Installing MATRIXX Traffic Routing Agent

Topics below provide information about how to install the Traffic Routing Agent (TRA) software.

The number and type of Traffic Routing Agents you deploy depends on the number of engines or engine chains (if multiple customer sub-domains are implemented) in your MATRIXX environment. Traffic Routing Agent clusters perform different functions based on how they are deployed and configured. Regardless of their function, TRAs are always deployed as two-node clusters that are active-standby pairs.Table 1 lists the different TRA functions:

Table 1. Traffic Routing Agent Servers
Traffic Routing Agent Servers Description
TRA-PROC

TRA-PROC with RCP

 (Required) TRA load balancer for MATRIXX Engine processing servers.

These TRA servers balance payload traffic between the processing servers in the MATRIXX Engine. A TRA-PROC or TRA-PROC with RCP is usually collocated with the processing servers that it is balancing the load for. This is the preferred configuration.

A TRA-PROC with RCP (Route Cache Proxy) is required if multiple sub-domains are implemented.
TRA-PUB (Required) These TRA servers route to the active MATRIXX Engine publishing server all internal MDC traffic (the transaction stream, transaction stream control messages, and Init database requests), event streaming, CMI traffic from cluster managers on other clusters, and SNMP traffic.

The TRA-PUB is collocated with the publishing servers of the MATRIXX Engine.
TRA-SI

TRA-DR

 These TRA servers are engine disaster-recovery agents. They ensure that the payload traffic they forward always arrives at an active MATRIXX engine.

The TRA-SI (Site Independence) and TRA-DR (Disaster Recovery) functions operate at the Application Protocol Layer. The TRA-SI performs this task for TCP/UDP data on the transport protocol layer, and the TRA-DR performs this task on Diameter data.

The TRA-DR and TRA-SI functions can be deployed within a single deployment configuration (where one TRA installation performs the combined functions (TRA-RT/SI)). For sub-domain routing, these functions are usually collocated with a TRA-RT (TRA-RT-(SI/DR)).
TRA-RT-(SI/DR) Application Layer Routing Agent for sub-domain routing.

A TRA-RT server routes the network payload traffic based on subscriber information contained in request messages; a request is forwarded to a MATRIXX Engine that hosts the subscriber. The TRA-RT usually also includes the TRA-SI or TRA-SI functions as a TRA-RT-(SI/DR).

A TRA-RT-(SI/DR) is only required when multiple customer sub-domains are implemented in a MATRIXX environment. It connects with the Route Cache Controller for sub-domain routing. This server works with the TRA-PROC with RCP.

The following diagram illustrates a TRA deployment in a MATRIXX environment that has two customer sub-domains. The TRA-RT-(SI/DR) routes requests to the correct sub-domain, identifying the sub-domain by querying the Route Cache. It also handles any disaster-recovery (engine shutdown) scenario by routing traffic to the newly active engine at whichever site it is running.

The illustration below shows how the TRA functions are deployed in a MATRIXX environment that has two customer sub-domains:

Figure 1. Example TRA Deployment

Note the following points about installing Traffic Routing Agents:

  • For high availability, all Traffic Routing Agents nodes are deployed as two-node clusters (TRA HA cluster) where the nodes are active-standby peers and each node resides on its own server.
  • While a TRA-RT HA cluster could be deployed separately from a TRA-SI or TRA-DR HA cluster, typically these functions are deployed as a combined TRA-RT-(SI/DR) HA cluster because less server hardware is required.
  • A TRA-RT-(SI/DR), TRA-SI, or TRA-DR must be installed on a non-MATRIXX Engine server.
  • All TRA nodes can be installed on the same physical hardware as the Network Enabler (when Traffic Routing Agent and Network Enabler are both installed in an environment).
  • For a TRA-PROC or TRA-PROC with RCP HA cluster, the preferred configuration for a given engine is to install one TRA-PROC on processing blade server 1 and its peer TRA node on processing blade server 2. Alternatively, the TRA nodes of a TRA-PROC HA cluster could be installed on their own servers at the same site as the engine they serve.

After you install the Traffic Routing Agent RPM, you defined that TRA function by editing the TRA configuration file. If the TRA is configured as a TRA-RT (sub-domain router), you must also enable the Route Cache Controller service and ensure that the MATRIXX configuration file on the TRA-RT server contains the required Route Cache Controller configuration settings. For information about configuring Traffic Routing Agents, including example configuration files, see the discussion about configuring MATRIXX Traffic Routing Agent in Installation and Configuration.