Speaking with Spanner engineers, sometimes they mention that they feel confident that they can get their epsilon down to 2ms with regularity. I beg them to release a paper about the engineering story for getting it as low as they have been able to. I think geo-distributed clock skew minimization is something that will have fascinating applications outside of google within a few years.
I wonder if the reason distributed systems are “surprising” for western engineers is because of the western conception of time. Do eastern engineers have any problems dealing with the uncertainty that distributed systems pose?
Speaking with Spanner engineers, sometimes they mention that they feel confident that they can get their epsilon down to 2ms with regularity. I beg them to release a paper about the engineering story for getting it as low as they have been able to. I think geo-distributed clock skew minimization is something that will have fascinating applications outside of google within a few years.
I wonder if the reason distributed systems are “surprising” for western engineers is because of the western conception of time. Do eastern engineers have any problems dealing with the uncertainty that distributed systems pose?
Eastern(+ greek) notion of time circular. As it’s a very big circle, it’s treated as a straight line for most practical purposes.