Comments on: The Oracle DBMS is a “legacy technology”

By: Dave H

Dave H — Fri, 28 Sep 2007 09:09:05 +0000

This is not from some random disgruntled crackpot. It’s from Mike Stonebraker

You have a “not” in there where you shouldn’t have.

By: Martyn Richard Jones

Martyn Richard Jones — Thu, 20 Sep 2007 10:39:52 +0000

Turning the tables … ??? … Mike spins a good old yarn. What will he come up with next, inverted lists?

By: Jim Polizo

Jim Polizo — Wed, 19 Sep 2007 18:33:17 +0000

Efficiency, longness of tooth, or whatever are not the issue. The issue is that if you are a BIG business, you are not going to buy the latest, coolest, fastest, thing that runs on the cheapest hardware with the best algorithms because, if it fails, you will be fired and never work in this town again.

Things like Oracle are there because they have hundreds of thousands of man-hours of development and testing time plus gazillions of hours of real world use. If I put a “2″ into Oracle, it is pretty well understood how to make sure that “2″ goes in, stays in, looks like a “2″ to everyone querying it, and is recoverable if the earth is hit by a comet.

New technologies are welcome but the problems that things like Oracle best lend themselves too will be slow to convert.

By: Pedro Oliveira

Pedro Oliveira — Wed, 19 Sep 2007 00:07:37 +0000

Actually, back in 2004, I have benchmarked Sybase IQ against Oracle 9i on a datawarehouse environment, and I’ve found that, using all of Oracle’s features – namely, data compression and bitmap indexes – we have achieved equal or better response times with Oracle in all operations but one.

The benchmarks were a mix of quite standard operations in our datawarehouse environment – queries and load processes – and were all run on the same machine, a HP9000 with 24 PA/RISC 1GHz CPUs and 32 GBytes memory.

Maybe Sybase IQ has better response times by now. As I remember, Sybase was best with some aggregation operations. Oracle was faster on all the remaining chores, and it took less disk space – which is probably the reason it was faster.

Still I agree with David. The data access will remain purely on a RDBMS model. For most people it doesn’t matter how data is organized inside the database, only how it is accessed.

By: David

David — Mon, 17 Sep 2007 18:21:38 +0000

Stonebraker is talking purely about the technology used for current DBMS *implementations*, not about a replacement for the relational model or the SQL model itself. The columnar technology that Stonebraker has championed for years isn’t intended as an alternative data model at all; it’s a particular mode of storage that in principle could be used by any DBMS, whether relational or not. As far as I know it has been used pretty much exclusively in SQL DBMSs (Sybase IQ for example).

By: Brian Fedorko

Brian Fedorko — Mon, 17 Sep 2007 15:37:12 +0000

I’m afraid that column-oriented DBMS will not take the lead from traditional, relational DBMS. You see, Object Oriented DBMS have promised to outsell traditional Relational DBMS for quite some time – So someday soon, I expect CO-DBMS will have to take the reigns from the OO-DBMS.

Personally, I think I’ll stick with relational, unless I decide that I don’t want to add or delete rows with any sort of efficiency.