Abstract:Emerging edging computing services and applications bring an unprecedented demand for bulk transfers at the network edges. However, the expensive access cost makes it difficult to deliver bulk data between geo-distributed-edge-datacenter. In this paper, storage in edge-datacenter is introduced into the transmission scheme for temporarily storing delay-tolerant bulky flows. We formulate the upstream resource allocation as a stratified multi-objective optimization model, which can adjust the spectrum and storage allocation between latency-critical flows and delay-tolerant flows. Our studies reveal that 60% of the system cost can be saved by trading cost-effective storage for expensive spectrum resources.

Performance and Cost of Upstream Resource Allocation for Inter-Edge-Datacenter Bulk Transfers