Cisco Cisco E-Mail Manager Unity Integration Option Dépliant

Pagina di 428
 
3-30
Cisco Unified Contact Center Enterprise 7.5 SRND
Chapter 3      Design Considerations for High Availability
Peripheral Gateway Design Considerations
Unified ICM Redundancy Options
Duplex/Redundant Unified ICM servers can be located at the same physical site or can be geographically 
distributed. This applies specifically to the Central Controller (Call Router/Logger) and Peripheral 
Gateways.
Under normal operations, the Unified ICM Call Router and Logger/Database Server processes are 
interconnected through a Private Network connection that is isolated from the Visible/Public Network 
segment. These servers should be configured with a second NIC card for the Private Network 
connection, and the Private connections should be isolated from the rest of the Visible/Public Network 
in their own Cisco Catalyst switch if they are located at the same physical site. If the Central Controllers 
are geographically separated (located at two different physical sites), under normal operations the same 
Private Network connections must continue to be isolated and connected between the two physical sites 
with a separate WAN connection. For normal operations, this Private Network connection should not be 
provisioned on the same circuits or network gear as the Visible/Public Network WAN connection 
because that would create a single point of failure that could disable both WAN segments at the same 
time.
The Unified ICM Peripheral Gateway duplex pair of servers is also interconnected through a Private 
Network connection that is isolated from the Visible/Public Network segment under normal operations. 
If the two sides of the duplex pair (Side A and Side B) are both at the same physical site, the Private 
Network can be created by using an Ethernet Cross-Over Cable between the two servers to interconnect 
their Private Network NIC cards. If the two servers in the duplex pair are geographically distributed 
(located at two different physical sites), the Private Network connections must be connected with a 
separate WAN connection between the two physical sites.   This Private Network connection should not 
be provisioned on the same circuits or network gear as the Visible/Public Network WAN connection 
because that would create a single point of failure that could disable both WAN segments at the same 
time.
For additional details on the ICM network requirements for this connection, refer to the Unified ICM 
Installation Guide
, available at 
For additional details on the Unified ICM network requirements for clustered over the WAN, see the 
section on 
Within the Agent PG, two software processes are run to manage the connectivity to the Unified CM 
cluster:
JTAPI Gateway
The JTAPI Gateway is installed on the PG by downloading it from the Unified CM cluster at the 
time of the PG installation. This ensures compatibility with the JTAPI and CTI Manager versions in 
the system. Note that, when either the PG or Unified CM is upgraded, this JTAPI Gateway 
component must be removed and re-installed on the PG.
The JTAPI Gateway is started by the PG automatically and runs as a node-managed process, which 
means that the PG will monitor this process and automatically restart it if it should fail for any 
reason. The JTAPI Gateway handles the low-level JTAPI socket connection protocol and messaging 
between the PIM and the Unified CM CTI Manager.
Agent PG Peripheral Interface Manager (PIM)
The PIM is also a node-managed process and is monitored for unexpected failures and automatically 
restarted. This process manages the higher-level interface between the Unified ICM and the JTAPI 
Gateway and Unified CM cluster, requesting specific objects to monitor and handling route requests 
from the Unified CM cluster.