Re: Microsoft destroys TPC-C records!

jahorsch_at_my-deja.com wrote in message <8batfq$7da$1_at_nnrp1.deja.com>...

>SQL Server can scale to the TB level.

And you can use a Harley to plough the fields, but a tractor works much better.

Scalability has little to do with pure database size alone.

A TB database means not only having 1 TB or more of data, but having VLT's or Very Large Tables. Microsoft proved nothing with their terra server. Sure it has a 1 TB database, but the biggest table ito of rows is/was something like 250,000.

With databases, data volume/size problems are usually solved using hardware. Faster controllers. Spreading the data across more disks. Making i/o faster. Make the pipe between the disks and hardware box bigger.

What is critical is the number of rows. Do you think that SQL-Server can handle 1 billion row tables? Do you think that you can scale SQL-Server to be able to process this volume of data in less than 12 hours?

Oracle has been there, and done that. I personally processed 170 million rows in a VLT on Oracle 7.3 a few years ago in -less- than 2 hours. And not, it is not just a question of throwing more hardware to make it go faster. It is a question of how good the core design of the db engine is to be able to process such large amounts of data.

As for the billion row VLT - has been running in a production data warehouse in Germany for some years now (that warehouse was performance tuned by two of my German colleagues back in the mid 90's). I would not be surprised if the that VLT has grown significantly larger in size since.

The fact is and remains that Microsoft is the new kid on the block. Oracle, Informix, Sybase.. well, they have been around the block a couple of times and it takes more than fancy marketing and M$ bs to catch up to them.

Billy Received on Thu Mar 23 2000 - 07:44:00 CST