Intellectual scaling in a distributed cloud application architecture: A message classification algorithm

摘要

The service systems deployed in a cloud environment have a volatile business logic which cannot be fully described by messaging patterns now available. Firstly, the work of distributed service systems is heavily influenced by the type, format, topic, and content of messages. Secondly, the existing middleware solutions for message exchange have a very narrow functionality focused exclusively on message passing between the nodes of a distributed system. They pay no attention to the features of the messages sent. However, scaling in hybrid and geo-distributed cloud solutions requires a more comprehensive approach. So does the analysis of the state of system modules. This paper offers an algorithm allowing to classify messages which are sent via the same protocol but have various content and do not correspond to a pre-set format. We provide an example of using such algorithm in middleware architecture. The major advantages and drawbacks of the algorithm are described with regard to a distributed cloud system.