You can do clustering over the Wan, but you need to consider below prerequisites. And you need to make sure all required ports are opened through the wan. if you meet all prerequisites, switch over will be same.
if your TSP supports, you can have HA on PRI, and it should be done by TSP.
https://www.ciscopress.com/articles/article.asp?p=2743756&seqNum=6
Port Usage
Clustering over the IP WAN
Cisco supports CUCM clustered over an IP WAN. Figure 2-7 shows the publisher and two subscribers at one location while another pair of subscribers from the same cluster resides at a different location. The QoS-enabled IP WAN connects the two sites. Note the requirement of a round trip time less than 80 ms between the sites. This requirement is in support of database replication occurring between the publisher and all the subscribers in the cluster.
Figure 2-7 Clustering over the WAN
Some of the characteristics of this model include:
- Applications and CUCM servers of the same cluster can be distributed over the IP WAN.
- The IP WAN carries intracluster server communication and signaling.
- Limited number of sites:
- Two to four sites for local failover (two CUCM servers per site)
- Up to eight sites for remote failover across the IP WAN (one CUCM server per site).
The cluster design is useful for customers who require more functionality than the limited feature set that is offered by Cisco Unified SRST. This network design also allows remote offices to support more IP phones than SRST if the connection to the primary CUCM is lost.
Design Guidelines for Clustering over WAN Deployment Model
Although the distributed single-cluster call-processing model offers some significant advantages, it must adhere to these strict design guidelines:
- Two CUCM servers in a cluster must have a maximum round-trip delay of 80 ms between them. Because of this strict guideline, this design can be used only between closely connected, high-speed locations.
- A minimum of 1.544 Mbps (T1) of bandwidth is required for Intra-Cluster Communication Signaling (ICCS) between each site and every other site that is clustered over the WAN. This bandwidth supports up to 10,000 busy hour call attempts (BHCAs) within the cluster. The BHCA represents the number of call attempts that are made during the busiest hour of the day.
- In addition to the bandwidth required for ICCS traffic, a minimum of 1.544 Mbps (T1) of bandwidth is required for database and other inter-server traffic between the publisher and every subscriber node within the cluster.
- Up to eight small sites are supported using the remote failover deployment model. Remote failover allows you to deploy one server per location. (A maximum of eight call-processing servers are supported in a cluster.) If CUCM fails, IP phones register to another server over the WAN. Therefore, Cisco Unified SRST is not required in this deployment model (although it is supported). The remote failover design may require significant additional bandwidth, depending on the number of telephones at each location.
This round trip delay 80 ms threshold is default/fix or we can change this threshold to prevent frequent splitting in between Publisher & Subscriber? This is the recommended round trip delay between publisher and subscriber to avoid db replication issues when u do the clustering through WAN. This value is not a setting on CUCM.
subscriber to publisher, every configuration that have been made on subscriber will be shown/implement on publisher after coming back or we need to reconfigure on publisher..? its recommended to make configurations changes on publisher. what ever configurations you make on publisher will be replicated to subscriber if DB replication is good. if publisher is down you cannot make any administrative changes Only end user facing features can be modified.
1- Is Cisco BE6k gives application base redundancy? if publisher goes down phone will work with subscriber(based on configuration) and vice versa. if publisher goes down u can make limited changes from subscriber until the publisher is up.
2- (2 xBE6k & 2 xISR 4431 is the number devices) In case of Publisher failure then ISR 4431 router will work as a phone registrar with SRST or Subscriber will be take over as a redundant? and what is recommended scenario? That's based on your configuration. Normally we configure the phones to fallback on subscriber. In case of both publisher and subscriber down and for added redundancy u can think of SRST on router.
3- Which interfaces are involved OR recommended for ISR 4431 to configure HA? either to use "NIM-1GE-CU-SFP" OR we can use LAN ports via L3?
when u think VG Redundancy, your ISP should also do HA. example if you have two PRI and one PRI connected on each gateway, in this case ISP should send calls on both the PRI. second case if you have only one PRI, u need to manually shift the lines.
