Telecommunication networks: protocols, modeling and analysis
Telecommunication networks: protocols, modeling and analysis
Two classes of performance bounds for closed queueing networks
Performance Evaluation
Models and algorithms for the design of store-and-forward communication networks
Models and algorithms for the design of store-and-forward communication networks
Mean-Value Analysis of Closed Multichain Queuing Networks
Journal of the ACM (JACM)
Optimal routing in closed queuing networks
ACM Transactions on Computer Systems (TOCS)
A tree convolution algorithm for the solution of queueing networks
Communications of the ACM
Modeling and analysis of flow controlled packet switching networks
SIGCOMM '81 Proceedings of the seventh symposium on Data communications
Pamma Noniterative Approximate Solution Method for Closed MultichainQueueing Networks
Pamma Noniterative Approximate Solution Method for Closed MultichainQueueing Networks
Modelling and analysis of flow-controlled computer communication networks
Modelling and analysis of flow-controlled computer communication networks
Communication nets; stochastic message flow and delay
Communication nets; stochastic message flow and delay
Research: Performance of end-to-end flow control in LAN/WAN interconnection
Computer Communications
Hi-index | 0.00 |
Closed queueing networks have been advocated by several authors to be a more desirable model than open queueing networks (Kleinrock's model) for network design. We compare open and closed network models and demonstrate the accuracy of a particular closed network model with experimental results. The proportional approximation method (PAM) is presented for evaluating performance measures of closed queueing networks. PAM algorithms have computational time and space requirements of O(KM), where M denotes the number of queues and K denotes the number of virtual channels in the network. Thus, PAM is the first (and only) method that can be used for solving industrial-strength network design problems using a closed network model.We formulate the following optimal routing problem: Find a route for a new virtual channel to be added to a network with existing flow-controlled virtual channels. A fast heuristic algorithm is presented. The algorithm uses PAM and exploits the following empirical observation: The route that maximizes the individual throughput of a virtual channel coincides in most cases with the route that maximizes the total network throughput (this is not true in general). We present statistical results from studies of 100 randomly generated networks to demonstrate the accuracy of PAM algorithms and the effectiveness of the optimal routing algorithm. (Exact solutions obtained by the tree convolution algorithm were used as benchmarks in our statistical studies.)