER, 2004
CICS
– the oldie but goodie
It
is interesting to see Peter Abrahams, a senior analyst with Bloor Research,
getting excited about IBM’s recent announcement that it is to introduce
a web services requester application for CICS, its long-standing Transaction
Processing Manager. And he’s right to, for a number of reasons that
make both immediate and, possibly, more long term sense to solutions architects.
The
IBM move, as Abrahams has observed, is in direct response to user pressure.
"The previous release of CICS TS," he writes, "had support
for SOAP but only as a service provider. This meant that new web services
applications could request services from existing CICS applications; this
wrapping of CICS gave the impression that IBM expected all new applications
to be non-CICS web services solutions and that the CICS applications would,
over time, be phased out and replaced by modern solutions."
But
users have taken to this rather more than IBM appears to have expected.
Big Blue obviously saw CICS as not just `mature’ but as a technology
so far past its sell-by date as to be ready for the great archive in the
sky. But users have thought differently and have been asking the company
to come up with new functionality, particularly the ability to develop
CICS applications that are requesters of web services. So IBM has obliged,
with a new implementation that provides support for CICS as a service
requester with full support for WSDL and SOAP.
It
also includes other improvements, such as removal of the 32K limit on
messages between services (which also applies to the size of the COMMAREA
used to communicate between CICS transactions), improved support for C
and C+ giving better performance, and provision of all system management
through a browser interface, rather than TSO. This latter provides a more
effective interface for the operations department and also better integration
of CICS system management with other operations functions.
This
announcement may seem to run counter to the general thrust of technology
development, but is a good indicator of why the service-based concepts
behind SOA put technology firmly into a subservient place. CICS is not
new – it is older than most of the developers and architects that
will have to deal with it. But deal with it they will have to do, that
is the point of service-based architectures. CICS-based applications are
still one of the cornerstones of so many businesses, particularly where
the secure operation of financial transactions is concerned. It is not
going to go away quickly, if at all.
Introducing
a requester application also shows that users see as still playing
an on-going part in delivering business services into the future, rather
than just a system to surface passive historical data.
All
of this possibly points to some longer term potential for the technology
and its application in service delivery. All service provision is about
transactions, and usually remote, loosely coupled, disjointed transactions.
Providing robust, reliable management of complex, disjointed transaction
is still, arguably, one of the weak points of web services and, therefore,
SOA. It is just possible to see `son of CICS’ coming along to fill
just such a gap.
www.ibm.com
|
