Convergence Coming for Cloud Computing, Semantic Web
Uche Ogbuji
SemanticWeb.com Contributor
Cloud computing is yet another one of those old realities repackaged as a new idea, with a fancy new buzzword. The Internet has long been represented in network diagrams as a cloud, to indicate that users and applications don’t have a precise picture of what lies out there, nor do they need one.
When you make a request — to send e-mail, for example — or to get a Web page, you are never sure what path of servers and infrastructure the request will wander in its fulfillment. Meanwhile, on the desktop and in the enterprise we’ve become used to very tightly controlled and understood systems for software services. The idea of cloud computing is that even more traditionally controlled applications can evolve into Internet applications — supplied by an indeterminate and constantly changing service environment.
At first, e-mail was something your ISP or employer closely controlled for you. Then more and more people started using services such as Hotmail, Yahoo mail, and later GMail. These services can be used by any user independently of any other commercial relationship, and thus work the same when the user changes employer or ISP, or wherever they might travel. This is a classic characteristic of cloud computing. It involves a relatively loose relationship between the user and the service provider. Eventually even entire organizations were outsourcing their corporate e-mail to remote vendors.
What’s more interesting is the degree recently to which people and organizations embrace the cloud from the sorts of applications where they have traditionally cherished control. The idea of shipping sales support services to SalesForce.com, or front office apps to Google apps would have seemed unthinkable just a few years ago, but now these are very real trends.
Cloud computing is an even more natural concept in Semantic technology. One of the first benefits from improving the context of information flowing through the Web is to make rich services available on the Web. Back in 1997 a Scientific American article by Tim Berners-Lee and other Semantic Web pioneers outlined the ambitions of the technology. The article highlighted a scenario of complex personal travel coordination — a utopia of cloud computing as well as Semantic Web. In the scenario the various information and action components to fulfill the task were assembled from a variety of sources and providers in an opportunistic way.
Contemplating such scenarios, the Semantic Web community immediately recognized some of the issues that follow, such as how to enforce policies, and whom to trust. As such, the community took up policy management and the Web of trust right away, and now you are seeing such matters dominate discussion of cloud computing. This is especially important, because one of the things driving cloud computing is its connection to the mobile enterprise. With more and more people using Blackberries, iPhones and such for serious work, it’s more and more practical to put the services they need in the cloud, and reuse the wealth of cloud-friendly applications available for such devices.
This of course makes anyone concerned with policy control very nervous, and organizations will need to concern themselves with how well usage of the cloud is governed by business rules, and the Semantic Web’s head start might prove useful.
The advantage of closely controlled services is that you make a focused determination of that provider’s worthiness, and then you just proceed with belief in high enough service availability and tractability. Cloud computing involves looser ties with your service provider, and so you have to constantly think of whether they are reliable and transparent. You want to be able to rapidly migrate from one service to another, which means that you take close control of your data flow, and of the interoperability of the service interfaces. Yes, this means that you haven’t really eliminated the need for close control, but you’ve just shifted its focus from management of the provider relationship to management of data and workflow.
Looking at it through that lens, you can see why Cloud computing is booming in the era of SOA. This is the direction the industry is taking: looser coupling, but better transparency and interoperability. What you lose in control of vendor relationships you gain in cost, thanks to market forces, and flexibility that improves how you yourself can respond to market forces.
But again, this improvement comes with better management of data and workflow. Rather than crossing your fingers and jumping into the cloud, it’s important to work towards enough maturity of data and workflow to gain the broader benefits of the cloud. Cloud computing pursuits often go hand-in-hand with SOA implementations for enterprise applications. Cloud computing, however, accentuates the extent towards which you should not just be preparing for a set of services interfaces for your internal IT, but also for focus on information and agnosticism to interfaces.
As I’ve advocated in this column, this preparation involves placing identifiers and data relationships at the heart of services. Expose Web-friendly identifiers for any data involved in cloud interactions, and for the components of the relevant policies and workflow, and for the users, sources, and relevant history. This will allow you to pull together a framework for control of cloud usage.
Cloud computing puts applications in their place — incidental to the information they process. Increasing the information quality of our raw data reduces the cost and complexity of cloud computing. This provides a natural opening for semantic technology and, you can expect to see a greater convergence between the two technologies.

The 
Eric Franzon
VP Community
Jennifer Zaino
Contributor
Angela Guess Contributor
semanticweb.com Twitter feed loading...