Most applications today go for Load Balancing Methodology to distribute
work across different servers, and to enhance continuity of business. Going
for a load balancing methodology increases system reliability by providing
redundancy. A load balancer is a software program that is listening on the
port and forwards the request to the server which responds to the load
balancer.  Mechanisms are available to forward request to backup servers in
cases of failure of the main server. There are a variety of scheduling
algorithms in place for distributing the work among different servers.


The diagram explains how a simple load balancing architecture looks like
and works.
The request from the client is forwarded to either of Server 1 or Server 2
whichever is relatively free. The decision is made by the load balancer
based on the algorithm. The algorithm could be as simple as random choice,
round robin or sophisticated based on the resource utilization and
performance. Suppose Server 1
fails, then there is a process set up to handle such situations “ON FAILURE
SWITCHOVER TO BACKUP SERVER”. The switchover happens automatically in most
times; In case it fails then the administrators do a manual switchover. The
Backup server acts as the main server and responds to the requests till the
main server is completely restored. This ensures continuity of business.
The failure of one of the servers is not a hindrance to progress or
production. A load balancing architecture could include more than 2 servers
as well based on business need.