We present schemes for disseminating information in the n-dimensional hypercube with some faulty nodes/edges. If each processor can send a message to t neighbors at each round, and if the number of faulty nodes/edges is k(kn), then this scheme will broadcast information from any source to all destinations within any consecutive n+[(k+l)/t] rounds. We also discuss the case where the number of faulty nodes is not less than n.

We describe two information disseminating schemes, t-disseminate and t-Rdisseminate in a computer network with N processors, where each processor can send a message to t-directions at each round. If no processors have failed, these schemes are time optimal. When at most t processors have failed, for t1 and t2 any of these schemes can broadcast information within any consecutive logt+1N2 rounds, and for an arbitrary t they can broadcast information within any consecutive logt+1N3 rounds.

We discuss fault tolerance of an information disseminating scheme, t-disseminate on a network with N processors, where each processor can send a message to t directions at each round. When N is a power of t+1 and at most tlogt+1N-1 (at most t) processors and/or edges have hailed, logt+1N+(f1)/t rounds (logt+1N+2 rounds) suffice for broadcasting information to all destinations from any source by t-disseminate. For a arbitrary N, logt+1N2f/t1 rounds (logt+1N+2 rounds) suffice for broadcasting information to all destinations from any source by t-disseminate if at most t(logt+1N1)/2 (at most t/2) processors and/or edges have failed.

A novel cell-level FEC (forward error correction) scheme at SSCS of AAL type 5 for error-free data transmission services in ATM networks is proposed and evaluated. In the proposed cell-level FEC scheme, both the length of user data (e. g. , IP packet) and attached redundant data can be modified based on sender's local decision without any end-to-end parameter re-negotiation procedure. The writing and reading order regarding an interleave matrix for cell-level FEC algorithm are the same, in order to perform pipelining data transmission. The prototype implementation with software processing achieves few Mbps end-to-end throughput. The end-to-end data transmission latency and the amount of retransmission packets due to packet error at the receiver entity are evaluated by computer simulation with correlated cell dropping process. Simulation results show that the benefits of cell-level FEC scheme for the error-free data transmission services, i. e. , by use of cell-level FEC scheme, the amount of retransmitted packets can be reduced, even if the average latency for end-to-end data transmission increases slightly.

This paper discusses and evaluate a high speed multimedia communications in ATM networks. Communication platforms below AAL layer in ATM networks can provide fairly high speed communication channel for ATM-UNI. On the other hand, the conventional data processing architecture in the workstations can not provide sufficient throughput compared to the throughput of ATM networks at this time. Also, regarding error-free peer-to-peer data communications, conventional transport protocol, e.g. TCP, can not provide small latent communication, when the product of communication bandwidth and propagation delay is large. This paper focuses on protocol architecture, especially on error control policy at SSCS in AAL layer, and on performance evaluation of high speed multimedia communication based on the current data processing architecture.