Following up on its unveiling in June of last year, San Francisco-based MemSQL has released a distributed version of its eponymous in-memory database, aimed at introducing big data scalability to the low latency performance provided by the initial release.
With its distributed version, MemSQL is providing in-memory performance, with access via the common SQL database access language. The company says the offering is already in use for applications, such as operational analytics, network security, real-time recommendations, and risk management.
MemSQL scales out across commodity hardware, and has already been deployed in production use across hundreds of nodes, with sub-second response times on terabytes of data. Data redundancy and security is provided by duplication across nodes, by checkpoints to physical disk, and across data centres.
The company has worked with Morgan Stanley to create a real-time bond data application that is used by 25,000 financial advisors nationwide. MemSQL allows the development team to balance high-velocity data streaming into the system with a large number of concurrent queries. With MemSQL, Morgan Stanley has been able to manage big data workloads, accelerate development and reduce total cost of ownership by scaling on commodity hardware.
Also new is MemSQL Watch, a browser-based interface for software and hardware monitoring, and system configuration.