CUBE High Availability: Clustering with Load Balancing (how does it work for ISR?)

[chotima095]

I know that Clustering with load balancing provides local and geographical redundancies. Load balancing can be achieved using Cisco Unified SIP Proxy (CUSP) or using a Service Provider (SP) call agent.

I understand that CUBE Box-to-box redundancy uses the HSRP protocol to form an HSRP Active/Standby pair of routers. The Active/Standby pair share the same virtual IP address and continually exchange status messages. CUBE session information is check pointed across the Active/Standby pair of routers. This enables the Standby router to immediately take over all CUBE call processing responsibilities if the Active router should go out of service for planned or unplanned reasons.

But how does the Clustering with Load Balancing (for ISR) really work when one active pair is down? Much Appreciate if anyone could suggest me this please?

[R0g22]

Irrespective of the fact whether the CUBE is on an ISR or a ASR, HA should work the same way. Can you be a bit more clear when you say how does it work for ISR ?

[Dennis Mink]

you would necessarily need HSRP for this to be redundant from an outbound perspective, you could put both cubes in a RG and do a round robin between the two, make sure you turn OOB options ping on on the trunk to enable keepalive

[Ayodeji Okanlawon · Accepted solution]

Dennis, no this is incorrect. When deploying CUBE HA, you use a single sip trunk and you point that to the virtual IP defined in either your HSRP config or redundancy group.

[chotima095]

Thanks Ago.

[Ayodeji Okanlawon]

Clustering with load balancing using a load balancer such as F5GTM would work as any other application. The intelligence is on the load balancer. You will usually define a health monitor for servers or server pools using certain criterias ie port to monitor etc. When the load balancer detects that port is down, any request for the wide IP assigned to that server pool will only be sent to the active server in the pool. It's that simple

[Dennis Mink]

Thank Ayo for stipulating that, I am not sure if there is a load balancer in the mix though. so in your opinion what is the benefit of using CUBE in a HA deployment in conjunction with HSRP, as opposed to having the cucm do the LB using RGs, like I suggested?

[chotima095]

You're right Dennis, they finally decided not to implement HA for CUBE...

[Ayodeji Okanlawon]

Hi Dennis,

These are two very different solutions. The main advantage of using CUBE with HA is media and signalling check pointing. With CUBE HA using HSRP ( on ISRs) and redundancy groups on (ISR4k and ASR) you actually get a copy of the media and signalling traffic on the standby gateway in the event that the active gateway fails. With CUCM using multiple SIP trunks, you only get redundancy for call setup. Your actual media is not preserved.

CUBE box-to-box HA is very powerful and useful. It also comes in very handy when you need to do maintenance on one of the cubes as your medial is preserved hence you dont have to worry about calls dropping during maintenance.