Archives of the TeradataForum
Message Posted: Fri, 21 Feb 2003 @ 09:40:05 GMT
The difference is in the architecture. Teradata uses a "Shared Nothing" architecture and Oracle uses a "Shared Everything" architecture.
The main reason for this, is history, in that Teradata was developed with "Parallelism" as the keyword from inception. Oracle however was developed for the OLTP market, which at the time did not require any form of "Parallelism", now they are playing catchup. Bolting on "Real Application Clusters (RAC)" to provide a form of "Parallelism".
On the other side of coin, you have to also look at the Hardware.
Oracle has been driven by an ever expanding SMP hardware market. They are one of the main reason Sun, HP, IBM offer servers with up to 128 CPU's. It has only been in that past 12 months, that Oracle have been encouraging people to move to using many small boxes, rather than one large box. This alone can cut costs. The main reason for there change of heart, is that they can finally see that growing a single box will not provide the scalability they want. Teradata has known this for many years, which is why they produce the Worldmark hardware, which is designed for MPP, and more importantly the "Shared Nothing" architecture of Teradata.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|