tag:blogger.com,1999:blog-4490444256563865262.post876858846196423421..comments2023-12-26T06:47:22.192+01:00Comments on This long run: You Do It Too: Forfeiting Network Partition Tolerance in Distributed SystemsNicolas Liochonhttp://www.blogger.com/profile/07943925485349697034noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-4490444256563865262.post-29764305016374897652015-07-10T12:30:52.353+02:002015-07-10T12:30:52.353+02:00I updated the post to make clear that the partitio...I updated the post to make clear that the partitions mentionned here are network partitions. That was clearly confusing. Also, there was no discussion around CP vs. AP. The discussion is around CA/CP. I changed a sentence that was unclear. And I added a few details here and there.<br /><br />CP and CA overlap because CA does not mean that partition are impossible. Brewer: "It is best to think about this probabilistically: choosing CA should mean that the probability of a partition is far less than that of other systemic failures."<br /><br />For the data itself, we need the probability of a partition. The paper pointed in the response does a pretty good job at showing that networks are not reliable, but not at estimating probabilities.<br /><br />And the same paper confirms you can't generalize anyway: "On the other hand, some networks really are reliable. Engineers at major financial firms have anecdotally reported that despite putting serious effort into designing systems that gracefully tolerate partitions, their networks rarely, if ever, exhibit partition behavior."<br /><br />I stick to my conclusion: “the whole space is useful.”<br />Nicolas Liochonhttps://www.blogger.com/profile/07943925485349697034noreply@blogger.comtag:blogger.com,1999:blog-4490444256563865262.post-59022637416056250592015-07-08T23:34:20.870+02:002015-07-08T23:34:20.870+02:00Curious as to what you think of this response: htt...Curious as to what you think of this response: https://aphyr.com/posts/325-comments-on-you-do-it-tooAnonymousnoreply@blogger.com