Hello,

Please suggest a best way to access and preserve 180 million records in both database perspective and hardware perspective.

Regards,
Siva

Too many ambiguities . what do you mean by "access" , "preserve".
In the jargon of database,use read only ,import files and etc

regards,

mujeeb

RAID-5 will provide good data protection and will not penalize reads. There will be a penalty on writes (about 20% depending on the number of disks in the array). Some advanced disk systems like IBM Shark do not have write penalty when using RAID-5.

When setting up tablespaces and containers on DB2 it is very important to do this in an optimal manner, taking into consideration stripe size, extents, and prefetch quantity.

Using multiple containers for a single tablespace, with each container on a separate physical disk or separate array, may enable intra-partition query parallelism. It helps to have at least one CPU for each container in a tablespace.

Using DB2 ESE with DPF and creating multiple partitions will enable inter-partition parallelism. You should have at least one CPU per partition.

Having very large buffer pools (backed up with real system memory) is also important.

We have a 5CPU, 4GB RAM machine with database residing on EMC SAN(RAID-5 Implemented).We do have partitions created for the tables. But still we are skeptical about the performance of querying through 180 million records.

Average size of the row is 600 Bytes, the kind of access pattern is "Select Only" , the kind of load will be "more than average (through screens)" and our application is a transactional application.

Please suggest on hardware and database configuration.