L. Lovén, E. Peltonen, E. Harjula and S. Pirttikangas, "Weathering the Reallocation Storm: Large-Scale Analysis of Edge Server Workload," 2021 Joint European Conference on Networks and Communications & 6G Summit (EuCNC/6G Summit), 2021, pp. 317-322, doi: 10.1109/EuCNC/6GSummit51104.2021.9482593
Weathering the reallocation storm : large-scale analysis of edge server workload
|Author:||Lovén, Lauri1; Peltonen, Ella1; Harjula, Erkki1;|
1University of Oulu, Oulu, Finland
|Online Access:||PDF Full Text (PDF, 0.9 MB)|
|Persistent link:|| http://urn.fi/urn:nbn:fi-fe2021081643333
Institute of Electrical and Electronics Engineers,
|Publish Date:|| 2021-08-16
Efficient service placement and workload allocation methods are necessary enablers for the actively studied topic of edge computing. In this paper, we show that under certain circumstances, the number of superfluous workload reallocations from one edge server to another may grow to a significant proportion of all user tasks — a phenomenon we present as a reallocation storm. We showcase this phenomenon on a city-scale edge server deployment by simulating the allocation of user task workloads in a number of scenarios capturing likely edge computing deployments and usage patterns. The simulations are based on a large real-world data set of city-wide Wi-Fi network connections in 2013—2014, with more than 47M connections over ca. 800 access points. We identify the conditions for avoiding the reallocation storm for three common edge-based reallocation strategies, and study the latency-workload trade-off related to each strategy. As a result, we find that the superfluous reallocations vanish when the edge server capacity is increased above a certain threshold, unique for each reallocation strategy, peaking at ca. 35% of top ES workload. Further, while a reallocation strategy aiming to minimize reallocation distance consistently resulted in the worst reallocation storms, the two other strategies, namely, a random reallocation strategy, and a bottom-up strategy which always chooses the edge server with the lowest workload as a reallocation target, behave nearly identically in terms of latency as well as the reallocation storm in dense edge deployments. Since the random strategy requires much less coordination, we recommend it over the bottom-up one in dense ES deployments.
European Conference on Networks and Communications
|Pages:||317 - 322|
2021 Joint European Conference on Networks and Communications & 6G Summit (EuCNC/6G Summit)
Joint European Conference on Networks and Communications & 6G Summit
|Type of Publication:||
A4 Article in conference proceedings
|Field of Science:||
113 Computer and information sciences
This research is supported by Academy of Finland 6Genesis Flagship and DigiHealth programs (grants 318927, 326291); the ECSEL JU FRACTAL (grant 877056), receiving support from the EU Horizon 2020 programme and Spain, Italy, Austria, Germany, France, Finland, Switzerland; Infotech Oulu research institute; the Future Makers program of Jane and Aatos Erkko and Technology Industries of Finland Centennial foundations; and the personal grant for Lauri Lovén on edge-native AI research by the Tauno Tönning foundation.
|EU Grant Number:||
(877056) FRACTAL - A Cognitive Fractal and Secure EDGE based on an unique Open-Safe-Reliable-Low Power Hardware Platform Node
|Academy of Finland Grant Number:||
318927 (Academy of Finland Funding decision)
© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.