InfiniDB 3

The Simple, Powerful Platform for Big Data Analytics is Faster and More Flexible

Try InfiniDB 3

Read the Press Release

Building on its recognized leadership for simplicity, scalability, and speed, InfiniDB 3 makes it simpler to take advantage of cloud deployments and faster loading of massive data sets.

With new parallel data loading, InfiniDB’s native, in-database map reduction operations and user defined functions (UDF) can more quickly be put to work across distributed hardware. Consuming Big Data for analyses is faster (as much as 1 terabyte per hour !), which means time to decision making is shortened.

To support the need for testing and deploying InfiniDB in the cloud, InfiniDB 3 is available as a prepackaged Amazon Machine Image (AMI), automatically provisioning InfiniDB user and performance modules in private EC2 instances.

InfiniDB 3 New Capabilities

Functionality Characteristics
  • Parallel Data Load designed for Big Data
  • Simple:
    • One simple command maintains operational cost low
    • Several data load configurations possible make it possible to adapt to different deployments and workload needs
  • Scalable:
    • Linear performance is maintained as more nodes participate in the loading process
  • Fast:
    • No query performance degradation during data load allows user to continue their work unaffected
  • Transparent provisioning and run-time operations on Amazon EC2
  • Simple:
    • Prepackaged AMI allows automatic provisioning of InfiniDB nodes on EC2
    • Transparent support of EC2 virtual storage and data redundancy (EBS) polices lower administration costs
  • Scalable:
    • Compatibility with Amazon provisioning allows users to benefit from on-demand provisioning and de-provisioning
  • Fast:
    • Same amazing performance as an on-premise deployment, but without the hardware maintenance overhead

InfiniDB 3 Parallel Data Load Configurations

Parallel Load Configurations Description
  • Bulk Load from a central location
  • 1 data source (single file)
  • 1 single command
  • Automatic data distribution across a suite of shared-nothing disk resources
  • Single Bulk Load Command, Partitioned Data Source
  • N data sources
  • N PM nodes
  • 1 single command
  • Independent, Parallel Bulk Load Command, Partitioned Data Source
  • N Partitioned load files
  • N PMs
  • N bulk load commands
  • 1 file/source can be loaded onto 1 PM without any coordination with a central point. N concurrent loads into 1 table will be allowed, assuming N source files.