|AustinTek LVS page||| LVS homepage | AustinTek homepage ||
OLS this year was moved forward one month from July to Jun. The people at the Youth Hostel told us to expect slightly cooler, wetter weather. The weather was quite fine by me (temps in low 20s, C), but as is routine for summer in North America, the conference rooms were air conditioned down to temperatures that no-one would put up with in winter. Ottawa spends half of the year below freezing and it never gets hot in summer. It would have been fine with me if they'd pumped in the air from outside. As it was, I had to go outside every now and again to warm up.
Lars, Ratz and I came for the annual LVS meet again. Ratz and I stayed in the solidly built jail, while Lars is stayed in a flimsy modern hotel, which I won't be standing in its current condition in 150 yrs (like our jail).
Ratz had arrived a few days earlier in Canada, to visit his lovely (now sadly ex-) girlfriend and a friend in Waterloo and appeared exhausted in Ottawa, late the evening before the conference. For the first time in 3 conferences Ratz got to sleep before midnight.
Since last year, Ratz has figured some things out, namely that he needs paper credentials. Sooner or later he will meet a potential employer, who will not be able to appreciate the skills acquired in 10yrs of writing kernel code and implementing network security as a valued member of a team working for a world class bank, but who will instantly recognise the tenacity of a person who survived 3yrs of classes taught by people who know that you need their piece of paper to get a job. Ratz visited the people at Waterloo U. to find out about their courses. (Waterloo U. is contemplating a 1 year compulsory course in M$ C# for engineers). Part of the coursework is requirement for a year's work experience. Did Ratz's work experience count? Nooooo, it had to be experience which you gained while paying them to be enrolled. Waterloo is no longer on Ratz's short list.
Lars gave his talk early the first morning on the Open Clustering Framework and was relieved to have his responsibilities discharged so early in the conference.
Here we are; L-R Lars, Ratz, Joe, and Harald Welte, in a somewhat sombre mood apparently. I'm wearing (on my jacket) one of the few remaining LVS buttons from OLS 2000 (they're a collector's item already).
Harald is an honourary member of LVS by virtue of his role in netfilter. Note Harald's white laptop - it's a Mac, one of the few I see at OLS. Harald gave a talk on plans for saving state in netfilter. This is needed for failover of machines with iptables rules e.g. firewalls, LVS directors. When these machines fail-out, all of the tcp state information has to be transfered to another machine without interruption of service. This is not an easy problem. Harald also discussed the duplication of code that has occured for all the commands that tag or manipulate packets. Each filtering command should be able to call a library which tags or inspects packets. Harald hopes to write a libiptables.
Although difficult to tell from the photo above, Lars and Ratz are sporting the latest geek-chic fashion statement, a short haircut (<1cm, sticking out radially from your head). About half of the participants of OLS seemed to have this haircut. Here's a better view of the Linux haircut (modelled by Ratz). This haircut is best suited to people who don't see a whole lot of sun.
By contrast here's the standard issue 8 bit grey scale haircut seen on a US defense contractor (modelled by Joe, who is running pine). This haircut is descended from the Jimmy Carter haircut (Carter was the first US president with hair long enough to cover his ears).
Ottawa this year had a distinctly unfriendly air, in contrast to the atmosphere which has helped bring me back each year. The usually tranquil Canada was girding its loins for the onslaught of peaceful protests about G8 conference being held the same week (25-27 Jun) in a wilderness area in western Canada, several 1000km from Ottawa. This meeting is of the world's richest countries who gather occassionally to determine the rules other countries have to follow before the rich countries will trade with them. Despite the remoteness of the G8 meeting, downtown Ottawa was swarming with riot police, equiped with helmets and transparent shields ready to do battle with people in tie-died shirts.
I didn't see any protesters the whole time I was in Ottawa, even though OLS was in the centre of downtown and only a block from the Parliament house. To be ready for the ravaging of the town, police were standing in groups at all intersections, and the town parks were closed off and patrolled by police talking to their wrists. Instead I went for a walk along the St. Lawrence R., and was told I to leave ("it's for your safety"). Even the pub where previously I'd sat outside, warming up and watching the people walk by, was closed off. The town was under occupation.
The Ottawa newspapers uncritically praised George Bush and applauded loudly when he made his first statement about the 6month old revelations about Enron. A photo of people protesting about G8, a dancing circle of people wearing tie-died shirts and holding hands, were described as "left wing fanatics". Fanatics fly planes into buildings, they don't dance in a circle, holding hands. There was only one column written by anyone showing critical thinking. The rest of the columns were written by bullies and sycophants.
Ottawa reminded me of Germany in the 1930s, as described by William Shirer. It was a thoroughly unpleasant place to be and the veneer of liberalism I'd seen previously in Canada, cracked severely when I talked to Canadians who didn't seem at all concerned about what I'd seen.
A strange incident that I thought nothing of at the time, had repercussions for one of the speakers. On entering Canada in Toronto, Immigrations asked why I was in Canada. On finding that I was at a conference, they asked if I was speaking or just attending. I thought this normally social question a little out of place, since is was asked with the demanding tone of one in authority. I couldn't see what it had to do with immigration. I said "just attending" and forgot about it. I kinda remember them asking the same question last year.
The first speaker of day 1, didn't make it because of "immigration problems". When asked the same question, he'd replied "speaker". On the assumption that he would be paid for his speaking, they asked him for his work permit, which of course he didn't have. He wasn't allowed into the country. He arrived late at night. OLS is not particularly well organised (presumably being run on a shoe-string and with minimal sponsorship) and I doubt if anyone at OLS would have been available late at night for a phone call to help in the situation. (I've had a lot of problem getting receipts for registration, for re-imbursement from work). Most attendees are from out of town and compared to the air-fares and other expenses involved, the registration fees for OLS are small. If it would help in situations like this, I'd be happy with a substantial increase in registration fees.
A major deficiency in the Linux world showed up the first afternoon, when I was looking for a suitable talk. The first session I tried, was introductory material that no-one at this conference needed to hear about. In the next room, a speaker confused the audience, who were there for information, with people who were there to be entertained by in-jokes. This is the style used in our local LUG, whose meetings I avoid. Speakers should give professional grade talks, remembering that the audience has devoted some time and money to be there and already knows a lot about computers. It's the speaker's job to deliver the maximum amount of data the audience can absorb, not to entertain. It wasn't till I tried a 3rd talk that I found a speaker who was delivering content. (The talks for the rest of the conference were better.)
Various Linux luminaries were present, and it amazes me that many of these people, whose names are known world wide, are in their early 20's. Marcelo (from Brazil), who's taken over from Linus to produce the 2.4.x kernels, is 19.
The first night, I left Ratz and Marcelo at a pub, where Ratz was helping Marcelo with the 2.4.19 kernel. Ratz was working his way through to the bottom of his personal pitcher of beer. (photo by Maddog, of Linux International).
I could see from the look in Ratz's eyes, that if it was humanly possible to get 2.4.19 out that night, he was going to do it. To help, I took his laptop and backpack with me, so that later that night, he wouldn't be overloaded with unneccessary responsibilities, while wending his way back to the jail. I also checked that Ratz had the key code written in ball-point pen on his hand. Ratz made it back at 4am, returning to his normal sleep cycle after having recovered from his preconference 4 day vaction.
Ratz has offered to be the official LVS ambassador. (He gets my vote.) Later, Lars and I discussed our impressions of the Ratz scheduler: it keeps functioning under conditions where other schedulers go to sleep for lack of resources, but the log files have large gaps in them. It doesn't remember what it was doing.
One of the talks next day was an update by Jeff Dike on user mode Linux (UML). The amazing this was that Linux could run a copy of itself as a job, without this being one of the original design criteria. The problem is that it doesn't do it well, e.g. all of the UML processes show up in the host kernel process table. Clearly only one process, the UML should show up in the host ps table, while the UML processes whould show up in the UML process table. Another problem is cooperative memory management. With only one kernel running, there is no need to swap out memory, till it is needed. You may as well leave it data in memory for a few weeks, in case it is ever needed, saving you an IO. With the host kernel and the UML kernels all taking all memory possible, there is going to be collisions for memory. The UML processes need to take minimum memory and the host needs someway of issuing an ultimatum to the UMLs that it will be needing memory soon and to either release memory or have it swapped out by the host kernel. (Note Sep 2002: UML is now in the 2.5 kernel tree.)
A talk by Wayne Meretsky of AMD promised 64 bit computing in a year, with the AMD hammer CPUs, at the same price as 32 bit compuing today. With the faster hyperlink hardware, the difference in access time for memory on another CPU on a multi-way SMP mobo, has been reduced from a factor of 10:1 to about 1.5:1, meaning that the memory attached to each CPU (your 4G of RAM) can be treated as shared memory.
At the reception, I sat next to Adam Richter of Ygddrasil. Ygddrasil only exists now to service requests from previous customers. Previously it was a viable company employing 8 people. He regrets that he didn't start looking for venture capital when it was available. However he realises that this could have set him up to fail in a larger and more spectacular way. I mentioned that it wasn't so long ago that the nerds were in charge of computing, that it needed someone techinally competent to set up a secure shopping cart, handle credit card data securely, track orders etc. Now that is all canned for you. He said we were all naive to think that the big companies who were handling retailing, wouldn't be able to handle on-line ordering or that customers who were familiar with the large companies would go elsewhere to order on-line. So the nerds and geeks had it for about 2yrs, now we're back to the big guys owning it all again.
A film about Linus/Linux was shown at the reception. Wensong and I feature prominently - we were in some footage shot at OLS_2000 - Wensong and I are just standing around. I think we get about 15 frames of fame. I didn't believe Ratz when he pointed it out, but he was expecting it and reran it for me from his laptop.
For some afternoon exercise, since it was unsafe to walk along the St. Lawrence R., I walked off in the other direction up the Rideau canal, and then back down the river (about 10km for each section). These photos are from Lars, who tackled the loop by bicycle. After walking up the Rideau canal, you come to a dam which forms a lake for the Ottawans to sail on.
The amount of water coming out is quite impressive. Here Lars was having trouble adjusting to the difference in latitude between Germany and Ottawa.
Here's looking downstream from the same spot.
Unfortunately from the sediment, color, smell and foaming, this is no mountain stream and I wondered what the Ganges must smell like as it flows past Benares.
About 10km downstream, the Rideau canal empties onto the St Lawrence R. at the Rideau Falls.
On the last evening at the party at the BlackThorn Cafe, one of the hot topics was preparations for the soccer world final which would be on TV at about 3am the next morning (due to the time difference between Japan and Ottawa). The only thing to do about this of course, was for someone to volunteer his $120/night room in one of the swank hotels, and for a group of people to go get refreshments. Fortunately this being Canada, all resources were within walking distance, so money was collected and in an hour or two, our band of heroes returned to the OLS party with the mission accomplished (it's easy to get a couple of cases of beer into a hotel, and have 30 people yelling at a TV set at 3am in your room, without anyone noticing).
It turns out that one of the people at the conference can juggle. Here's a few frames
Here's a single frame.