GridIron Systems Reveals a Big Data TurboCharger
GridIron Systems this week, came out of stealth mode with its TurboCharger SAN-attached data acceleration appliance, and announced its immediate availability.
The GridIron TurboCharger, which was just unveiled, was designed to transparently solve performance and scalability problems associated with dealing with Big Data, according to the company.
"Installed transparently in the standard storage network between servers and storage, TurboCharger uses patent-pending "Set-Ranking" technology (with unique silicon and software analytics) to speed up existing storage performance by up to 100 times ... [enabling] applications to run up to 10 times faster," GridIron Systems said in a statement.
To date, setting up many data centers to handle Big Data has been a task that often requires redoing existing infrastructure to accommodate it.
"[GridIron's] TurboCharger appliance is the only solution in the industry that speeds up access to Big Data with zero changes to server, storage, application software, or operational practices," the statement added.
The TurboCharger takes advantage of solid-state memory to stage data to Flash or DRAM in order to match the throughput capabilities of the CPU in a manner designed to eliminate traditional storage bottlenecks.
To do so, the TurboCharger attaches to a SAN switch and is invisible to the servers and storage arrays, acting as a Fibre Channel proxy. Servers act as if they are connected directly to the storage arrays and the storage arrays function as if they are receiving requests directly from the servers, according to GridIron.
The TurboCharger's CPU, in turn, analyzes real-time storage requests and then structures the data into what are called "Bayesian Trees" that are used to predict what data is likely to be needed next. The appliance thus "learns" the optimal access patterns for the data and builds a data model.
GridIron Systems claims that one early customer -- the Shopzilla online price comparison service -- used the TurboCharger to speed up the performance of processing a 40 terabyte Oracle RAC 10g data warehouse by an order of magnitude.