Are Mashups leading the way to better understanding SOA?
by
Alex Cameron
The publication of the recent Gartner Emerging Technologies Hype Cycle has Mashups snuggling up next to the ill defined Web 2.0 technologies, near the peak of the hype cycle. Mashups are a “popular” and interesting manifestation of how a simple approach to integration of public Web services can be achieved to provide new and innovative applications and functionality.
It is sobering to realise that Mashups are demonstrating to the business community the power of integrating services. Essentially they are similar to intermediary services that produce facades with user interfaces. Even though these basic services have been made public, they are no less integral to the businesses that are exposing them. With appropriate security safeguards in place, they are gaining advantage by extending their reach into the global network indirectly developing social and business networks and allowing applications to be developed that would not be otherwise possible. Probably the leaders in the field are eBay, Google and Amazon, to mention a few. The majority of these derived applications are free. A powerful example that integrates data from both public information (Craigslist) and proprietary data (Google Maps) is http://www.housingmaps.com.
If you need to be convinced as to the power of Mashups, and the rate of adoption and growth, take a look at The Programmable Web. The example above is one of a growing number of Mashups being referenced at this site. They also have some interesting metrics - there are 2.73 Mashups being produced every day (or 500 over the past 6 months)!
The problem these mashers are solving is of course fundamental to what we in the IT industry are trying to solve with SOA. Gartner in their latest application development hype cycle have SOA approaching the bottom of the trough of disillusionment, but give guidance that SOA is inevitable. As most of us know however, businesses have to live with legacy at all levels in their organisation and have to deal with traditional business imperatives, so the implementation of SOA must be slow and evolutionary.
Also hindering SOA is the lack of support it is getting in the area of tools and methodologies. Most of the mashers out there are developing these cool applications on the back of some very fundamental and straightforward technologies, such as PHP, PEAR, JavaScript and AJAX. Whilst arguably these technologies are not enterprise ready (although Microsoft with .NET ATLAS templates is possibly addressing this), in the hands of innovative and creative developers, they can be used to show us some of the way forward. These tools are not only solving the integration problem but also the traditional limitations that we are all aware of with the thin client technologies (browsers) – when coupled together provide not only the utility but the ease of use we need.
Although I don’t think there is much risk that the adoption of SOA (as a business enabler) will not cross the “chasm,” I do think Mashups in the hands of the early adopters, will play an important part in making the technology compelling and getting it accepted in “main street.”
