Introduction

In order for an end user to be able to comfortably use a cloud gaming service, it must be within a reasonable latency range of a cloud gaming server and the server must host the user's requested game. In this project, we investigate various heuristics for server and game placement in order to reach as many potential end users as possible and satisfy these end users with their requested games.

Server Placement

Many cloud gaming providers use few datacenters to provide their services; however, this could result in a significant fraction of the target population being beyond the 80ms RTT of a datacenter. Therefore, in addition to using few, large datacenters, EdgeCloud proposes the use of servers that are located on the edge (such as CDN servers). However, there is a cost of incorporating an edge server into a cloud gaming architecture. Therefore, we aim to reach as many end-users as possible with as few edgeservers. These edge servers must also be intelligently placed in order to satisfy as many end users as possible. We explore various heuristics such as random placement, placement according population distribution, region-based placement, and client-voting for a particular location

Game Placement

Edge servers, such as those found at CDNs, have limited storage. Given that some games may require 10 to 20GB of storage, an edge server may have the potential to serve perhaps 100 or at most 200 games. We investigate various heuristics for game placement such as random, most popular, mixing between popular and unpopular games, and a client-voting based approach.