LinkedIn Twitter
Vice President, Research at Gartner, focusing on Data Management and Integration. Three decades in the IT Industry, most recently running his own IT Market Strategy, prior to that with Forrester as Senior Vice President, earlier with the Giga Information Group.

One response to “Oracle Exadata: Early Signs Promising”

  1. rjh

    Super Computer T4 (Oracle)

    Watched that Oracle webcast “again”, extreme parallelism and compression means faster transactions (10 to 50x faster) than the largest business computer (IBM P 795), at a lower cost than the cheapest computer (X86). This was stated as their two primary goals. With data compressed at a factor of 10, 10 x faster means effectively 100 times faster (10x faster “times” 10x compression), moving the data faster but the data is ten times less. There stating10x faster for OLTP, 50X faster for queries for DW/BW/BI (more data). Database server and storage server executing in parallel both running part of the database.

    1000 Exadata servers installed, inspect to triple this year to (4000). Case studies such as Teradata 36 racks where replaced by 3 realizing reduction in floor space, 250TB data base is one tenth of its original size, 1/10 of power consumption. Backup time reduced by a factor of 60x. They point out memory is a single point of failure on P Series, shared system boards, shared CPU and Memory, single point of failure, Exadata has no single point of failure.

    IOPS one Exadata server does more IOPS than 18 million dollar max configuration from IBM. (Exadata 8.8 million for 12 million IOPS verses 1.2 million IOPS at 18 million dollars max configuration on IBM 795).