Cisco Cisco IP Contact Center Release 4.6.1 Leaflet

Page of 428
 
3-23
Cisco Unified Contact Center Enterprise 7.5 SRND
Chapter 3      Design Considerations for High Availability
Cisco Multi-Channel Options with the Cisco Interaction Manager: E-Mail Interaction Manager (EIM) and
Additionally, the application server of Cisco Interaction Manager establishes a connection with the 
Unified CCE Administration Workstation (AW) database server to import relevant configuration data 
and to map the configuration to Cisco Interaction Manager objects in the Cisco Interaction Manager 
database. Note that Cisco Interaction Manager does not make use of the Configuration API (ConAPI) 
interface.
When Cisco Interaction Manager is integrated with Unified System CCE, the multi-channel controller 
of Unified System CCE is installed on the services server. 
In parent/child configurations, there is no multi-channel routing and integration through the parent ICM. 
Media Routing PGs need to connect to the child or Unified System CCE. A separate Cisco Interaction 
Manager or partition is required for each child.
Likewise, in hosted ICM/CCH environments, there is no multi-channel routing through the Network 
Application Manager (NAM) layer, and integration is at the individual Customer ICM (CICM) level 
only. The Media Routing (MR) PGs need to connect to the CICM.
High Availability Considerations for Cisco Interaction Manager
The Cisco Interaction Manager offers high availability options using additional web and application 
servers and using load balancing equipment to distribute agents and contact work more evenly across 
the platform as well as to provide for failover in redundancy models.
Load Balancing Considerations
The web service component of a Cisco Interaction Manager deployment can be load balanced to serve a 
large number of agents accessing the application at the same time. The web (or Web/Application) servers 
can be configured behind the load balancer with Virtual IP, and an agent can access Cisco Interaction 
Manager through Virtual IP. Depending on the selected load balancing algorithm, the load balancer will 
send a request to one of the web/application servers behind it and send a response back to the agent. In 
this way, from a security perspective, the load balancer serves as a reverse proxy server too.
One of the most essential parameters for configuring a load balancer is to configure it to support sticky 
sessions with cookie-based persistence. After every scheduled maintenance task, before access is opened 
for users, Cisco recommends verifying that all web/application servers are available to share the load. 
In absence of this, the first web/application server could be overloaded due to the sticky connection 
feature. With other configurable parameters, you can define a load-balancing algorithm to meet various 
objectives such as equal load balancing, isolation of the primary web/application server, or sending 
fewer requests to a low-powered web/application server.
The load balancer monitors the health of all web/application servers in the cluster. If a problem is 
observed, the load balancer removes the given web/application server from the available pool of servers, 
thus preventing new web requests from being directed to the problematic web/application server.