Today, Xeround officially announced the release of the private beta of its "MySQL for the Cloud" service--an elastic, linearly scalable, relational database designed to run applications in cloud environments.
Xeround is based on an in-memory database and has been tested in a number of telco production environments, according to CEO Razi Sharir. The software utilizes virtual partitions where data partitions are decoupled--or abstracted--from physical resources. These virtual partitions hold copies of both the data and the indexes, in order to ensure high availability and performance.
Despite the ubiquity of open-source MySQL, the database has in the past suffered from scaling issues, which has often led developers down one of two paths: using a NoSQL database to front-end applications; and/or sharding, a process that splits the database into smaller chunks, but requires a high-degree of user sophistication.
According to Xeround CEO Razi Sharir, NoSQL databases like Cassandra and HBase have proven to be a good approach for some projects. But most users simply want to keep using MySQL and SQL in general, but remove the risk of scale issues, rather than replacing or changing applications to follow a NoSQL model.
There is obvious merit to be able to scale MySQL and no question in my mind that a service such as Xeround that allows MySQL to be scaled easily and consistently will have a great deal of interest.
What's less clear is whether enterprises will want to run their database off site in the first place. And, as a fairly staunch supporter of the NoSQL approach, I have to wonder if scaling MySQL is really the issue, or if the issue is more related to application and database design, wherein scale becomes an issue.
Certainly there are limitations with MySQL but it's a bit hard to tell which use cases would benefit from this elasticity instead of using a NoSQL database that's designed for scaling across multiple nodes. For example, would Twitter or Facebook be better served by using a more scalable MySQL, or do their apps really make more sense with a NoSQL model?
MySQL has such a huge installed base and wealth of applications running atop of the database that I have to believe that if the Xeround software works as promised, they will be well positioned to take a leadership position in the database market. Improving and innovating on a proven product is just as often a great business model than inventing something entirely new.