Re: Oracle TPC Benchmarks and "discrete transactions"

From: bob devine <devine_at_olympus.CS.Berkeley.EDU>
Date: Wed, 2 Sep 1992 20:32:12 GMT
Message-ID: <1992Sep2.203212.22856_at_pasteur.Berkeley.EDU>


tcox_at_netcom.com (Thomas Cox) writes:
: If dt's are useful in a TPC environment, then they will be useful in
: those real world applications that resemble the TPC environment. If
: there are no such applications, then the TPC is useless. Which position
: are you taking?

The discrete transactions that are in Oracle 7 are so constrained that they only seem good for TPC style uses. As to whether the TPC benchmark that Jim Gray developed is close to reality depends on your reality. For an OLTP system, maybe the constrained discrete transactions can be used (hard to say for the general case) but for complex queries it seems doubtful (full-blown transactions are needed there).

The question of whether it is possible to invent a faster mechanism for special cases is a simple "of course it is possible!" But the question is still open about if the constrained mechanism is useful for more than a handful of cases. The IMS approach fastpath was useful, and in my first method I was agreeing that something similar could be usefully added to relational systems (though the more religious may brand that as heresy).

Bob Devine Received on Wed Sep 02 1992 - 22:32:12 CEST

Original text of this message