Dynamic adaptive advance bandwidth reservation in media production networks

Media production networks deal with large and predictable video transfers and streaming sessions which need to be scheduled to optimally use the limited network capacity. As the time and locality of transfers are predictable, advance bandwidth reservation results in more effective use of the underlying network. To offer reliable reservations, the incorporation of fault-tolerance related features in bandwidth reservation strategies resulting in redundancy is a necessity, but this incurs additional costs and extra performance overheads as network capacity remains unused to offer this protection. In order to mitigate these side-effects, we present an efficient event-driven runtime adaptive approach that continuously monitors the network transfers, dynamically updating the transfer schedule. This dynamic approach re-uses leftover network capacity to improve network utilization. The resulting management system offers protection from failures using resilient advance reservation, while also improving the network utilization and request admittance ratio. Our evaluation shows that the proposed approach leads to significant improvements, up to 13.9% in percentage of admitted requests in stable network conditions, compared to resilient advance reservation algorithms.