SIGCOMM '88 Symposium proceedings on Communications architectures and protocols
Congestion avoidance and control
SIGCOMM '88 Symposium proceedings on Communications architectures and protocols
Dynamic adaptive windows for high speed data networks: theory and simulations
SIGCOMM '90 Proceedings of the ACM symposium on Communications architectures & protocols
Performance analysis of a feedback congestion control policy under non-negligible propagation delay
SIGCOMM '91 Proceedings of the conference on Communications architecture & protocols
A hop by hop rate-based congestion control scheme
SIGCOMM '92 Conference proceedings on Communications architectures & protocols
Analysis of a rate-based feedback control strategy for long haul data transport
Performance Evaluation - Special issue on performance modeling of high speed telecommunication systems
Interaction among virtual circuits using predictive congestion control
Computer Networks and ISDN Systems - Special issue on high speed networks
Feedback control of congestion in packet switching networks: the case of a single congested node
IEEE/ACM Transactions on Networking (TON)
Pathwise Optimum Policies for ATM Cell Scheduling and Rejection
Pathwise Optimum Policies for ATM Cell Scheduling and Rejection
On Characterizing Optimal Buffer Control Policies in ATM Nodes
On Characterizing Optimal Buffer Control Policies in ATM Nodes
| Hi-index | 0.00 |
Limited buffers in ATM network elements will inevitably result in cell loss. We model an ATM network as a tandem network with finite buffers and investigate the performance of a cell-level control policy (the HILO heuristic) that is designed to minimize cell loss rates. The HILO heuristic is based on limited feedback about the cells awaiting transmission in adjacent network nodes. We demonstrate via analysis and simulation that this scheme can reduce cell loss by substantial amounts over the simple FCFS scheme. The parameters of the control policy can be modified adaptively to maximize the reduction in loss rates. The heuristic is robust in the sense that it is relatively insensitive to values of the control parameters, and is effective for different traffic loads and arrival patterns.