Network layer based redundancy for time-critical VoIP applications
ABSTRACT Voice over Internet protocol (VoIP) applications are making their way into safety-critical environments. If voice data has to be transmitted in such environments, the connection must not fail and the network has to recover quickly. Additionally, the use of specialized systems that are tailored to a certain application is not a preferred solution. Instead, the use of commercial-off-the-shelf (COTS) hard- and software is required. This paper discusses the use of standardized hard- and software as well as network protocols to allow failover times within the range of milliseconds. A redundant end-system and its network coupling are presented. The Layer 3 protocol OSPF and the combination of OSPF with the protocol-independent Hello protocol BFD are investigated to fulfill the requirements. The tests and the achieved results show a significant reduction of the convergence time.
Conference Paper: Reliable data distribution in voice communication environments[Show abstract] [Hide abstract]
ABSTRACT: Due to lowering equipment and maintenance costs as well as optimization towards fast failure detection and convergence time, Ethernet based technologies have found their way into safety critical communication systems. This article focuses on message-oriented transfer of data packets matching ordering, reliability, and group membership criteria. Messaging systems, available under open source licenses, are discussed with respect to a specified use case; two of them - Corosync and Spread Toolkit - are analyzed regarding their scalability within a loss-free Local Area Network. A functionality analysis in case of network partitions is out of focus for this work. The results show slight advantages of Spread over Corosync for the defined use case. Even though, Corosync enables sending significantly higher message rates for a low number of recipients, Spread catches up and succeeds in case these numbers increase.AFRICON, 2011; 10/2011