Posted by: Jack Vaughan
Cloud Services, data architecture
A part of many discussions of massively-scaled cloud computing architecture these days is the notion of “NoSQL.” That is because the trustworthy and ubiquitous SQL data base seems to be playing a less-than-central role in big cloud apps built around Gooble’s Big Table, Facebook’s Cassandra and Amazon’s Dynamo data schemes.
Some of the NoSQL crowd can be pretty strident about SQL’s shortcomings, much as was the object data base crowd when it stood ready to undo the relational data base back in the 1990s. But a toning down of rhetoric is probably due.
”The idea of ‘NoSQL’ started out in a bit of a negative way. But now people tend to mean ‘Not Only SQL’ rather than just ‘NoSQL.’ That is the discussion I am seeing,” said distributed computing veteran Nati Shalom, CTO and Founder of GigaSpaces.
With long experience in the type of applications that require impressive scaling, Shalom is in a unique position to view emerging data architectures for the cloud. He sees scaling issues and RDB issues driving the push to NoSQL – or, Not Only SQL.
”The NoSQL thing came from the realization that scaling comes first. The likelihood that you are going to need to scale is much more acute than in the past,” said Shalom, who added that the demands of social networking apps, particularly, have made distributed application scaling issues more vivid. Like others, Shalom foresees distributed applications evolving that employ both SQL and non-SQL data stores.
For its part, GigaSpaces continues to add capabilities to its flagship XAP application and caching server.
This week marked the roll out of XAP 7.1, which includes Elastic Middleware Services, described by GigaSpaces documents as a simple high-level abstraction of a deployment now exposed via a GigaSpace Administration and Monitoring API. Agents running on machines take care of provisioning and job partitioning. XAP 7.1 also employs an updated version of the light-weight Jetty Web container. Joe Ottinger has more detail.