On Wed, Sep 14, 2016 at 1:49 PM, jerome <jeromefroel...@hotmail.com> wrote:
> I was curious if anyone had any kind of statistics or ballpark figures on
> how long it takes information to propagate through a cluster with Gossip?
> I'm particularly interested in how fast information about the liveness of a
> node spreads. For example, in an n-node cluster the median amount of time it
> takes for all nodes to learn that a node went down is f(n) seconds. Is a
> minute a reasonable upper bound for most clusters? Too high, too low?
Dahlia Malkhi gave a talk on gossip protocols at the Papers We Love
conference last Thursday (http://pwlconf.org/dahlia-malkhi/), and she
answered this better than I ever could. The video of her presentation
hasn't been posted yet, I'm told it should be as early as later today
though. You can look for at on the Papers We Love Youtube channel
(https://www.youtube.com/channel/UCoj4eQh_dZR37lL78ymC6XA), and it'll
be announced on the website (http://paperswelove.org/).