« July 11, 2006 | Main | July 25, 2006 »

July 19, 2006 Archives

July 19, 2006

Real World Composite Services

Pragmatic BPEL Implementation

Six months ago I demonstrated to a customer how they could take a shiny new E-Business Suite implementation, wrap it with BPEL services and expose thse as web services to a Delphi application.  A couple of months later fired with enthusiasm they had plans to use BPEL for everything from composite applications, to FTP transport to making the tea for the managing director.  I visited the customer again this week to discover that they were still happy with BPEL but they were now more realistic and their initial use of BPEL was to be just for composite applications.  I find this journey instructive because I think this customer has travelled that well known technology curve the Gartner Hype Cycle in a period of just over 9 months.
Gartner Hype Cycle Chart
A graphical representation of the hype cycle is shown above.  The hype cycle can be used to describe the introduction of new technology into the market place but it can also be used to describe the introduction of new technology into an organisation.

Technology Trigger - Hello BPEL!

For BPEL customers the hype cycle starts with their introduction to BPEL and their vision of what it could do for their business.  This is helped along by the vendors sales force who paint a picture of a few quick diagrams leading to automated end to end business orchestrations.  Fired by this the customer starts looking at all the places BPEL could be used.
The customer in point is a good example.  I spent 30 minutes demoing how they could introspect an E-Business suite instance (using the EBS adapter) and create a service that could be readily invoked from their current development tool (Borlands Delphi).  When I returned a week later they were already building their own simple BPEL processes without receiving any training, the techonology was well and truly triggered!

Peak of Inflated Expectations - Everything can be BPEL!

As the customer embraces BPEL every problem they have seems to become a BPEL opportunity.  The BPEL hammer looks set to usher in a new era of hyper-productivity.
My customer looked at all the different things they were doing - FTP transfers, Post Code lookup, data exchange with business partners - and said 'we'll use BPEL to tie all this together'.  This was not necessarily a bad outlook for the long term but the reality is that BPEL was acquired as part of a specific project with specific deliverables and a very specific timeline.

Trough of Disillusionment - Even BPEL takes time!

Ah timelines.  There comes a point in the hype cycle when the customer realises that not everything is necessarily a nail.  Also this shiny new hammer has its own quirks, it must be managed ina production environment, procedure must be put in place for managing release cycles and patches, operations staff need training, the business needs to understand the value of the new tools and approaches.  Oh and what about high availability and disaster recovery.
The trough was probably reached for my customer when they started to have some performance problems retrieving 400 records with 40 fields in each field.  They realised that although BPEL was very productive and generally very efficient it was not a magic bullet to all their problems.  A clever consultant friend of mine (thanks Alan) solved their performance problem (an innovative approach I will share in a future blog) but the seeds of doubt had now been planted.
My customer realised that they had a lot of 'ility' issues to solve before they replaced the whole world with BPEL and they shifted their focus quite rightly from all the things they could do, to all the things they needed to do to get into production.

Slope of Enlightment - Composite Services are low hanging fruit!

Over time as new technology becomes more mature in a user organistation they realise what is the best way to apply it in their own environment.
For my customer the initial 'sweet spot' for BPEL was in the creation of composite services.  This was needed for the project at hand, the other ideas could wait until a later date, leaving existing solutions in place or putting more appropriate solutions in where BPEL was being squeezed to hammer in a screw.
Using BPEL for composite services gave them what they needed now, allowed them to focus on moving to production and provided a nice framework for extended error handling to help in fault resolution.

Plateau of Productivity - Use BPEL where appropriate

Moving forward customers after their first few succesful implementations of a new technology can begin to review the more ambitious usages, but only after they have gained some solid experience in using the technology and gained an appreciation of its strengths and limitations.
When their current project completes in October I hope my customer will be able to review what worked for them, what didn't and look at how BPEL ysage can be extended in the organisatoin to improve developer productivity and business visibility of their operations.
BPEL is a powerful tool that needs to be used appropriately to get the maximum return from investment.  It is difficult to avoid the hype cycle, but at least be aware of it and figure out where you are on the curve!

About July 2006

This page contains all entries posted to Antony Reynolds' Blog in July 2006. They are listed from oldest to newest.

July 11, 2006 is the previous archive.

July 25, 2006 is the next archive.

Many more can be found on the main index page or by looking through the archives.

Powered by
Movable Type and Oracle