Optimizing Local Area Network Performance: Insight from Riverbed Modelling

Abstract

Designing and deploying an efficient and cost effective network that will satisfy end users requires a careful consideration of the different network devices that are available and selecting network parameters that will guarantee optimum performance. In this study, Riverbed modeling software was used to simulate and study the effects of varying different network parameters and devices on the performance of a local area network (LAN). Starting off with the designated Simple Network and Busy Network, the network parameters were varied for other scenarios to investigate the effects of network background utilization, link type, as well as the number of servers. Results from the simulation show that for a local area network, the web server demonstrated the highest CPU utilization. While increasing the background utilization of the network increased the remote server access time, FTP get download time, and HTTP request page response time, replacing the three servers with a single server did not have any significant impact on the CPU utilization. Further results from the simulation show that using cables enhanced the data transmission rate, and rather than deploying multiple servers, effective selection and distribution of network services can be used to achieve optimum network performance using a single server. This is achieved by striking a balance between frequently utilized network services and non-frequently utilized network services.