Summary
Media‐centric networks deal with exchanging large media files between geographical distributed locations with strict deadlines. In such networks, resources need to be available at predetermined timeslots in the future and thus need to be reserved in advance, based on either flexible or fixed timeslot sizes. Reliability of the transfers is also important and can be attained by advance provisioning of redundant reservations. This, however, imposes additional costs, because redundant reservations are rarely in use, causing network resources to be wasted. Further adaptation and network utilization can be achieved at runtime by reutilizing unused reservations for transferring extra data as long as no failure has been detected. In this article, we design, implement, and evaluate a resilient advance bandwidth‐reservation approach based on flexible timeslots, in combination with a runtime adaptation approach. We take into account the specific characteristics of media transfers. Quality and complexity of the proposed approach have been extensively compared with that of a fixed timeslot algorithm. Our simulation results reveal that the highest admittance ratio and percentage of fully transferred requests in case of failures are almost always achieved by flexible timeslots, while the execution time of this approach is up to 17.5 times lower, compared with the approaches with fixed timeslot sizes.