Configuring Stat Server

See the Stat Server Deployment Guide for deploying Stat Server and RTME Options Reference for information about Stat Server configuration options.

Additional Stat Server configuration information

Standard configuration of client connectivity (for reference)

The configuration of a Stat Server cluster in a small standard environment is fairly simple: you add a Stat Server node to its client’s connections in Configuration Server as shown in Figure 1. Unfortunately, this approach has some drawbacks for nonstandard environments that have a large number of Stat Server clients. For one, you must pay close attention that loads are distributed evenly among all Stat Server nodes. In an environment with a large number of clients—such as Interaction Workspace (IWS), where each agent has an IWS and the number of such clients is as many as the number of agents), this is no simple task. Another problem is the number of ports that can be opened on single network interface is limited to 65,535. In theory, this limit could be reached in a deployment with comparable number of agent desktops. To further complicate matters, all nodes of a Stat Server cluster must be installed on the same host.

This document describes a configuration workaround that resolves both problems.

Figure 1: Connectivity in standard configuration

Configuration of client connectivity with load balancing

Generally speaking, this approach is based on a well-known principle when all clients use the same Fully Qualified Domain Name (FQDN) of the server’s host and receive real and different IP addresses from the DNS service—the entity responsible for balancing connections. The problem with this approach, however, is that all Stat Server nodes end up sharing the same IP address (but with different ports)—this is because all Stat Server nodes must run on the same host.

To overcome this obstacle, perform the following steps:

1. Add multiple IP addresses to the TCP/IP properties on the host for the Stat Server cluster. The number of IP addresses must equal the number of Stat Server nodes that you plan to run on the host. (The explanation of how to provision multiple IP addresses on the same network interface is beyond the scope of this document.)
2. Configure all Stat Server nodes in accordance with rules described in the Configuring Stat Server Applications as Part of a SIP Cluster section of the Stat Server Deployment Guide. Each Stat Server node should have a port ID of default with a different value.
3. Configure an additional “access point” port (the name of this port is unimportant) on every Stat Server node, and specify for it a number that must be the same for all “access point” ports in Stat Server nodes.
4. For the access point port of each Stat Server node, set the Transport Protocol Parameters on the Advanced tab to address=<dedicated IP address of node>, where <dedicated IP address of node> is the unique IP address created at Step 1 for the node.

Figure 2: Connectivity in configuration with unique IP addresses and DNS

5. Create an “access point” Stat Server application (no actual Stat Server executable will ever call this application) , and specify a new FQDN as its host name. Specify a port with the common port number configured for the “access point” ports of the Stat Server nodes (from Step 4).

Figure 3: Port configuration in the fake Stat Server Application

6. Provision the DNS service to resolve the FQDN from the “access point” Stat Server application host into the IP addresses added on Stat Server host at Step 1.

Figure 4: Connectivity in configuration with unique IP addresses and DNS

As a result, all clients use the same Stat Server application for connection, but in reality they are evenly connected to all Stat Server nodes.

Configuration of primary and backup Stat Server clusters

A backup Stat Server cluster must have the same number of Stat Server nodes as the primary Stat Server cluster. Everything stated in the previous section also applies to a backup Stat Server cluster as well. You must have two physical hosts with two Stat Server clusters and two “access point” Stat Server applications in Configuration Server—one to be used as the primary Stat Server application and the other to be used as the backup.