Nt1330 Unit 1 Exercise 1

Words: 466
Pages: 2

Distributed systems will help us to finish jobs faster by using multiple machines as one machine. This will be done using job scheduling and load balancing between multiple servers or machines. The machine, which request the jobs may, may divide it to multiple chinks, which the load balancer will send every chink to a server to be calculated then send response to the client. Other times, these servers could be used to serve multiple request from multiple users. In the case, the load balancer will divide the requests between servers and send response to the client.
Round robin is a technique for load balancing using the method of divide traffic equally between multiple destinations. This technique will send packet to first available server in the queue then send it to the end of the server queue. In this way, it will guarantee that servers have one or zero difference between numbers of jobs. In this way, the load balancer guarantees equal load between servers but this works will if all servers have the same processing power. Processing power mean the maximum number of processes that a server can handle before it crashes, reboot, or start to drop request. Processing
…show more content…
With time Servers processing power will decline every year and new servers will have different processing power than old servers. In this way, old server or servers with less processing power than other server will be overloaded very fast. To overcome this problem a new algorithm introduced will take this factor in mind. Weighted round robin will compute this processing power of every server. The difference between this way and round robin that it will check if the server at top reaches maximum requests number of connections, if not it will send it more requests otherwise it will send it to the end of the queue. In this way, every server will have his processing power