From e7c322d956e14560c8eee9f28f2a2b81273703a3 Mon Sep 17 00:00:00 2001 From: martinsumner Date: Mon, 10 Jul 2017 09:49:20 +0100 Subject: [PATCH] Clarify pointer to research --- docs/ANTI_ENTROPY.md | 1 + 1 file changed, 1 insertion(+) diff --git a/docs/ANTI_ENTROPY.md b/docs/ANTI_ENTROPY.md index db57ef1..2a65951 100644 --- a/docs/ANTI_ENTROPY.md +++ b/docs/ANTI_ENTROPY.md @@ -179,6 +179,7 @@ Even with this mitigation, the volume of false positives and negatives needs to The approach considered for Leveled has been to modify the Merkle trees used to ease their implementation, as well as specifically separating out anti-entropy for recent changes as a different problem to long-term anti-entropy of global state. +[Recent research](http://haslab.uminho.pt/tome/files/global_logical_clocks.pdf) has been released which examines using dotted version vectors at a node level to improve the efficiency of managing key-level consistency, reducing the risks associated with deletes (without permanent tombstones), but also provide an inherently integrated approach to active anti-entropy. The Global Logical Clock approach does assume that durability is not mutable: