Computer Clusters
Computer clustering involves the use of multiple hardware, typicallypersonal computers (PCs) or UNIX workstations, storage devices, andredundant interconnections, to form what appears to users as a singleintegrated system (Cluster computing). Clustering has been available sincethe 1980s when it was used in Digital Equipment Corp.'s VMS system. Today,most leading hardware and software companies support clustering because ofits demand for parallel processing, batch processing, load balancing and Parallel processing is the processing of program instructions bydividing them among multiple processors with the objective of running aprogram in less time. Parallel processing is normally applied forrendering and high computational based applications. Rather than usingexpensive specialized supercomputers for parallel processing, implementershave begun using a large cluster of small commodity servers. Each serverruns its own operating system, to take a number of jobs, process them, andsend the output to the primary system (Grama, 2003). Clusters provide theability to handle a large task in small bits, or lots and lots of smalltasks across an entire cluster, making an entire s
In a business network for Internetapplications, a cluster, often called a Web farm, might perform suchservices as providing centralized access control, file access, printersharing, and backup for workstation users (Server farm). The first PC cluster to be described in scientific literature wasnamed Beowulf and was developed in 1994 at the NASA Goddard Space FlightCenter (Beowulf clusters compared to Base One's batch job servers). For only $40,000 in hardware, Beowulf produced the processingpower of a small supercomputer costing about $400,000 at that time. Just as clustering has reduced the important of supercomputers forparallel processing, clusters are making the mainframe less relevant forbatch applications. Clustering has been one of the beneficial innovations in technology. 999 percent availability in some cases (Cluster computing). It has already fulfilled its promise for improving parallel processing,batch processing, load balancing and high availability. Grid computing adds detailed scheduling, high levels of serviceand distributed control (Douglas, 2002). Ideally, the resources of thefailed hardware are taken over by the remaining nodes in the system andtherefore there is no system downtime. Load balancing is commonly used in applicationswhere the load on the system cannot be predicted and is variable from timeto time. Beowulf initially consisted of sixteen PCs, standard Ethernet, and Linuxwith modifications and achieved seventy million floating point operationsper second. Before clustering, batch applications were typically thedomain of expensive mainframes. One example where load balancing is often used is for web serverswhere two or more servers are configured in such a way that when one servergets overburdened with requests, they are passed on to the other servers inthe cluster, thus evening out the work. This is implemented by mirroring thecritical components in the system such as the storage sub-systems and thepower supply.
Common topics in this essay:
Pentium III,
Corp's VMS,
Ethernet Linux,
PCs UNIX,
Base One's,
,
parallel processing,
load balancing,
Center Beowulf,
University Mexico,
cluster computing,
supercomputers parallel processing,
cluster computing clustering,
parallel processing batch,
computing clustering,
grid computing,
supercomputers parallel,
processing load balancing,
processing batch,
batch processing load,
batch processing,
floating operations,
d'souza 2001,
|
Acceptance Essays
