Architecture of the library
The purpose of this page is to present briefly the global architecture of the c-cote library. This can be useful to understand the implementation, use it to port the library to another language...
Cote is based on discover and axon architectures.
The following schema illustrates how the library is implemented.
The [ xxx ] annotations mean the link is performed for the selected cote configuration only: PUB for publisher, SUB for subscriber, REQ for requester, REP for replier and MON for monitor.
The cote part contains the exposed functions of the API to start/send/receive data. The purpose of this layer is to start axon and discover instances. Particularly, the discover instance is started immediately for SUB, REQ and MON configurations, and started after port bound by the axon instance for PUB and REP configurations.
The discover library used to perform discovery of the other nodes.
The axon library used to perform communication between nodes.