Towards Reliable Integrated Services for Dependable Systems. Henrik Schi ler

Reliability issues for various technical systems are discussed and focus is directed towards distributed systems, where communication facilities are vital to maintain system functionality. Reliability in communication subsystems is considered as a resource to be shared among a number of logical connections and a reliability management framework is suggested. We suggest a network layer level reliability management protocol RRSVP (Reliability Resource Reservation Protocol) as a counterpart of the RSVP for bandwidth and time resource management. Active and passive standby is discussed as well as utilization of passive standby redundancy by background applications residing on alternative routes. Details are provided for the operation of RRSVP based on reliability slack calculus. Conclusions summarize the considerations and give directions for future research.