Most of you who read my writings know that I don't often write about hardware. And in fact, I'm not writing about it today. I'm only pointing out a post by one of the most famous hardware bloggers, Linchi Shea. His recent post refers to a paper written by Goetz Graefe, which describes current research and re-evaluation of the 5-minute rule.
I found this post very intriguing for several reasons.
1. The 5-minute rule is one of the few hardware oriented rules that I discuss in my classes so it's great to see an update to the famous paper.
2. The original 5-minute rule was first proposed by Jim Gray, who wrote the forward to all of my books except for the most recent, and I blogged about Jim's disappearance here. My latest book was dedicated to Jim.
3. Goetz Graefe was one of my favorite developers on the SQL team at Microsoft, and I didn't realize he had left MS and was now working for HP labs. I only found out when I looked at the paper that Linchi posted. Since he is gone, I can now mention that he was the developer I was referring to in my post on query parallelism, when I mentioned:
So instead of saying 'We are going to be able to PARALLELIZE your queries', he would say 'We are going to be able to PARALYZE your queries'
So take a look at Linchi's post, and both the original and revised papers, as a good way to start your technical explorations this new year.