Individual Failures

Individual Failures

Failure of distributed systems can cause errors that may have results of something that can be repaired easily or crashing of a system. The importance of a dependable distributed system ensures there will be a no fault tolerant system. A description of four types of failures in a distributed system will be discussed, as well as those applying to a centralized system. A solution on how to isolate and fix two of these types of failures will also be discussed.

Failures of hardware, software and networks cannot be fault free. Achieving fault tolerance should be with the least involvement of users or system administrators. Centralized systems are usually are more fault tolerant than a distributed system. Although, when more processor hosts are used there is more of a chance of having individual faults occurring more often. Four types of failures in a distributed system are; crash failures, response failures, omission failures, and byzantine failures (Distributed Systems (2012).

Crash failures occur when a server repeatedly fails to respond to inputs. When this failure occurs operation of the system is halted. The loss of all history will be lost and system reboot is necessary. When a crash failure occurs in the hardware all the hardware must be replaced or reinstalled. Fault tolerant systems can be developed by using these operating system failures (Distributed Systems (2012).

Response failures involve three different types of failures including value and state transition failures. The initial response failure is due to the server's response being incorrect. A value failure is when the value of the response is incorrect. Whereas, the state transition failure is caused when the server deviates from the correct flow of control (Distributed Systems (2012).

Omission failures occur throughout the server because of the lack of, reply, or response from the server throughout the distributed system. Different issues that occur are due to omission failures and the main reason for this is the server not listening or because of buffer overflow errors throughout the distributed system. This includes failure from the server responding to incoming requests, incoming messages the server fails to receive, and the sending of a message by the server (Distributed Systems (2012).

Byzantine failures are also known as arbitrary failures. These failures occur across the server of the distributed systems. This type of failure will cause a server to act in a nonrestrictive manner occurring at random times throughout the distributed system. The output that will be received from the server will be unreasonable. Thus, causing opportunities for Output from the server would be inappropriate and there could be chances of detrimental occurrences and double messages on the server side (Distributed Systems (2012).

Failures that occur in both distributed and centralized systems are