Four Failures That Matter Pos 355

Four Failures That Matter

James K. Miller

POS/355

January 30, 2014

Pinchas Smolianski

Four Failures That Matter

Distribution failures occur for multiple reasons in a system and below those reasons will are discussed in its entirety.

Some Network systems can have the same type of failures but not every network system failures occur in the same manner because they are different. The two types of network systems that people use distributed and centralize systems. Distributed system failures occur for multiple reasons and discussing those reason in its entirety will provide a better understanding in how the system failures occur, why they happen and also how to correct a system failure issue and reconfigure the system after a failure occur in its components. A distributed system is designed to be fault tolerant and there are three types of common failures network, link failure, storage medium failure, and host failure. There is also other failure like halting failure that just causes a component to stop. Centralized systems share two of the three failures with a distributed system, the host failure and the storage medium failure. The network link failure can only occur in the network distributed system. To detect link and network failure the handshaking procedure is used by sending out an “I am up” message between link and the host with a certain time period. If “I am up” message does not return within the time allotted to the site it will send out “Are you up?” message. After the allotted time has pass and the site still has not received the process can be repeated, if this does not work the site can safely a conclusion that some sort of failure occurred and a “Time out scheme” is displayed. Although the failure occurred, the site can try to figure out if it’s a link failure or a site failure by sending a “Are you up” to the other site by a different route if there is another route and reply completely. When reply is received this tell the opposite site that...