Continued, previous part is HERE: https://www.artezio.com/pressroom/blog/flamenco-computational-cluster-part-1-introduction
Flamenco is an infrastructure for the distributed task performance. Any task that can be effectively solved in a distributive manner and compiled in JVM bytecode can be implemented in the form of the corresponding Flamenco plug-in.
The key features of Flamenco system are:
Flamenco has peer-to-peer topology (P2P) which practically stands for the full equality between cluster nodes. Each of them can receive and send out tasks that it gets from its client. Such an approach simplifies the configuration and increases the fault-tolerance due to the lack of a node which in case of the crash failure will lead to the system disorder. Nevertheless, it must be remembered that even within such organization of the interaction, there are a number of elements in the system the operating performance of which is very critical, among these are the network equipment (hubs, network switches, routers, etc.). The breakdown of a network element results in so-called network partitioning; the Flamenco cluster falls into several independent parts, and each of them can continue operating but with reduced productivity. After the recovery of the work of the network element, the parts of the cluster are again consolidated.
To be continued…