Edge Provisioning with Flexible Server Placement

• Published in: IEEE transactions on parallel and distributed systems Vol. 28; no. 4; pp. 1031 - 1045
• Main Authors: Hao Yin, Xu Zhang, Liu, Hongqiang Harry, Yan Luo, Chen Tian, Shuoyao Zhao, Feng Li
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.04.2017; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Balances (scales); Cloud computing; Complexity theory; Computer simulation; Coordinates; decoupling; Delays; Edge computing; Edge provisioning; Electronic mail; Fault tolerance; New technology; optimization; Placement; Provisioning; Servers; unforeseen locations
• ISSN: 1045-9219; 1558-2183
• DOI: 10.1109/TPDS.2016.2604803

We present Tentacle, a decision support framework to provision edge servers for online services providers (OSPs). Tentacle takes advantage of the increasingly flexible edge server placement, which is enabled by new technologies such as edge computing platforms, cloudlets and network function virtualization, to optimize the overall performance and cost of edge infrastructures. The key difference between Tentacle and traditional server placement approaches lies on that Tentacle can discover proper unforeseen edge locations which significantly improve the efficiency and reduce the cost of edge provisioning. We show how Tentacle effectively identifies promising edge locations which are close to a collection of users merely with inaccurate network distance estimation methods, e.g., geographic coordinate (GC) and network coordinate systems (NC). We also show how Tentacle comprehensively considers various pragmatic concerns in edge provisioning, such as traffic limits by law or ISP policy, edge site deployment and resource usage cost, over-provisioning for fault tolerance, etc., with a simple optimization model. We simulate Tentacle using real network data at global and county-wide scales. Measurement-driven simulations show that with a given cost budget Tentacle can improve user performance by around 10-45 percent at global scale networks and 15-35 percent at a country-wide scale network.