• Post Reply Bookmark Topic Watch Topic
  • New Topic

How is Load Balancing done in case of App Servers  RSS feed

 
Chris Wagnor
Greenhorn
Posts: 21
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
A little curious abt how the Load Balancing done in case of Application Server is done.
For example when a server is hosting multiple clients, to share the load you create multiple instances of App Server. I just know that this is Load Balancing, but dont know exactly how it is done.
Please throw some light on this.
Regards
New Java Rancher
 
Jeroen Wenting
Ranch Hand
Posts: 5093
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
typically there's one machine whose task it is to spread requests across all the others.
In a very simple system that could be a process running on one of two servers which forwards say 60% of requests to the other machine and handles 40% itself (giving some thought to the load the loadballancer itself places on the machine).
In a somewhat more complex system, one computer runs the load ballancer and assigns each new client that connects to one of several others to process from now on based on session ID.
It would keep track of sessions assigned to each of the other machines making sure that each gets an identical share (or whatever distribution is configured).
More complex systems still would also take into account the number and speed of requests from each of these sessions so that not only sessions but also actual requests are spread around (so one server might get more sessions than another if the sessions it does get yield less requests per second).
 
It is sorta covered in the JavaRanch Style Guide.
  • Post Reply Bookmark Topic Watch Topic
  • New Topic
Boost this thread!