- مبلغ: ۸۶,۰۰۰ تومان
- مبلغ: ۹۱,۰۰۰ تومان
Advance reservation services are being used by a range of applications to schedule connection bandwidth resources at future time intervals. To date many different algorithms have been developed to support various point-to-point reservation models. However, with expanding data distribution needs there is a need to schedule more complex service types to provide connectivity between multiple sites/locations. In particular, these offerings can help improve network resource utilization and help expand carrier service portfolios. Along these lines, this paper presents a novel, scalable optimization solution to schedule (virtual) overlay networks with fixed end-point nodes. An improved re-routing heuristic scheme is also proposed and analyzed for comparison purposes.
6. Conclusions and future work
Network advance reservation (traffic scheduling) is an important area, and a range of studies have looked at reserving point-to-point connection demands. However, as user applications continue to evolve, there is a further need to schedule more complex overlay topologies at future intervals. This paper addresses this concern and presents a novel dynamic optimization scheme for overlay scheduling. In addition, some advanced heuristic solutions are also developed by using connection re-routing strategies. Overall findings confirm that the optimization approach gives the lowest blocking and resource utilization. However, the proposed re-routing heuristics also provide very good gains over a baseline greedy heuristic scheme. In particular, strategies which use thresholding to minimize the amount of re-routed capacity give the highest request acceptance rates. Future efforts will look at extending these schemes to address the more general virtual network scheduling problem for cloud-based settings.