The 2 billion entry directory tree

It's deja vu. I got a note from Oracle's Technology media marvel (and soon to be media mother) Rebecca Hahn touting a new study, commissioned by Oracle, from the Performance Engineering Group at Persistent Systems. It's titled "2 Billion Entry Directory Benchmark." I thought I was back in the 90's!

Youngsters may not remember, but back in the mid to late 90’s there were frequent performance tests done - and more frequent performance claims made – favoring one or another of the various LDAP-enabled directory servers then being sold, but primarily for the Netscape directory server (the grand-daddy of today’s Java Enterprise Directory server from Sun as well as the Red Hat directory server), and Novell’s NDS directory server (now called eDirectory). There were directory load tests, directory read tests, mixed operation tests – whatever test, in short, could show off your directory to it’s best advantage.

The directory shootouts came to a head – and disappeared from the landscape – just about eight years ago when, in March, 1999, Novell announced the 1 billion entry directory tree. One billion entries – who would ever need more?

Oracle would. Larry Ellison likes everything of his to be bigger, better, faster, smarter (but definitely not cheaper) than anyone else’s. So now we have the 2 billion entry tree. (Take that, Microsoft!)

The 22 page white paper doesn’t document a competition, though. As it says in the introduction: “The benchmark objective was to evaluate the scalability of Oracle Internet Directory at very large Directory Information Tree (DIT) sizes, and determine the scalability characteristics of OID under various Lightweight Directory Access Protocol (LDAP) operation workloads. The results demonstrate OID scalability at DIT size of 2 billion entries, with high sustained LDAP operation throughputs.” It goes on to do just that.

Now there’s not a lot of drama in the reading of this report – you know going in that the hero (OID) is going to be triumphant. In fact, the paper chronicles one success after another – there is no struggle. But, then, if you have to build the identity infrastructure for your enterprise you probably don’t want to struggle either.