Distributed Systems

Distributed systems are connected by a network of servers, and computers usually in different locations; whereas, a centralized system is in a single location. The physical proximity of servers can be a problem concerning redundancy and replication of information. In a centralized system, servers, and computers are usually kept in one location. Therefore, if there was an event such as a fire, all components of the businesses information could be lost. Data backups could help this, but if physical equipment is lost the business could be non-functional in a centralized system for an extensive time. With a distributed system redundancy and replication is distributed across the network, therefore, an event such as a fire could enable the system to take over in a different location enabling business to continue.

As described above, physical location can affect failure rate, although there are different types of failures that can happen in a system beyond the type of system. There are many types of failures commonly discussed, although this is a basic list. There are many more that could be outlined and discussed. Here are several:

• Network Failure: A distributed system must have network communication between each other. At times, “a failure can have total loss of communication along parts of the network” (Redwine & Holliday). If part of the network fails (a network link breaks), the idea is to have an alternate route for communication to continue. This creates network redundancy.

• Hardware Failure can be a problem in a distributed network as well as a centralized network. Fortunately, hardware has become less of an issue in recent past due to durability, and improvements in design of hardware.