Skip to Main Content
Existing congestion control and network management techniques have been designed to cope with problems such as unexpectedly high offered load and hardware failures. On rare occasions, networks have failed despite the deployment of such techniques. The authors survey the current research in congestion control and network management techniques and their role in ensuring network stability in the presence of software errors. What situations are these techniques intended to address? What are the characteristics of the problems that have led to network collapse? What areas of research offer the best prospect for further improving network reliability?