An Introduction To Clustering Under Linux
				=========================================

Shanker Balan
http://shankerbalan.com/
Tue Apr 29 18:13:10 IST 2003

Linux based clustering is big these days for the following reasons

 - Ability to use COTS (Cheap Off The Shelf) hardware
 - GPL software with source code
 - Easy to setup and maintain
 - Support from Linux newsgroups and user groups
 - Extensive documentation
 - Availability of cluster optimized applications - both GPL and commercial
 - Low TCO

 There are many types of clusters, the most common ones are 

 - Beowulf clusters
 - MOSIX" clusters
 - High-Availability cluster
 
 Which one should you use depends on what you wish to use your cluster for.
 Applications usually fall in one of the following X categories -
 Computational Intensive, IO intensive and High Availability.

Beowulf clusters -

 Beowulf clusters are the most common and popular clusters currently. These
 are used for computationally intensive jobs like simulations and other number
 crunching problems typically found in research centers and laboratories.
 Beowulf clusters use parallel computing to reduce the time required for a CPU
 intensive job. By distributing the workload across the clusters and
 paralleling executing the instructions, it is possible to reduce the job
 execution time considerably depending on the number of nodes in the cluster.

 Beowulf clusters can use either PVM (Parallel Virtual Machine) or MPI
 (Message Passing Interface) (there are others) as distributed application
 programming environments.

  PVM - 
   
   PVM (Parallel Virtual Machine) is a freely-available, portable,
   message-passing library generally implemented on top of sockets. It is
   clearly established as the de-facto standard for message-passing cluster
   parallel computing

	See http://www.linuxdoc.org/HOWTO/Parallel-Processing-HOWTO-3.html#ss3.4

  MPI - 
   
   Although PVM is the de-facto standard message-passing library, MPI (Message
   Passing Interface) is the relatively new official standard. 

   PVM Vs MPI has always been a religious war akin to Vi Vs Emacs. The author
   of the Linux Parallel Processing HOWTO has provided an unbiased summary of
   the differences b/w the two at

	http://www.linuxdoc.org/HOWTO/Parallel-Processing-HOWTO-3.html#ss3.5

   Though the MPI has clearly established itself as a standard, PVM still has
   an upper edge over MPI as it has more applications and coder base is also
   quite strong. But this would soon change as MPI becomes popular.

MOSIX Clusters:

 Now lets get to another type of clusters - MOSIX clusters. MOSIX is a
 software package that was specifically designed to enhance the Linux kernel
 with cluster computing capabilities. The core of MOSIX are adaptive (on-line)
 load-balancing, memory ushering and file I/O optimization algorithms that
 respond to variations in the use of the cluster resources, e.g., uneven load
 distribution or excessive disk swapping due to lack of free memory in one of
 the nodes. In such cases, MOSIX initiates process migration from one node to
 another, to balance the load, or to move a process to a node that has
 sufficient free memory or to reduce the number of remote file I/O operations. 

 MOSIX clusters are typically used in data centers and data warehouses.

  See http://www.mosix.org/txt_whatis.html

High Availability Cluster:

 As the name suggests, these are clusters which provide a high degree of fault
 tolerant server systems where 100% up-times are required. In the event of
 failure of one node, the other nodes which form the cluster will take over
 the functionality of the failed node transparently. HA clusters are typically
 used for DNS server, proxy and web servers.

  See http://www.linux-ha.org/