Feedback
Table Of Contents
Load Balancing in PEP
When multiple PDPs are configured to communicate with the PEP, the request- response time may be considerably higher because the PDP is down or the PDP is busy handling other requests. Load-balancing addresses this problem by allowing the protected application server to distribute the load to different PDPs based on a load-balancing policy or algorithm that is configured for the PEP.
The CEPM implements the round robin load-balancing policy.
Figure C-1 Loadbalancing in PEP
When the PEP sends a request to the PDP, the load-balancer component decides to send the request to the first PDP registered within the cluster. If the specified PDP is down (not alive) or busy handling any other request, the request is redirected to the second PDP as per the round robin policy configured in the PEP component.
In this way, when the last PDP (PDP 4 in the diagram) is alive and responds to the PEP Request 3, the very next request (Request 4) is transferred to the first PDP (that is PDP 1) as per the round robin algorithm.
Unlike in Java PEP, the load-balancing process for the .NET PEP is not protocol-independent. The rule is implemented within any number of the PDPs having SOAP as the transport protocol.
