Using technology originally developed to accelerate HTTP traffic, Varnish Software has launched Varnish API Engine, an API management platform that the company claims can process as many as 20,000 API calls a second.
Steve Raby, vice president of sales and marketing, says Varnish API Engine makes use of open source caching software already being used by 2.3 million websites to create an API management platform.
Because the platform is based on open source software, Varnish Software has opted to charge only for the API management features it is including, such as support for throttling, high availability, authentication, authorization, metering and, of course, caching, Raby adds.
In addition, because Varnish API Engine is based on open source software, Varnish Software has opted to charge customers a flat $47,000 annual subscription to use Varnish API Engine versus charging customers per API call made to the platform. That latter pricing model essentially penalizes customers for having a successful application that generates lots of API calls, notes Raby.
Varnish API Engine caches objects in memory to accelerate HTTP performance. Now Varnish Software is applying that same concept to APIs, says Raby. The result is not only a faster API management platform, but also one that needs much fewer server resources to support and can be installed in less than an hour, says Raby. By way of comparison, Varnish Software claims that rival platforms need 50 servers just to support 10,000 API calls a second, which serves to dramatically increase the total cost of deploying an API management platform.
The next version of Varnish API Engine, says Raby, will add support for a graphical user interface to make it easier to navigate the API management platform. The current version provides support for an API and a command line interface to manage the platform.
Raby says Varnish Software is primarily looking for organizations that have applications that make a lot of API calls. As the API economy continues to develop, it’s apparent that not all applications are going to generate the same number of API calls. In addition, different classes of applications have higher sensitivity to latency than others, which may mean they don’t necessarily require an API management platform that can process thousands of API calls per second.