Reinventing the wheel, with different tyres
Are we ready for a cloud computing applications market? It's coming, if major vendors' plans are anything to go by.
Cloud computing isn't an entirely new concept, but like many reinvented ideas in corporate computing, it's the approach that is changing. Hosting applications remotely and accessing them using thin clients is something that we've been doing since the first video terminals began appearing in the 1960s. Today, cloud computing brings fluidity to the concept, enabling us to better cope with variations in client demand by virtualizing the physical resources (such as storage and computing power) that we give to those applications.
As an industry, we're just getting to grips with cloud computing. Companies are busily applying the principles behind the firewall, creating 'private clouds' of consolidated, virtualized servers and storage. Service providers are simultaneously applying the same concept to applications and infrastructure designed to be accessed by multiple customers outside the firewall. Corporate customers who recognize the benefits but who are nervous about this idea will first engage in 'hybrid clouds', in which private cloud resources are linked to public ones, enabling them to draw on the benefits of public cloud, while keeping their sensitive data and resources closer to home.
One potential development which pervades all three models of cloud computing is the idea of application marketplaces. The idea is that like-minded communities will form, bound together either organisationally (such as departments under the same company) or by special interests (a community could revolve around accounting, public sector computing, or business intelligence). Users and service providers may begin publishing cloud-based services as services that can be accessed, perhaps for a fee.
The services may be fully-formed applications, or perhaps individual functions, such as a currency exchange lookup, or the batch geocoding of a collection of postal addresses. They will be available, catalogue-style to whoever wants to use them, to avoid people re-inventing the wheel. Companies such as Fujitsu, which is staking its future on cloud computing, are actively working towards these goals.
The interesting part, though, is that it has been attempted before. A decade ago, the web services movement formed the foundation of this push. The Universal Description and Discovery Interface registry was an XML-based specification that would create directories of these services.
It never really took off. Microsoft removed support for it. Vendors such as SAP and IBM removed their public UDDI nodes, and the system was largely mothballed.
Customers were not ready for this - or indeed for the grandiose service-oriented architectures that systems like UDDI were supposed to support. Now, in an era of cloud computing, the industry is gearing up for another stab at it. The question is, in an age of flexible back-end computing, will we succeed a second time around? And can we overcome the political and cultural barriers to this way of working, even after we have solved the technological ones?