Amazon Web Services' efforts to capture the enterprise storage market continues with the official launch of its Redshift cloud data warehousing service that's been in invitation-only beta since November last year.
The cloud-based Redshift is targeted at rival offerings from companies like IBM and Oracle, with simple pricing that is a fraction of what competitors charge for on-premises solutions and with substantially better performance, the company claims.
Thanks to columnar data storage, compression and high-performance disk and network input/output technologies, Redshift promises to deliver ten times better performance at one-tenth of the cost of traditional data warehousing solutions hosted on premises, AWS evangelist Jeff Barr said.
The effective cost for storage and processing using Redshift is US$3723 (A$3615) per terabyte per year for on-demand use according to Barr's calculations.
This can be brought down further if customers pick one or three year contracts.
Two types of nodes are available.
High Storage Extra Large comes with 15GB RAM and 2TB of locally-attached compressed storage split across three hard drives and two virtual Intel Xeon E5 cores. It provides 4.4 elastic compute units (one ECU is roughly equal to the CPU capacity of a 1 to 1.2GHz AMD Opteron or Intel Xeon processor from 2007) moderate performance network and disk I/O.
For larger tasks, Redshift has the High Storage Eight Extra Large node with 120GB RAM, 16 terabyte of locally-attached storage over 24 hard drives and 16 virtual Intel Xeon E5 cores for an ECU rating of 35. This node type uses 10 Gbps Ethernet connectivity and features very high disk I/O performance.
Redshift integrates directly with other offerings such as S3 and DynamoDB and will be available in the US East region to start with and will expand to other areas in the coming months, Amazon said.