SMP vs MPP Architecture

Reading Time: 2 minutes

Databases such as Oracle, DB2, Sybase

Symmetrical Multi-Processing Architecture (SMP) was once the champion of the Data Warehouse. It was rivalled by MPP as SMP had the below disadvantages:

• Failures of components did not result in graceful decline in performance. Rather, the whole system failed and data was unrecoverable until the failure was resolved
• Upon recover, the failed components were unable to rejoin the system with ease
• Failures should not result in data loss
• The system should be scalable and hence support increased load capacity and performance agility

MPP Architecture = Share Nothing = Divide and Conquer

Databases such as Teradata

MPP systems consist of very large numbers of processors each processor has its own memory, backplane and storage. The no shared-resources approach of pure MPP systems allows nearly linear scalability

High availability is another advantage – when one node fails, another can take over.

In Teradata’s MPP Architecture , processor-RAM-storage disk pairs (“nodes”) operating in parallel divide the workload to execute queries over large sets of data. Each processor communicates with its associated disk drive to get raw data and perform calculations. One SMP Host collects intermediate results and assemble the query response for delivery back to the requesting application. With no contention for resources between MPP nodes, this architecture does allow for scalability to petascale database sizes. A major weakness of this architecture, however, is that it requires significant movement of data from disks to processors for BI queries. MPP Architecture