Geographic Separation of Active and Standby Cisco MGC Hosts Feature

The Geographic Separation of Active and Standby Cisco MGC Hosts feature is described in the following sections:

Related Documents

This document contains information that is related strictly to the Geographic Separation of Active and Standby Cisco MGC Hosts feature. The documents that contain additional information related to the Cisco Media Gateway Controller (MGC) are listed below:

Publication Date	Comments
August 28, 2003	Final version of this document.
May 19, 2006	Added information regarding XECfgParm.dat that should be modified to support geographic separation.

Release	Modification
9.4(1)	This feature was introduced on the Cisco MGC Software Release 9.3(2).

•

Regulatory Compliance and Safety Information for the Cisco Media Gateway Controller

•

Cisco Media Gateway Controller Software Release 9 Installation and Configuration Guide

•

Cisco Media Gateway Controller Software Release 9 MML Command Reference Guide

•

Cisco Media Gateway Controller Software Release 9 Operations, Maintenance, and Troubleshooting Guide

Supported Platforms

The hardware platforms supported for the Cisco MGC software are described in the Release Notes for Cisco Media Gateway Controller Software Release 9.4(1).

Scope of Documentation

This document describes the Cisco PGW 2200 Geographic Separation of Active and Standby Cisco MGC Hosts feature. This document covers the deployment requirements for remote operation of the Cisco PGW 2200 under the following conditions:

Overview

The Cisco PGW 2200 Geographic Separation of Active and Standby Cisco MGC Hosts feature is the geographical separation of the active and standby Sun Netra hardware host platforms with connectivity over a campus or a wide area network (WAN). This feature provides the basic network environment needed to geographically separate the Cisco PGW 2200.

Requirements

Following are limitations, implications, and caveats of remotely locating (physical separation) the Cisco PGW 2200:

Table 1 Geographic Separation Requirements
Condition	Requirement
Software release version	Cisco PGW 2200 software 9.3(2) or later (with associated operating system and hardware requirements).
Total end-to-end delay, one way	Delay is the length of time it takes to send a message from a source to a destination, such as from Cisco PGW 2200 to SLT, Cisco PGW2200 to GW, or Cisco PGW 2200 to Cisco PGW 2200. Must be less than 150 milliseconds.
Packet loss	Packet loss is defined as missing packets with a message. Must not exceed 1% (preferably, less than 0.5%). Note For packet loss rates below 0.5%, increase the RUDP receive window size (*. rudpWindowSz) to 64 for increased performance.
Bandwidth limitation	Directly related to end-to-end delay. There are no special requirements.
If improved SIP automatic switchover support is enabled...	Both the Cisco PGW 2200s should be part of the same IP subnet.
*.AllISDNLinksFailCausesFailover added in Release 9.5(2)	Controls the severity level of the All ISDN IP Conn Fail alarm. Default: false Valid values: false (alarm severity is Major) and true (alarm severity is Critical). This property should be set to true if your Cisco MGC hosts are in separate geographic locations. You can also set this parameter to true if your system is not processing SS7 calls and you want your system to perform an automatic switchover should all of the ISDN IP connections fail.
*.AllLinksFailCausesFailover added in Release 9.4(1)	Controls the severity level of the following alarms: •All C7IP Links Fail •All M3UA Assoc Fail •All SUA Assoc Fail Default: false Valid values: false (alarm severity is Major) and true (alarm severity is Critical). This property should be set to true if your Cisco MGC hosts are in separate geographic locations. You can also set this parameter to true if your system is not processing SS7 calls and you want your system to perform an automatic switchover should all of the links fail.
Setting for XECfgParm.dat parameter replicator.reconnectInterval	This parameter defines the reconnect interval in number of seconds for the replicator during a switchover. Value: Any integer Default: 15 seconds Note Set this value to 0 for a standalone Cisco PGW. For geographically separated Cisco PGW pairs, the following replicator timer values are recommended: •On one PGW, replicator.reconnectInterval = 15 seconds •On the other PGW, replicator.reconnectInterval = 20 seconds If the timer settings are the same on both Cisco PGWs, the additional latency between the separated Cisco PGWs may cause a problem in which the replicator links are continually reconnecting and then immedately disconnecting. The timer change prevents this problem.

Configuring the Cisco PGW 2200 for Geographic Separation of Active and Standby Cisco MGC Hosts

Configuration of the Cisco PGW 2200 for Geographic Separation of Active and Standby Cisco MGC Hosts is required if the Active and Standby PGWs are located in different subnets. To configure the Cisco PGW for Geographic Separation of Active and Standby Cisco MGC Hosts, follow the procedures for specifying IP Addresses in the chapter "Configuring the MGC Software," located in the Cisco MGC Software Release 9 Installation and Configuration Guide. The section on "Specifying IP Addresses" is provided below and has been adapted for this feature.

Note If the Active and Standby PGWs are located in the same subnet, there is no additional configuration required for Geographic Separation of Active and Standby Cisco MGC Hosts.

Specifying IP Addresses

To specify IP addresses, modify the XECfgParm.dat file parameters listed in Table 0-2.

Note If there are two Ethernet interfaces defined on the Cisco MGC, it is mandatory to have these on separate subnets.

This is not a valid combination because they are on the same subnet. The following example illustrates a valid combination:

When Two Ethernet Interfaces are on the Same Subnet

If two Ethernet interfaces are on the same subnet, then one of them must be physically disconnected from the existing subnet and then connected to a different subnet. The new IP address must be appropriately configured on the system. Refer to the manual pages for the UNIX command ifconfig for more information.

Specifying Next Hop Router IP Addresses for Geographic Separation of Active and Standby Cisco MGC Hosts

To specify next hop router IP addresses, modify the XECfgParm.dat file parameters listed in Table 0-2 and do the following procedures:

Step 1

Determine whether the Active and Standby hosts are in different subnets. Type the following command and press Enter:

If the Active and Standby Cisco PGWs are located in different subnets, go to Step 2.

For the above example, the following next hop address should be configured by editing the XECfgParm.dat parameter as follows:

Step 2

Configure the next hop router IP address. Use an editor such as vi to configure the next hop router IP address by editing the *.IP_NextHopx in the XECfgParm.dat file.

Note You can configure *.IP_NextHopx before starting the PGW while configuring the IP address in XECfgParm.dat.

Note Make sure that the next hop router can be reached from both Active and Standby hosts.

XECfgParm.dat file.parameters for Specifying IP and Next Hop Router IP Addresses

Table 0-2 XECfgParm.dat File Parameters (Specifying IP and Next Hop Router IP Addresses)
Parameter	Modification
*.ipAddrLocalA	Enter the first local IP address; used for checkpointing and switchover heartbeats. Caution This address is the same value as .IP_Addr1, and is the hme0 interface. Caution* No other machine on the network should have *.ipAddrLocalA set to 0.0.0.0.
*.ipAddrPeerA	Enter the first corresponding peer IP address; used for checkpointing and switchover heartbeats. Note If you have two Cisco MGC hosts in a fault tolerant configuration, this value is set to the IP address of the second host.
*.ipAddrLocalB	Enter the second local IP address; used for checkpointing and switchover heartbeats. This is the address of the hme1 interface. Note If your configuration does not use a secondary Ethernet adapter, leave this address set to the default value, 0.0.0.0.
*.ipAddrPeerB	Enter the second corresponding peer IP address; used for checkpointing and switchover heartbeats. This is the address of the hme1 interface on the second host. Note If your configuration does not use a secondary Ethernet adapter, leave this address set to the default value, 0.0.0.0.
*.IP_Addr1	Enter the IP address of the hme0 interface.
*.IP_Addr2	Enter the IP address of the hme1 interface (if configured).
*.IP_Addr3	Enter the IP address of the hme2 interface (if configured).
*.IP_Addr4	Enter the IP address of the hme3 interface (if configured).
*.IP_NextHopx	Enter the next hop router IP address. The next hop router IP address is determined through the .IP_NextHopx parameter. This parameter allows you to assign eight different next hop router addresses that are different for the active and standby PGWs. The .IP_NextHopx parameter (where x is any number from 1 to 8 because there are eight entries in XECfgParm.dat) is used with the active and standby PGWs that are on different subnets and therefore have different IP addresses for the routers to get to a given destination. For every destination that needs a different next hop router, an unused .IP_NextHopx parameter should be chosen. The appropriate IP address to reach the destination should be entered in each PGW's XECfgParm.dat. If a given destination is on a locally connected subnet of one of the PGWs, the .IP_NextHopx parameter for that PGW would be 0.0.0.0. Default value: 0.0.0.0.

Table Of Contents