InfiniDB™ Enterprise Edition is built to scale to meet your most demanding user, data, and performance requirements for business intelligence and data warehousing. Scaling the capabilities of the MPP architecture is as simple as plugging in another node to the system. This approach delivers a future-proof platform that you can rely on to meet your needs as your business grows along with your data.
- Massive Parallel Processing (MPP) capable: InfiniDB can use multiple commodity hardware machines to achieve linear increases in overall performance. Adding inexpensive hardware to an InfiniDB configuration allows the database to increase processing power so response times can sometimes be reduced in half just with the addition of new nodes. This makes it easy to adjust for growing data volumes, increased user activity, or the desire to meet better overall performance goals. Moreover, this can be accomplished in a dynamic fashion where the InfiniDB system doesn’t go down or be taken offline when new nodes are added.
- Distributed shared-nothing data cache: In a multiple-node InfiniDB configuration, data is distributed among the various nodes and their data caches. No node shares data with the other, however all are accessed in the InfiniDB MPP architecture when data is read to satisfy queries. In essence then, InfiniDB creates one large logical data cache that is accessed in a distributed fashion in parallel by all participating nodes. This allows InfiniDB to literally cache large databases when enough nodes are present with generous amounts of memory.
- Distributed Hash Joins: InfiniDB utilizes hash joins – which are tailor made for processing large data volumes – and distributes the join operation among all participating nodes so query processing speed is heighted as much as possible.
- Automatic failover: InfiniDB’s architecture contains built-in automatic failover for the software modules responsible for managing user queries and performing the actual database I/O operations.
- Automatic performance scale-out: Performance can be scaled by the simple addition of additional nodes.
- Automatic concurrency scale-out: Nodes can be added to a configuration that allows concurrency to be automatically scaled across 1-n machines.
- Automated software patch management: When patches or upgrades need to be applied to more than one InfiniDB server, an automatic process takes the software from the first node where an upgrade is applied and automatically upgrades all other participating nodes.
