On communication cost vs. load balancing in Content Delivery Networks

Abstract

Abstract-It is well known that load balancing and low delivery communication cost are two critical issues in mapping requests to servers in Content Delivery Networks (CDNs). However, the trade-off between these two performance metrics has not been yet quantitatively investigated in designing efficient request mapping schemes. In this work, we formalize this trade-off through a stochastic optimization problem. While the solutions to the problem in the extreme cases of minimum communication cost and optimum load balancing can be derived in closed form, finding the general solution is hard to derive. Thus we propose three heuristic mapping schemes and compare the trade-off performance of them through extensive simulations.Our simulation results show that at the expense of high query cost, we can achieve a good trade-off curve. Moreover, by benefiting from the power of multiple choices phenomenon, we can achieve almost the same performance with much less query cost. Finally, we can handle requests with different delay requirements at the cost of degrading network performance.Index Terms-Content delivery networks, distributed load balancing, power of multiple choices, query cost.

A Novel Communication Cost Aware Load Balancing in Content Delivery Networks using Honeybee Algorithm

“…Considering the communication cost along with the maximum load of servers (that translates to response time) as the two critical decision-making parameters affecting request dispatching is the main goal of [4] and [5]. These papers show that there is a trade-off between communication cost and response time in a distributed network.…”

“…These papers show that there is a trade-off between communication cost and response time in a distributed network. To handle the abovementioned trade-off, [5] introduces three basic algorithms each of which leads to a different trade-off curve. Furthermore, it assumes that each server has a limited cache size, and therefore, it can respond to a specific group of requests.…”

“…Its results show that more constraints on communication cost lead to worse response time. Despite [5]'s assumption about limited cache size, the authors of [15] consider every server is capable of responding to every request which is a simplified model of real CDNs.…”

“…However, it is obvious that in a CDN, a single server cannot hold all the available data, so having a suitable content distribution algorithm among all the available servers is necessary. There are some challenging problems and key features that have a direct effect on a CDN's performance such as content In this paper, we have used the stochastic optimization problem defined in [5] as our main problem definition. Our proposed load balancing method, which is based on the Artificial Bees Colony (ABC) algorithm (derived from the natural behavior of honeybees foraging for food resources [6]), aims to provide an acceptable solution of the above-mentioned optimization problem.…”

“…The more accurate this information is, the more effective and efficient the dispatching decisions are made. In [5], the authors supposed that servers are ideally aware of each others' load. However, this assumption is not valid in the real world scenarios.…”