Failures in a Distributed System

Failures In A Distributed System

During the design and implementation of a reliable distributed system is a very important. The failure of this system can result in easily repairable errors, such as hardware, software, or catastrophic meltdowns where hardware overloads, hardware crashes that were not protected. Faults can occur in any of the components of a distributed system.

There are four types of failures more commonly known as compute cluster which allows applications to run on more than one computer (node). Data centric systems that involve high volumes of access are more difficult to make fault tolerant. They require failure monitoring software and a special configuration. These types of systems are high availability clusters that allow two or more computers and a set of shared disks. Any application can be installed on the shared disk which allows all other computers access, with certain restrictions designated for each user or pc. The running application's node has exclusive access to the application's data on disk. Specific applications may run very well from this type of system. The application may fail on the shared disk and the nodes exclusive access is terminated and is granted to another node in the cluster. The application is restarted on a new node, which the user will notice a slight interrupt and lose of some data.

Design issues are very important in utilizing the multiplicity of processors and storage devices transparent to users. In a conventional, centralized system the user interface should be transparent to what actually happens beneath the interface. All the communications between hardware, the software, remote access and local access. The distributed system should be responsible for locating the resources and for arranging for the appropriate interaction.

Another look at transparency is for user mobility. Rather than have a user be forced to log onto the system on one specific machine, it would be more transparent to allow them to log in where he or she desires with the desktop that individual is used to. By having the user that facilitates user mobility by bringing over the user's environment. With this home directory set up on the distributed system, it will allow the user to access any parts of the system with their usual desk top applications that would be shared on the shared drives. This allows for a decrease in computers and faster access