Exersize Solutions Of Distributed Operating System By P.k.sinha
The distributed operating system by p.k.sinha is one of the most trending technologies nowadays because, it allows users to develop scalable and efficient systems which are robust with regards to failures, have high availability at a lower cost, reduce network traffic/latency, has reduced complexities of replication and supports different types of load distribution. There are many solutions for this type of system like consensus protocol (e.g., Paxos), leader election protocols (e.g., random or best-effort assignment), data replication protocols (e.g., single or dual writes), quorum access protocols (e.g. majority or weighted majority) and many other algorithms. This article will provide a complete exercise solutions of all protocol specifications included in text book distributed operating system. The author of this article have tried to provide the best possible solution by taking help from different open source programs available online. All the credits goes to the authors of these programs. If anyone have any query regarding any problem or solution please contact at contact@exersisedsolutions.com The first exercise is based on consensus protocols related to distributed operating systems, which are used for building fault-tolerant applications having multiple replicas, reliable execution of multistep transactions, consensus protocols are used in many other areas as well like e-commerce, network routing etc. The exercise is divided into two parts, first part deals with basics of consensus protocols and other deals with building distributed transaction. The purpose of this exercise is to give you an understanding of the consensus algorithms, here the problem statement provide you the basic understanding about consensus protocol design. Researchers can be implement different types of consensus protocol for example, leader election algorithm or random assignment algorithm etc. Leader election algorithm implements in single node (it means there will be only one leader in network), where as random assign algorithms implements in multiple nodes. For this exercise use random assignment algorithm to implement leader election in distributed system. This will allows you to understand how it works and what are benefits and draw backs using this type of algorithm. For this exercise, please note that in above figure the file system is implemented in four nodes A, B, C and D. The user can write data in any node and read back from that node or any other node. Here we assume the network is in communication and connectivity with each other. Now assume if we send a message to node A and make it as “leader” for this all nodes (A, B, C and D). So we need to make A as leader by implementing random assignment algorithm. Following are the steps for random assignment algo: 1) In first step of this algo get two integers using get two integers function from internet. 2) Now make the first integer as the node id (for example, Node A will be assigned new random value 1, B will be assigned new value 2, C will be assigned new value 3 and D remains 4). 3) Now keep new values of node ID in array (the array will store the name of all nodes in network), which you can use to assign each node with unique numbers. 4) After keeping all new values into array you need to sort it by using compare two values function of internet. 5) After sorting these numbers according to above steps you are now able to assign order for leader. So please assign A as leader, B second leader and so on.
Click468eeb4e9f3225 Click
