Everywhere you look on the Web these days, it seems as if somebody’s launching new application programming interfaces (APIs). There are over 5,000 documented APIs on the Web, according to some estimates, with another 5,000 expected by the end of 2012. Within four years, that number could increase to more than 30,000 documented APIs. Before long, in other words, it might prove impossible to have a credible Web application that doesn’t include an API that can be called by another application. What does this mean? We could be on the verge of a new “programmable” age of cloud computing, where developers will be able to invoke data inside an application without needing to actually move data into it. That approach may do more than save on storage and networking costs; it’s arguably only through the use of such APIs that cloud applications will be able to truly scale. Indeed, it remains impractical to move large amounts of data around extended networks. Short of shipping physical 1TB hard drives to people overnight, the most practical thing to do is expose an API that, at a bare minimum, gives another application developer access to a summary of that data. APIs and Big Data As massive stores of data (i.e., “Big Data”) become more prevalent, it could become increasingly critical to find a cost-effective method of sharing access to large datasets over networks via an API. In fact, the actual value of any given cloud application could end up directly proportional to either the number of other cloud-based applications that invoke its API, or the business value of the applications that leverage the data it provides. Recently, the explosion of APIs prompted MuleSoft, a provider of application middleware software, to launch Mule iON SaaS Edition, which provides an API for connecting to other applications via a managed service. That offering, which competes with similar services from Dell, Talend and Informatica, offers developers the ability to create their own APIs using MuleSoft connectors. “The end goal,” said MuleSoft CTO Ross Mason, “should not be to transfer data across bandwidth constrained networks, but rather facilitate access to data via well-defined APIs.” Different Ecosystems Carter Lusher, chief analyst for enterprise applications ecosystems at the market research firm Ovum, suggests that different types of ecosystems are beginning to emerge within the API economy. “Some organizations will opt to publish their own APIs to create point-to-point lightweight integration with another company or individual,” he said. “Others will leverage middleware services to integrate their offerings with multiple organizations via a hub and spoke model.” Meanwhile, developers are finding easier ways to stitch together combinations of applications in order to drive new business processes or reinvent existing ones. Much of this phenomenon, Lusher noted, can be attributed to the rise of the Apple App Store, which many developers see as a model for distributing highly integrated lightweight applications. That model could ultimately put a lot of pressure on traditional suites of applications that depend on application servers for integration. The whole notion of a programmable Web is likely to be one of the most profound IT trends of the decade. But oddly enough, the API economy itself may exist long before most companies ever come to realize they’re actually participating in it.   Image: lucadp/Shutterstock.com