TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Elle: Inferring Isolation Anomalies from Experimental Observations

43 pointsby blopeurabout 5 years ago

3 comments

aphyrabout 5 years ago
Oh! I didn&#x27;t expect this to show up here! Happy to answer questions, if folks have any.<p>While this is on frontpage, I&#x27;d like to ask: if anyone has experience with machine proofs, I&#x27;d appreciate your help in completing a formal proof of Elle&#x27;s correctness. I&#x27;ve been teaching myself Isabelle&#x2F;HOL in an attempt to formalize the proof sketch, and I <i>have</i> encoded most of the formalism and properties I care about, but actually proving lemmata has been... frustrating. Like, I burned hours one one lemma because it hinged on showing \forall x :: Nat, x \in N.<p><a href="https:&#x2F;&#x2F;github.com&#x2F;jepsen-io&#x2F;elle&#x2F;tree&#x2F;master&#x2F;proof" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;jepsen-io&#x2F;elle&#x2F;tree&#x2F;master&#x2F;proof</a><p>If you&#x27;d like to work on this, my email is aphyr@jepsen.io!
blopeurabout 5 years ago
Github code : <a href="https:&#x2F;&#x2F;github.com&#x2F;jepsen-io&#x2F;elle" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;jepsen-io&#x2F;elle</a>
mjbabout 5 years ago
This is really cool work, very exciting to see if published in paper form (I saw your talk about some of this last year)!<p>I had a question about your completeness argument (in 4.3.1), you say &quot;we typically observe enough of a history to detect the presence of non-cycle anomalies&quot;. I think I understood why that is after reading the rest of the paper, but I didn&#x27;t understand the worst case. Is it possible, for example, for a database to intentionally construct incorrect histories that Elle doesn&#x27;t detect?<p>I don&#x27;t have any real &quot;evil database&quot; threat model in mind, just trying to test my understanding.
评论 #22699353 未加载
评论 #22696669 未加载