Call-on-hold for improving the performance of dynamic channel-assignment strategies in cellular networks

Efficient use of the limited radio spectrum is of paramount importance to supporting the ever-increasing number of mobile terminals. There is a need for devising complementing techniques to improve the performance of dynamic channel-assignment (DCA) algorithms to increase the capacity of cellular systems. In this paper, we study the concept of call-on-hold to improve the performance of a class of DCA algorithms called coordinated assignment without measurement (CAWM). DCA algorithms such as the Geometric strategy, the Nanda-Goodman strategy, the borrowing with directional channel locking (BDCL) strategy, and the two-step dynamic priority (TSDP) strategy fall into the CAWM class. To gain insight into the CAWM DCA algorithms, we simulated those algorithms and monitored carrier availability and failure to assign a channel characteristics in a cell. After observing these characteristics, we formulated a hypothesis as follows: If a new or handoff call is put on hold for a short while in a cell in the absence of an available channel, it is highly likely that the local base station will soon find a channel for the call. In the proposed approach, a DCA algorithm is said to have failed to assign a channel to a call only if a waiting call is delayed for longer than a threshold period called maximum delay. Our simulation-based study shows that it is possible to significantly reduce failure rates of the CAWM class of DCA algorithms by putting those calls on hold which would otherwise be blocked or dropped. The impact of small values of maximum delay on the average delay suffered by all calls in a network is negligible, but the reduction in failure rate is significant. We have explained how the call-on-hold idea can be easily integrated with the GSM system.