These are the old pages from the weblog as they were published at Cornell. Visit www.allthingsdistributed.com for up-to-date entries.

June 03, 2004

Reruns

If you do not have time for new content you can of course do what television network do; have some reruns. Here are five of the most popular 'All Things Distributed' episodes of the past year.

  • Web Services are not Distributed Objects. This appears to become an all time favorite. Initially this was something Steve Vinsoki tricked me into writing one late night while in Budapest at WWW2004. But it grew out into a back-to-principles writing, which resonates particularly well with practitioners. It even has gotten me a few unexpected speaking engagements
  • Comparing CLR, Mono, SSCLI and Java Performance Part I. This was the first posting of five on the low-level performance comparison of the different VM's. It also was my first attempt to show how some simple research can grow over time into a paper. This work became really interesting when I figured out how to get the x86 out of the JIT and compare it. Other postings: II, III, IV, Method Size and CLR Performance and Floating Point Error.
  • HPTS Day II - How are Databases used at Big Customers. The High Performance Transaction Processing workshop was one of the best conferences I have been in years. Especially interesting were the details of all the big e-commerce shops and how they had to avoid to use databases for anything else but a persistent store to make them scalable. Everything, even referential integrity checking goes into the middleware.
  • On the Scalability of Feeds and Aggregators. My thoughts about how the whole weblogs feeds system is not likely to scale. Not at the human level, as reading feeds is a limited abstraction, and not at the infrastructure level, as the pull approach is limited in scalability when used in a closed feedback loop.
  • Putting a Value to a Conference. This was written in frustration about the inability to make a link between the registration fee and the quality of conferences (of the non-academic kind). So I decided to throw some analytical skills at it and come up with an approach to evaluate it. Unexpectedly it generated a lot of buzz, as it seems a lot of people are wrestling with the same issue. Of course it also helped that Chris Pirillio put it as a mysterious link on the FAQ of Geeks Gone Wild (formerly know as Gnomedex).
Posted by Werner Vogels at June 3, 2004 04:24 PM
TrackBacks

Comments