The DRA, as specified by 3GPP, allows MNOs to deploy multiple Policy and Charging Rules Function (PCRF) elements in their network. The DSC (operating as a DRA) correlates specific user equipment to IP-CAN sessions and ensures subsequent messages for the session are sent to the same PCRF regardless of which interface (e.g. Gx, S9, Gxx, and Rx) they originate.
The DSC also addresses the need for very high performance and reliability requirements for policy systems. Coupling DRA functionality and advanced load balancing, PRCF systems can be dynamically managed, avoiding congestion, and allowing incremental dynamic growth of servers.
As Diameter traffic volume rises and the number of Diameter server elements increase, MNOs are faced with the challenge of how to cost effectively operate and maintain policy (PCRF) in a scalable and resilient manner. There are considerations for traffic steering when multiple different PCRF systems are used, and for load distribution across multiple servers acting as a single PCRF. There will be challenges.
IP-CAN Session Routing – When a user establishes an IP-CAN session, for example when a tablet requests a video stream, all subsequent messages for this IP-CAN session (bandwidth policy/enforcement, charging, etc.) must be handled by the same PCRF that handled the initial session. Diameter messages to the PCRF can arrive on several different Diameter interfaces from different nodes and the IP-CAN session relationship is identified by Diameter parameters like IMSI and/or IP address. The challenge is that only the PCRF knows the IP-CAN session association making it impossible for the connected nodes to route the subsequent messages to the correct PCRF.
Policy Service Scalability – A PCRF is typically deployed on multiple servers to handle traffic volume requirements. When traffic increases and a new server is needed, many of the other Diameter network elements (Packet Data Network (PDN) Gateway, Serving Gateway, etc.) need to be updated with Diameter connections to this new server. Instead of being an isolated action to increase service capacity, it involves updating the entire network.
Policy Service Resiliency – A key element in deploying the policy service (PCRF) is its ability to survive in case of a failure. If all the PCRF servers are interconnected to the network, a single PCRF failure and the resulting recovery not only depends on the failed PCRF coming back online, but also re-establishing the connections to the other elements.
Interoperability – The number of vendors entering the Diameter signaling market is increasing, and despite Diameter being a standard, interpretations of the standard are numerous. When a new vendor enters the network an extensive amount of interoperability tests are required to minimize the number of incompatibilities and Diameter session failures.
The Mitel DSC Solution
The solution to these challenges is to deploy the DSC as a Diameter Routing Agent (DRA). The Diameter Routing Agent is defined in 3GPP specifications. In this configuration, the DSC “front-ends” a set of PCRF nodes, enabling an IT style server-farm deployment allowing IP-CAN session load sharing, very high service availability, topology hiding, and seamless back-end capacity upgrades. The DSC manages and routes traffic flow and optimizes application availability and performance.
3GPP defines an IP-CAN session as an association between the user equipment (phone, tablet, etc.) and an IP network. The DSC will load balance all IP-CAN establishment requests among the available PCRFs and store the corresponding IP-CAN session association information – IMSI, IP address, Access Point Name, etc. This enables the DSC to correctly route subsequent Diameter messages associated with an established IP-CAN session.
The DSC manages and routes traffic flow (IP-CAN sessions) between the available back-end servers. As traffic volume increases, new PCRF servers can be added without any impact to the Diameter network. This allows for a “start small” service deployment and the subsequent addition of capacity, in the form of new servers, as traffic volume increases.
Using Service-Oriented Routing technology, the DSC maintains service availability during loss of any server without involving the Diameter network, by load sharing all traffic between the other available PCRF servers. The DSC is backed by over 30 years of experience in providing high-availability solutions for telecom networks and is architected with all the HA characteristics of a carrier-grade solution.