Circuit allocation and overload control in a hybrid switching system
Computer Networks and ISDN Systems
Dynamic hierarchical control of resource allocation in an integrated service broadband network
Computer Networks and ISDN Systems - Special issue: telecommunication services for developing economics
Queueing study of a 3-priority policy with distinct service strategies
IEEE/ACM Transactions on Networking (TON)
Optimal fixed frame multiplexing in integrated line- and packet-switched communication networks
IEEE Transactions on Information Theory
Asynchronous transfer mode or synchronous transfer mode or both
IEEE Communications Magazine
Control of multirate synchronous streams in hybrid TDM access networks
IEEE/ACM Transactions on Networking (TON)
The parallel optimization of network bandwidth allocation based on generalized particle model
Computer Networks: The International Journal of Computer and Telecommunications Networking
The parallel optimization of network bandwidth allocation based on generalized particle model
Computer Networks: The International Journal of Computer and Telecommunications Networking
Hi-index | 0.24 |
We consider a DQDB network carrying both isochronous and asynchronous traffic flows. Different classes of isochronous traffic are distinguished, according to their speed and originating user site, and their dynamics is modelled by continuous time Markov chains; the asynchronous packet traffic is represented by a single class. The allocation of slots in each frame to the various traffic classes is dynamically decided by means of two control levels: a distributed one, acting on the admission of isochronous connection requests independently for each class; and a centralized one, periodically adjusting two sets of parameters that affect the admission control functions. One such set represents upper bounds on the number of slots that may be assigned to each isochronous traffic class, i.e. its maximum capacity; the remaining bandwidth is shared by the asynchronous packets by means of the distributed queue algorithm. The capacity assignment is performed by minimizing a cost function that takes into account call blocking probabilities and packet delay. The control problem, which is characterized by multiple time scales, is discretized with respect to events occurring in an equivalent overall embedded Markov chain. Simulation results show the effectiveness of the method in ensuring quality of service requirements.