Load Balancer / Cluster - Algorithms for balancing

A computer cluster is a single logical unit consisting of multiple computers that are linked through a LAN. The networked computers essentially act as a single, much more powerful machine. A computer cluster provides much faster processing speed, larger storage capacity, better data integrity, superior reliability and wider availability of resources.

Computer clusters are, however, much more costly to implement and maintain. This results in much higher running overhead compared to a single computer.

Many organizations use computer clusters to maximize processing time, increase database storage and implement faster data storing & retrieving techniques.

There are many types of computer clusters, including:

• Load-balancing clusters
• High availability (HA) clusters
• High performance (HP) clusters

The major advantages of using computer clusters are clear when an organization requires large scale processing. When used this way, computer clusters offer:

• Cost efficiency: The cluster technique is cost effective for the amount of power and processing speed being produced. It is more efficient and much cheaper compared to other solutions like setting up mainframe computers.
• Processing speed: Multiple high-speed computers work together to provided unified processing, and thus faster processing overall.
• Improved network infrastructure: Different LAN topologies are implemented to form a computer cluster. These networks create a highly efficient and effective infrastructure that prevents bottlenecks.
• Flexibility: Unlike mainframe computers, computer clusters can be upgraded to enhance the existing specifications or add extra components to the system.
• High availability of resources: If any single component fails in a computer cluster, the other machines continue to provide uninterrupted processing. This redundancy is lacking in mainframe systems.

Algorithms for balancing is one of the most important factors in this context, then we will explain three basic methods :

Least Connections

This technique redirects the requests to the lowest based on the number of requests / server connections. For example, if server 1 is currently handling 50 requests / connections, and server 2 controls 25 requests / connections, the next request / connection will be automatically directed to the second server, since the server currently has fewer requests / connections active.

Round Robin

This method uses the technique of always direct requests to the next available server in a circular fashion. For example, incoming connections are directed to the server 1, server 2 and then finally server 3 and then the server 1 returns.

Weighted Fair

This technique directs the requests to the load based on the requests of each and the responsiveness of the same (performance) For example, if the servers server 1 is four times faster in servicing requests from the server 2, the administrator places a greater burden of work for the server 1 to server 2.

Refer:

https://www.esds.co.in/blog/cluster-computing-definition-and-architecture-of-a-cluster/#sthash.SEcCghTH.dpbs