I saw a story last week which caught my eye about LinkedIn restricting its API to only a small number of approved partners and to a limited set of cases. They have legitimate commercial reasons in wishing to do this in seeking to drive traffic to their site rather than to their perceived competition Salesforce and Microsoft. Twitter, as mentioned in the article had done something similar a few years back.
Now one might pass this over as normal commercial behaviour and the usual rough and tumble between the major Internet properties. However, if you are a small company depending on this API for a feature in your application, this will have created a problem for you. Even more critical, if you are a small development company who has specialised in building applications around this API you are in even more trouble. As someone put it “don’t bet your business on a proprietary API from a large vendor”.
Now in the Cloud world there are also both Open APIs and proprietary APIs out there (most notably AWS) and whilst it would be hard to see someone like Amazon restricting access to its API without significant commercial impact, it is fully within its rights to do so or to make significant changes to it as it sees fit. As indicated, not a probable scenario. More interesting though is the implication within the story that this has resulted in concerns by LinkedIn of traffic flow and customer use to its competitors as the reason for shutting the drawbridge. This is certainly emulated in the Cloud world where different providers may support different APIs and interoperability is not so common place. Most large public Cloud providers want to get your data in and keep it there rather than make it easy for you to port that data between infrastructure and application Clouds.
It was interesting to read this further article highlighting the benefits that Open APIs can bring to the creation of Digital Services within Government.
At DataCentred we have chosen to work with OpenStack because we believe in open APIs and data portability under user control. So if you want to keep control of where your data is and retain control of it we think that using OpenStack and providers who provide full OpenStack API access is a an excellent idea. The alternative is replicating data in a number of silo’d repository with restrictions on access and interoperability.