Advantages of Distributed Data Processing
By John Papiewski
A distributed data processing system is one that uses several computers to host a website, crunch numbers or store documents in a company network. In the early days of mainframes, many users shared a single computer. When businesses adopted personal computers, each person had his own PC but was restricted to the capabilities of one machine. With a network of modern desktop PCs, file servers and mobile devices, you can divide tasks among many computers.
When you divide a computing function among several machines, you can fine tune each computer to suit the needs of each task. For example, a company's website requires a Web server and a database server. A lightly used site can manage having both functions on a single computer, but putting each task on its own, dedicated machine permits the site to handle more business. The Web server computer serves pages more efficiently because it doesn't have to process database requests at the same time. The administrator can set up the database computer with additional hard-drive space and memory to better store and process large amounts of information, optimizing it independently of the Web server.
Although modern desktop PCs have excellent calculation performance for spreadsheets and databases, some complex applications require the combined power of several connected machines to deliver results in a timely fashion. Three-dimensional simulations and video processing, for example, take many billions of calculations to solve. A distributed data processing technique called a Beowulf Cluster assigns part of the computing task to each of several computers. Beowulf Clusters consist of inexpensive, off-the-shelf computer hardware connected with standard network switches and routers. The commodity approach gives you a level of performance that historically took million-dollar, custom hardware to achieve.
Resilience and Reliability
For some companies, computer uptime is crucial; even small outages can cause problems for customers. You can improve uptime by using high-availability software such as Microsoft Windows Server with Failover Clustering. This version of Windows automatically shares tasks among several network-connected computers. If one computer fails, the software automatically removes it from the group and continues to run uninterrupted on the remaining machines.
Local Data Access
A company with several regional offices gets faster access to information if each has a local copy of the company's data. When the data resides on a single computer, distant offices can face lengthy response times due to Internet traffic, especially during daytime hours. Each office can have its own copy of the data, eliminating these delays. Periodically, a process synchronizes each branch to the main office, ensuring that the data remains current.
Chicago native John Papiewski has a physics degree and has been writing since 1991. He has contributed to "Foresight Update," a nanotechnology newsletter from the Foresight Institute. He also contributed to the book, "Nanotechnology: Molecular Speculations on Global Abundance."