Evolution of architecture in a geo-distributed system: what Cassandra has to do with it

Mission-critical systems sometimes require a particular level of fault tolerance, a ‘cluster of clusters’. Organizing the correct operation of this kind of solution means tackling a certain class of problems, which do not come up when handling a single data processing center. We will do a case study of fault-tolerant architecture, prerequisites for changing it, and decisions made to rise to the challenges.

This talk may be of interest to those wishing to find out more about the challenges of building geo-distributed systems, as well as those having already designed or developed similar solutions and wishing to compare experiences.