Cloud Data Analytics

Having pushed the boundaries of modern technology, ScaleOut Software, the distributed data grid start-up, can now link distributed data grids at different locations into a single logically coherent grid.

Seems a simple thing but it's been illusive and ScaleOut says it's the first to manage it.

So now with its widgetry developers can migrate data across geographies seamlessly without having to replicate it site-to-site.

CEO William Bain remarks that "Global data integration is key to large organizations, such as financial services firms and e-commerce companies, which store fast-changing application data at multiple locations and need to access it from remote sites. Replicating data to all locations is not practical for them, and manually tracking where it is stored and controlling access is costly, error-prone and inefficient."

But wait there's more.

With its shiny new StateServer 5.0, ScaleOut also says it's taken a big step in driving distributed data grids into the cloud and other virtual server environments.

Grids and the cloud are natural elastic allies and ScaleOut's widgetry's got an automatic global data integration feature that lets cloud environments form logical distributed data grids, spanning on-premise and cloud locations, which promises seamless data integration in and out of the cloud.

ScaleOut figures it's the foundation of next-generation parallel data analysis using distributed grids.

Rev 5 also enables VMware users to license StateServer with an installer that will automatically create a VMware distributed data grid appliance and bring up a distributed data grid on VMware virtual servers.

The upgrade apparently boosts StateServer's membership mechanisms for increased performance and robustness in virtual server environments. For example, the company says, heartbeat connections can now adapt to the highly variable network delays often experienced when using virtual servers.

ScaleOut has been something of a Microsoft groupie - well, Bain did sell a previous start-up to Microsoft, which is where Windows Server got its Network Load Balancing - so StateServer 5.0 also features full parallel Microsoft Language Integrated Query (LINQ) and can do fully parallel queries to retrieve .NET and Java-based objects alongside increased security features and enhanced virtual machine support.

That means that for the first time .NET developers can use parallel LINQ queries to access selected groups of data in a distributed data grid through a familiar and standard .NET query mechanism.

Users can automatically retrieve objects from the grid based on their properties, simplifying the structure of their queries while maintaining fast access from all grid servers. Java users have the same parallel query functionality with standard APIs.

The company expects enterprise application users and financial institutions to take to the 5.0 release since it advances the use of distributed data grids for data mining and analysis.

Version 5.0 is supposed to increase security for organizations with multiple applications accessing data within a single distributed data grid. Data that requires higher levels of security can now be protected from unathorized access using an extensible authentication mechanism. Access to named groups of objects in the grid is now controlled and named groups can be authenticated by default using the operating sytem's user accounts.

ScaleOut shipped its first product in January of 2005 and now claims 250 customers including a lot of very familiar names like, oh, Monster and T-Mobile, and says revenues have doubled year-over-year.

The widgetry sells for $1,200-$1,700 per server plus 25% maintenance with a typical installation going for $10k-$20k, Bain said.