UC San Diego

Today's data centers may contain tens of thousands computers. Manually monitoring these networks is time consuming, so there is an emerging need to automate these tasks by implementing network management systems. It is important for network engineers to have a tool enabling them to monitor the network to spot problems before users are affected. This need is also critical during the network design since it will make the network debugging easier. A well structured management software will also enables the designers to demonstrate the network behavior before development. One important component of most network management tools is visualization tool. Data centers' network architecture typically consists of a tree of routing and switching elements with more specialized and expensive equipment moving up the network hierarchy. When deploying the highest-end IP switches/routers, resulting topologies may only support 50% of the aggregate bandwidth available at the edge of the network, while still incurring tremendous cost. Non uniform bandwidth among data center nodes complicates application design and limits overall system performance. UCSD data center switch research group has proposed a new technique on how to leverage largely commodity Ethernet switches to support the full aggregate bandwidth of clusters consisting of tens of thousands of elements. The group argues that appropriately architected and interconnected commodity switches will deliver more performance at less cost than higher-end solutions. In this project a visualization tool was developed for the proposed network. This tool uses data captured from the network and use graphic techniques to visualize the data