original joint optimization of compute and network resource allocations. For 20% time- sensitive applications, we show that we could guarantee 200-times smaller service latencies while reducing DC utilization efficiency by no more than 30%.
To provide edge computing with pre-negotiated, guaranteed time budget, we propose an original joint optimization of compute and network resource allocations. For 20% time-sensitive applications, we show that we could guarantee 200-times smaller service latencies while reducing DC utilization efficiency by no more than 30%.