Towards Survivability of Application-Level Multicast
Gal Badishi, Idit Keidar, Roie Melamed · 2004
This position paper focuses on challenges in providing survivable and scalable multi-point to multi-point reliable application-level multicast systems (ALMs) for very large groups in wide-area networks. A protocol deployed in such settings must be able to withstand frequent node failures as well as non-negligible message loss rates. A survivable system should also cope with uncooperative users. Moreover, in typical wide-scale multicast sessions, users frequently join and leave. Rapid joining and leaving, also called churn, may effectively cause denial of service (DoS) if handling joins and leaves induces high overhead. Survivability also mandates withstanding attacks. One of the most devastating security threats faced by a distributed system is a DoS attack. Coping with DoS attacks is essential when deploying services in a hostile environment such as the Internet; in 2003, approximately 42% of U.S. organizations were faced with DoS attacks.