Distributed cron across the planet
Cron is used on every local machine out there – all systems have periodic tasks which have to be started at some time. But some tasks cover more than a single server, for example check another server...
View ArticleApache Bookkeeper or how NOT to design replication consistency scheme
Consistency in a distributed system is hard, it is even a limitation – everyone knows that one has to sacrifice a letter in CAP theorem to be have consistent distributed system: either it will not be...
View ArticleFly, baby, fly: initial PAXOS consensus daemon implementation commit
As promised (long-long-freaking-long ago) I started next project, and this is a distributed consensus algorithm written on top of PAXOS. I believe I started to understand its main feature, and thus...
View ArticleMulti-PAXOS consensus system implementation
We used to rely on management servers to handle our data. When it comes to huge amount of simple requests we can use NoSQL solutions, when we need more relations its time for more classic SQL...
View ArticleElliptics: datacenters, open functionality, changes, future
Our team has grown up a little and we started to make things faster. External projects First, our friends in Yandex open-sourced FCGI management daemon: fastcgi-daemon It allows a rapid deployment of...
View ArticleZab or Paxos
I really did not find a major difference between fast Paxos and Zab - it is the same message broadcast and quorum detection as long as there is only single master (leader) server. In elliptics...
View Article