Method and apparatus for load balancing in network based telephony application

In a system comprised of a network routing calls between a plurality of callers and at least one receiver using a session initiation protocol, wherein a load balancer sends requests associated with calls to servers, a method for directing requests associated with calls to said servers, the method comprising the steps of:

• assigning different weights to different request types based on an overhead associated with processing requests of the different request types, wherein the different request types comprise an INVITE request and a BYE request, wherein the INVITE request is assigned a first weight, and the BYE request is assigned a second weight, wherein the first weight is higher in value than the second weight;
• maintaining a counter for each of a plurality of servers, wherein the counter for a given server indicates a total count of a weighted number of requests assigned to the given server;
• receiving a request associated with a caller, wherein the caller corresponds to a device or client application making a call;
• estimating an amount of work that each of the plurality of servers has left to do for processing calls currently assigned to the plurality of servers, wherein said amount of work that a given server of the plurality of servers has left to do is estimated based on the total count of the weighted number of requests assigned to the given server as indicated by the counter maintained for the given server;
• selecting a server of the plurality of servers to receive and process the request associated with said caller based on the estimated amount of work that the selected server has left to do as compared to the estimated amount of work that at least one other server of the plurality of servers has left to do;
• receiving a subsequent request that may be associated with said caller or another caller;
• determining whether or not the subsequent request is associated with said caller;
• and
• sending the subsequent request to said selected server based on determining that the subsequent request is associated with said caller.

In a system comprised of a network routing calls between at least one caller and a plurality of receivers using a session initiation protocol, wherein a load balancer sends requests associated with calls to servers, a method for directing requests associated with calls to said servers, the method comprising the steps of:

• assigning different weights to different request types based on an overhead associated with processing requests of the different request types, wherein the different request types comprise an INVITE request and a BYE request, wherein the INVITE request is assigned a first weight, and the BYE request is assigned a second weight, wherein the first weight is higher in value than the second weight;
• maintaining a counter for each of a plurality of servers, wherein the counter for a given server indicates a total count of a weighted number of requests assigned to the given server;
• receiving a request associated with a receiver, wherein the receiver corresponds to a device or client application receiving a call;
• estimating an amount of work that each of the plurality of servers has left to do for processing calls currently assigned to the plurality of servers, wherein said amount of work that a given server of the plurality of servers has left to do is estimated based on the total count of the weighted number of requests assigned to the given server as indicated by the counter maintained for the given server;
• selecting a server of the plurality of servers to receive and process the request associated with said receiver based on the estimated amount of work that the selected server has left to do as compared to the estimated amount of work that at least one other server of the plurality of servers has left to do;
• receiving a subsequent request that may be associated with said receiver or another receiver;
• determining whether or not the subsequent request is associated with said receiver;
• and
• sending the subsequent request to said selected server based on determining that the subsequent request is associated with said receiver.