Introduction to Cluster Computing Seminar:
Congestion avoidance:
When congestion is detected the device responds by throttling back the rate at which it sends the data. The device then uses an algorithm to lower the rate of data transfer and when congestion is cleared then it gradually increases the rate.
Cluster computing examples:
Many applications use cluster computing for greater performance, and flexibility. Examples of these are Google search engine, Image rendering. The google search engine uses the cluster computing as it provides load balancing, availability, reliability. When a user enters a qauey in the webpage, the browser searches for the IP address via the www.google.com Domain Name Server (DNS).
Google uses DNS based load balancing system that matches the query to a cluster that is geographically nearest to the user. The IP address of the selected cluster is returned. The web browser then sends the search request in the form of HTTP to the selected cluster. The selected cluster then processes the query locally. A hardware based load balancer monitors the available set of the Google Web Servers(GWS) in the cluster and distributes the request evenly among the cluster. The GWS machine receives the request, executes the query and sends the search result back to the user browser.
The second example is image rendering. In this, the images which are very large to analyze, such as any explosion are analyzed by using the divide and conquer technique. The dataset of the image are broken into smaller images and processed in the different nodes of a cluster. After this these datasets are integrated to form the single image. So, the cluster computing is of great importance in today’s scenario.
Download Cluster Computing Seminar Abstract .