In The Press

Skipping ahead with IT flexibility

In times past, one of the most valuable facets of any technology application was its longevity. It was seen as vital that systems were 'built to last', capable of powering the business for years, even decades.

Today, however, the paradigm has shifted and the most important thing about any application is that it is 'built to change'.

Built to change means that a business process currently invoking your finance and warehousing systems can quickly be extended to include your CRM systems. Or your vendor community. Or your webshop. Or all of them.

Gifting existing IT assets with the power to change and grow rapidly is the promise of a service-oriented architecture (SOA), the software industry's attempt to finally align IT with the business.

Delivering on that promise quickly and cost-effectively is the function of the Enterprise Service Bus (ESB), an integration architecture that twins the disciplines of SOA and the power of Web services standards to change both the technology and the economics of application integration and the development of new, orchestrated business services.

Wax Digital is a pioneer of service-oriented architecture, having developed its own enterprise service bus and deployed it to deliver standards-based integration/web services across large enterprise networks and trading partner communities.

However, the complex nature of enterprise systems means SOA is not the complete answer to all ills.

If every business had the luxury of a greenfield IT landscape, i.e. throw it all away and start again, then a thoroughbred web services approach would be a perfect way to build an agile, distributed application set.

This has often been true in the past; it would have been great if applications could be have been written using a client/server methodology, then an object methodology, then an object broker methodology for example.

Equally, if the promise of SOA was that you could genuinely put a thin layer on top of every existing system and expose it effectively and securely so that other third party systems could re-use it, then that would be great. But in reality the technology is not yet at a point where this is either feasible or cost-effective.

So what is the best way to approach delivering integration and new services?

Essentially, it is to do what good IT practice has always done - make the best use of the latest technologies without discarding fundamentally sound existing systems with a huge sunk cost.

The monolythic applications that make up the enterprise IT infrastructure, however, don't behave like web services - to the point where they may only be able to process things once a day in batch mode, and may be unavailable for communication for large chunks of time. They are a mixture of often very old but reliable and efficient systems and it is simply not cost-effective or neccessary to expose these systems as web services when a simple ftp export can do the job required.

An ESB should deliver a platform for creating new services with functions such as security, roles, permissions and user interfaces are all handled under a service-oriented architecture.

The key then is a strategy of transition - evolution not revolution.

SOA, made possible and practical by the use of ESB, is a fantastic way to build and expose new applications but there simply isn't the benefit - yet - in turning everything into a web service.

So keep what works, integrate with it simply, and use web services where they really have benefit; to expose things to the right people with the right permissions - to bring agility to existing systems and new services.

Home	About us News Events Blog Careers Support Contact Site Map