IEEE Access (Jan 2019)
A Queuing Delay Utilization Scheme for On-Path Service Aggregation in Services-Oriented Computing Networks
Abstract
In services-oriented computing networks, packets in the process of routing to a data center must wait for a sufficient amount of data before service aggregation to reduce the network transmission load. However, packets must be uploaded to the data center as soon as possible to reduce delay. With the exponential growth in the number of IoT connected devices, the wait time for packets is longer at routers due to massive amounts of data, which causes a large queuing delay. If this queuing time can be utilized for service aggregation in a service-oriented computing network, the network performance will be substantially improved. Therefore, a queuing delay utilization scheme for on-path service aggregation (SAQD) is proposed in this paper. This scheme has the following innovations: 1) SAQD fully utilizes the queuing delay of packets for service aggregation, which can effectively reduce the transmission volume and communication overhead. Based on the proposed service aggregation algorithm, packets are divided into forwarding packets and aggregating packets, and the service aggregation of aggregating packets is completed by utilizing the transmission time of forwarding packets to ensure that the transmission volume and communication overhead are effectively reduced without additional latency. 2) SAQD can effectively alleviate the traffic pressure of the data center and balance the workload of routers. By the service aggregation and intranet cache of routers, some requests for the data center can be handled by routers, which reduces the traffic pressure of the data center, especially in the peak period. Compared with conventional schemes, the experimental results demonstrate that SAQD reduces the workload of the data center by 55.8%-66.26% and provides users with a better quality of experience by reducing the request response delay by 31.33%~51.41%.
Keywords