IBM today unveiled a range of products and services designed to advance the state of API integration, including the API Harmony service running on the IBM Bluemix integration platform that the company claims will make it simpler to identify the contextual relationship between different APIs.
IBM is using the IBM Watson cognitive computing service to both identify the relationship between APIs and make recommendations about what APIs to use to build certain classes of applications.
Meg Swanson, marketing director for Bluemix, says that API Harmony, launched at the IBM InterConnect Cloud 2015 conference, pulls public APIs into a central database that enables Watson to analyze their attributes. Additionally, organizations can create their own databases that can include a mix of private and public APIs that supposedly makes it easier for developers to bind up different classes of services.
By applying Watson to APIs, IBM claims it is revolutionizing API management by applying artificial intelligence in way that makes developers more efficient.
In addition to API Harmony, IBM also announced an orchestration service based on its implementation of OpenStack designed for hybrid cloud computing environments and a Secure Passport Gateway that the company says makes it easier to securely access data stored locally from within a cloud applications via a single point of control.
In addition, IBM announced that via support for Docker APIs, it is now possible to store data in the cloud or locally and then move it to be processed in either environment using patented network extensions to the Docker APIs and IBM DataWorks, a service through which organizations can isolate who gets access to a particular data set.
IBM also announced that it will make a local instance of its Bluemix integration software available in a form that can be deployed on premises as a private cloud.
APIs are clearly at the heart of IBM’s overall hybrid cloud computing strategy. To provide developers with flexibility, a workload can, for example, leverage containers that can run on Linux, Unix or even mainframes to move those workloads closest to where they are being consumed, which theoretically could eliminate bottlenecks in a way that improves overall application performance.
It’s unclear just how much artificial intelligence IBM plans to apply to managing those workloads in the future. But the one thing that is certain is that the more those workloads fall into distinct classes of application workloads, the easier it will become to automate the management of those workloads.