Thursday 17th November, 2011
4:50pm to 5:50pm
Almost all internet Enterprises are providing some sort of a Web API enabling the community at large to build amazing and cool mash-up applications. Increasingly, the same API/Services model is also internally used within the Enterprise, to increase the agility and productivity. These API/Service communications require (de)serializing the messages into some data format, such as XML or JSON. These formats suffer from performance issues at high message volumes and payloads. Binary XML formats such as Fast Infoset have not fully addressed this either. Better optimized and efficient formats and serialization techniques have emerged over the past few years. Examples include Protocol buffers, Avro, and Thrift. This talk covers how these different formats compare and contrast, how they address the performance problems, and how such formats can seamlessly be leveraged even when the API message definitions are expressed using XML schema. This will be demonstrated using the Turmeric open source SOA platform. The talk will also cover API versioning implications on these data formats and the potential approaches to address them.
Sign in to add slides, notes or videos to this session