I will cop to still being somewhat unclear on the precise definition of linearizability, even after reading aphyr’s original Jepsen articles, the wiki article, watching this, etc. Does anyone have a good resource they would recommend? Is the original paper very readable?
Here’s the best one I tell people: “Each operation appears to take effect atomically between its invocation and its conclusion.”
The original paper’s not bad if you’re used to reading papers. I recommend reading only the first 3-4 pages if you are looking for the definition of linearizability.
I will cop to still being somewhat unclear on the precise definition of linearizability, even after reading aphyr’s original Jepsen articles, the wiki article, watching this, etc. Does anyone have a good resource they would recommend? Is the original paper very readable?
Here’s the best one I tell people: “Each operation appears to take effect atomically between its invocation and its conclusion.”
The original paper’s not bad if you’re used to reading papers. I recommend reading only the first 3-4 pages if you are looking for the definition of linearizability.
What’s the paper you’re referring to?
Fortunately computer science preprints are relatively easy to find on the web, rescrv is probably referring to either the first link or the followup:
http://www.cs.cmu.edu/%7Ewing/publications/HerlihyWing87a.pdf http://cs.brown.edu/~mph/HerlihyW90/p463-herlihy.pdf