Optimized Workload Placement Using a New Batch Scheduling Algorithm

VNFs usually have more placement constrains than general cloud applications. Although NFV workload requests reach operator as a batch, NOVA scheduler only works on a first come first serve basis. This generally results in sub-optimal placement and stranding resource usage. In this talk, we will present a new offline placement algorithm. The tool dynamically collects resource usage of the cloud and takes the batch workload requests as inputs. All the special requests of each VNF, such as affinity