Blame the last man standing?

		Print
		Email weblog link
		Discuss
		Blog this

Tim O'Reilly
Jul. 14, 2003 05:46 PM
Permalink

I received the following message from Stuart Cheshire, architect of Apple's Rendezvous technology.

We frequently see the following pattern:

(a) someone's network is broken
(b) they look with a packet sniffer
(c) all they see is Rendezvous traffic
(d) they conclude that Rendezvous broke their network

The whole point of Rendezvous is that it is the networking that keeps working even when everything else is broken. Of course when your network is broken, and everything else stops working, all that's left is the Rendezvous traffic. Does that mean that Rendezvous broke your network? When all the world's oil reserves are exhausted, and all that's left running on our roads are the electric cars, does that mean that the electric cars caused the world's oil reserves to run out? No, I don't think so. Without actually being there, I can't tell what caused the problem in this case. Maybe it truly was caused by Rendezvous, or maybe it wasn't. We will probably never know now. The point is that just because Rendezvous is the only thing that survives your network outage, doesn't mean that Rendezvous caused your network outage.

See http://oscon.kwiki.org/index.cgi?WirelessProblems

My advice to Stuart was to put this idea out there for discussion. Like Stuart, I have no opinion on the cause of the problems with the network. However, I loved the idea he articulates here, that it's easy to blame the last man standing. A new version of "blame the messenger."

Tim O'Reilly is the founder and CEO of O'Reilly Media, Inc., thought by many to be the best computer book publisher in the world. O'Reilly Media also hosts conferences on technology topics, including the Web 2.0 Summit, the Web 2.0 Expo, the O'Reilly Open Source Convention, and the O'Reilly Emerging Technology Conference. Tim's blog, the O'Reilly Radar, "watches the alpha geeks" to determine emerging technology trends, and serves as a platform for advocacy about issues of importance to the technical community. Tim is an activist for open source and open standards, and an opponent of software patents and other incursions of new intellectual property laws into the public domain. Tim's long-term vision for his company is to change the world by spreading the knowledge of innovators. For everything Tim, see tim.oreilly.com.

">
Have you seen anything like this? I'm interested both in technical comments about the behavior of Rendezvous, and also social comments about other situations where you might have seen "blame the last man standing."

You must be logged in to the O'Reilly Network to post a comment.

Showing messages 1 through 8 of 8.

Inferior Rendezvous implementation?
2003-07-15 20:12:09 anonymous2 [Reply | View]

Given the tight coupling between Apple's hardware and its software perhaps some faulty non-OSX implementation of Rendezvous and/or a Rendezvous-based application caused the problem?

Just a possibility....

Rendezvous ill accused probably.
2003-07-15 00:19:35 joccam [Reply | View]

First of all, there were many powerbooks and ibooks at OSCon. It was great to see. I heard anecdotally this was the first year with such a trend. I'd guess upwards of half the portables were powerbook (plus a few ibooks). Amazing. So, there were plenty to go around.

However at Apple's WWDC the week before, there were probably twice as many people who virtually all had powerbooks or ibooks. There were network problems initially, but by the end everything was working well (when you could get an IP address from DHCP). IOW, no rendezvous issues in a much heavier wireless network situation.

Based on this (unscientfic) perspective, I'd guess Rendezvous had little if anything to do with the network outages. I noted that to my session neighbors during the break when the announcement was made, and saw nothing which cast Rendezvous in doubt.

Eventually they got the network working, and it wasn't immediately after the announcement (when everyone would have killed Rendezvous).

2c,

= Joe =

Rendezvous ill accused probably.
2003-07-15 17:35:59 anonymous2 [Reply | View]

"There were network problems initially, but by the end everything was working well (when you could get an IP address from DHCP). IOW, no rendezvous issues in a much heavier wireless network situation."

I was also at WWDC. One time I enabled Rendezvous in iChat, and the next thing I knew, the "mDNSResponder" Multicast DNS responder process was eating my PowerBook G4 17" CPU alive. 75% or thereabouts of the CPU was being eaten by this thing. I hate to say it, Stuart, but even if Rendezvous wasn't the cause of the problem with the network(s), it sure was a problem on my CPU. I immediately turned it off in iChat and things returned to normal.

Rendezvous
2003-07-14 19:17:27 stevegillmor [Reply | View]

The vast number of Macs at the conference and the widespread viral use of Hydra may well have produced the effect. Either way, it is a testament to the Mac conference connection and the growth of Rendezvous as a core protocol of the RSS/Hydra/WiFi virtuous circle.

Rendezvous
2003-07-15 10:09:28 anonymous2 [Reply | View]

I took Hydra based notes throughout the conference, but frankly I only saw three or five others at any given time doing the same. Now there were a huge number of folks attached to iChat, and I saw a decent number connecting up to either weblogs in realtime, IRC, or some other recording mechanism that I was reasonable sure wasn't Hydra.

Rendezvous
2003-07-14 20:40:33 Tim O'Reilly | [Reply | View]

Actually, there wasn't much use of hydra at OScon. Nothing like we saw at ETcon. In fact, a number of people seemed to start up Hydra documents, and then give them up because no one joined in.

Of course, I checked in very sporadically, since I was in the middle of a lot of things, and didn't have much time to look at what was happening in the surrounding ether.

Rendezvous
2003-07-14 21:24:33 stevegillmor [Reply | View]

Judging by the lack of posted results, you may be right. But my sense of the experimentation around Hydra and iChat AV in WiFi conference settings is that the explosive adoption is taxing the network. It would be useful to know what is causing the outage more from an allocation perspective.

Rendezvous
2003-07-17 20:54:46 ziggy-at-panix [Reply | View]

Actually Hydra and iChat weren't that heavily used at OSCon. Lots of people wanted to try them out in a conference setting, but I wouldn't call the use of either "explosive" or even "widespread".

Sure, there was a huge Mac OS X presence, but many people didn't want to run iChat (even after being prodded), and a lot of people were frustrated with the networking situation.

Plus, iChat was acting funny. On many occasions, I could see about 7 people logged in via Rendezvous, but the person sitting next to me could see dozens. I couldn't send messages to him, but he could send messages to me that I could respond to.

Showing messages 1 through 8 of 8.

Return to weblogs.oreilly.com.

Weblog authors are solely responsible for the content and accuracy of their weblogs, including opinions they express, and O'Reilly Media, Inc., disclaims any and all liabililty for that content, its accuracy, and opinions it may contain.

This work is licensed under a Creative Commons License.