Three Little Words

PVN? USN? Who was right?

• By Dian Schaffhauser
• February 01, 2001

The argument started this way: Two manuscripts came in, both from excellent writers who understand Microsoft technology, possess MCSE titles, work with Windows 2000 on a regular basis, and have fairly deep knowledge of Active Directory. First, we worked on Curt Simmons’ story on intra-site AD replication. Then we tackled Jeremy Moskowitz’s article on bringing AD accounts back once they’ve been lost. It’s when we were looking over the polished results that we realized something was amiss.

Both writers allude to the approach AD uses to keep track of the currency of its objects. AD uses an approach involving a combination of USNs (update sequence numbers) and PVNs (property version numbers). The two stories we had on hand appeared to contradict each other in how these two types of identifiers work.

In an effort to reach agreement, we questioned both writers, who dug in their technologist heels. Jeremy’s explanation: “USNs are to check out for when ‘things are in sync,’ and PVNs check out ‘what needs to be updated.’” Curt concluded: “What [sources] I’ve found always seem to talk about USNs in terms of replication and PVNs in terms of collision resolution.” Well, that certainly cleared things up.

Then they upped the ante. Jeremy quoted his technical mentor, Bill Boswell, esteemed author of Inside Windows 2000 Server from New Riders. Curt cited the Active Directory Technical Reference from Microsoft Press.

So we editors did as we’ve always done: We kept asking questions — in this case, of Microsoft. We turned to Waggener Edstrom, Microsoft’s primary PR company. Lisa answered the phone. That’s when I whispered those three little words: “USN or PVN?”

The publicist kicked into action. We exchanged daily e-mail for at least a week as she attempted to track down the final authority on AD. Apparently, that person must be at work on Blackcomb’s directory structure behind a locked door. The e-mail updates petered out. We were back where we started.

By now, it was just a matter of a few days before the issue went to the printer. What to do? Hold the stories until we could be absolutely positive of their accuracy? That would leave a few blank pages. Publish them as they were written and hope nobody noticed? Once a reputation for technical veracity is lost, it’s tough to win back. Besides, if ever I’ve seen an eagle-eyed readership, you’re it. You’d notice.

That’s when Mr. Boswell came to our rescue. He suggested that he write a sidebar explaining these two concepts. This offer arrived — like the cavalry — at the last possible minute, and I know he worked evenings and a weekend to pull it together in time. You’ll find Bill’s article in our online edition of this issue. We’ve left the articles by Curt and Jeremy intact. Do both authors have it right? Is anybody wrong? Weigh in at [email protected].