Use of application programming interfaces (APIs) is expanding rapidly because legacy application integration strategies fit business needs in the client-server world, but fail to meet today's need for real-time, distributed cloud and mobile applications, according to Dell Boomi general manager Chris McNabb. For CIOs, using APIs enables efficient sharing of information and data across applications, something that's been hard to do until now. The challenge is that API-centricity calls for big changes in traditional middleware stacks.
Simply put, an API is a toolset of protocols and routines for specifying how one application can interact with another application. "Think of your products being sold on Amazon.com, or collaborating with SharePoint files via a platform like Yammer," said McNabb in a recent SearchSOA interview. "When properly implemented, APIs make it far easier for software developers to bring together data from various sources to create today's modern Web applications."
Evidence that APIs are an economic force that merits the catchphrase API economy comes from many sources, said McNabb. Consider, he said, that Netflix has used its API to extend its service to over 200 different devices -- television, gaming devices, remote controls -- on all major mobile platforms in less than two years. Then look at predictions for APIs' future. For example, 50% of business-to-business collaboration will take place through Web APIs by 2017, and by next year 75% of Fortune 1000 firms will offer public Web APIs, according to IT research firm Gartner Inc.
McNabb elaborates on APIs' uses, API strategies and changes wrought by cloud integration in this Q&A.
What technology and business changes contribute to the big increase in API use?
Chris McNabb: In the traditional hub-and-spoke model a central server, or hub, performs the data transformations and communicates independently with each application, or the spokes. Today, however, enterprise IT is highly decentralized and includes data assets and huge volumes of real-time data that often travel great distances over the Internet.
In short, we are witnessing the end of middleware as we know it.
In this scenario, the central server creates a choke point and a single point of failure. In addition, with the centralized hub-and-spoke model, enterprises typically need to purchase a separate instance of a vendor's integration product for each integration process, and each product needs to be administered separately, creating an expensive and highly complex environment.
A cloud-based application integration platform turns the integration model on its head. To reduce latency, integration processes take place in close proximity to any number of places where the integration of the applications is required. And while the execution of the integration is distributed at multiple points throughout the enterprise, the development and administrative functions are centralized, providing the ability to develop, deploy and manage all the integration processes anywhere across the WAN from a single location. In short, we are witnessing the end of middleware as we know it.
What are the most common uses of Web APIs today?
McNabb: Web APIs are required for mobile and social applications, as well as for innovative forms of e-commerce that combine Web services and contextual information to deliver a compelling user experience. A Web-based API can expose an on-premises data asset, such as a product catalog or a SharePoint file, to a Web-based application, such as a marketplace or a social media platform. It can also connect on-premises to on-premises, and cloud to cloud.
What's the role of cloud computing in API strategies?
McNabb: A cloud-first [strategy] is appealing because it promises to lower costs and increase agility. However, CIOs should view their API strategy not just as the integration methodology that enables cloud-first, but also as a way to syndicate assets to new internal and external audiences. When considered this way, it becomes clear that a well-conceived and properly implemented API strategy also has the potential to drive innovation, increase productivity, and create new channels and markets.
What software development and integration challenges do moving to cloud as an integration platform address?
McNabb: Adopting an API strategy requires effective management and governance in order to achieve smooth and cost-effective integration. Even enterprises that recognize the value of APIs often treat them as an afterthought on a project list rather than a core feature. As a result, the APIs are not well designed or properly built and can wind up costing tens of thousands of dollars in ongoing maintenance due to infrastructure costs and the drain on engineering resources. A cloud-based application integration platform can help solve these challenges as well by providing a consistent development environment that provides a centralized place to create, manage and monitor APIs and their performance.
Where will API and services dominance impact existing systems most?
McNabb: You'll see a decomposition of ERP systems, as an example, that have been on-premises and in the cloud, or hybrid for a long time. You'll see a constant migration from 'I own and can control everything,' to constant changes from on-premises to cloud; some may come from some areas outside of IT, and that's the architecture of the future.
For example, Gartner is saying that by 2017, CMOs will spend more on IT than CIOs. What will they be doing with that? They will be providing marketing services and answers to the rest of the business -- a sentiment analysis to sales, for example -- to offer up those solutions. And they may have bought it from Salesforce or Radian6, but they'll be using APIs to make those services and data available.
Could you offer advice to CIOs about what assessment and planning stages in enterprise architecture, as that relates to the API economy?
McNabb: Honestly, my advice to CIOs is to really listen to and understand the market landscape and varying business needs. As this services economy comes, they'll need to understand the meaning and significance of all the things they've traditionally been doing.
Talking to industry analysts is a good starting point, allowing CIOs to think through what fits for their business. There is no one-size-fits-all -- the CIO of a Fortune 50 manufacturing firm will have a different take on what type of enterprise architecture works best than the CIO of a midmarket company. So my counsel would be for them to work with analyst firms and industry experts to determine what's best for the organization. There is no one single thing that I would state to that question, I would advise CIOs to read what analysts have to say about platforms, about integration and about mobile and big data, and work with them to tailor their planning and strategy for all that stuff to their organization at their pace.
What's on your wish list for API functionality? E-mail us at [email protected]
About the author
Jan Stafford plans and oversees strategy and operations for TechTarget's Application Development Media Group. She has covered the computer industry for the last 20-plus years, writing about everything from personal computers to operating systems to server virtualization to application development.
How to find happiness in the API market