Optimization of DDS for Peer To Peer Communication over a Constrained Network

Abstract

The Object Management Group (OMG) standardized Data Distribution Services (DDS) is considered to be one of the best middlewares with maximum number of the Quality of Service (QoS) parameters. DDS works on Publish-Subscribe paradigm utilizing different topics (information about single data) to provide specified services to systems in a distributed environment. The number of topics increases with the number of services. These topics come with their respective overheads which consume the available bandwidth on per topic basis within a network. It becomes more severe in the case of communication over constrained networks. In case of communication over constrained networks, where resources are limited characteristically, the Peer To Peer (P2P) paradigm is chosen. Generally, the choice of communication paradigm of publish-subscribe and P2P are mutually exclusive for any schema. We propose to arrive at an engineering compromise to have the benefits of DDS and P2P. In this paper the design towards hybrid objective of having P2P over DDS is presented. The design caters for synchronous P2P service invocation primitive over DDS. The DDS communication has been optimized for the P2P communication among multiple services while limiting overheads of topics on constrained network. In order to achieve the desired objective we have designed an application logic which is akin to Multiplexing-Demultiplexing.