Revisiting the Gamma paper for my grad DB class. Struck by how ambitious the project was and how professional the work still feels 30+ years later.
Prikaži ovu nit
They innovated on redundant disk failure layout analysis before RAID. They did parallel dataflow deacdes before MapReduce. They wrote their own OS. They implemented a version of ARIES. They built a competent query optimizer.
-
-
Speedup/Scaleup results look good, blemishes admitted and explained. Great model for ambitious, long-term academic system building. https://scholar.google.com/scholar?cluster=8912521541627865753 …
Prikaži ovu nit -
I understand why the MapReduce and Big Data hype was frustrating for DeWitt. Students: if you're looking for 20th century systems work that doesn't read like archaeology, read the Gamma paper.
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
-
-
Correction: The RAID paper is SIGMOD 1988. The Chained Declustering paper TR is 1989, and cites the RAID paper and lays out differences. Apologies to all the fine researchers on both sides who were doing their scholarly jobs!
Prikaži ovu nit -
What kind of jerk tweets out easily disproven falsehoods late in the night? *sigh*
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.