HA100 + 2x UCM 6510: Concurrently active both E1 PRI and VoIP SIP WAN trunk inbound links from the telco?


#1

With HA100 + 2x UCM 6510: Can the active 6510 have concurrently active both E1 PRI and VoIP SIP WAN trunk inbound links from the telco (in one hunt group)?

If not, what about in the absence of the HA100?

Answer not found in manual or in


#2

if you do not receive advice, you may want to open a ticket to Grandstream Support:


#3

I am not sure I follow, but the E1 provider and VoIP provider are presumably different providers. You specifically referenced trunk inbound links from the telco and a hunt group.

The UCM has no control over a provider and how they may send or route calls to the UCM. If the concern is truly at the provider level, then you will need to confer with them.

The UCM merely sees a call as it comes in from the provider, looks at the DID portion of the message and then determines if the number is valid and if so routes same according to the inbound rules that were set for that DID. If the number is not found (it must be an exact match or a wildcard accept all), then the UCM will reject the call and return a 404 not found message back.

If your concern is about the failover aspect and if both services will continue to function, then refer to -
https://documentation.grandstream.com/knowledge-base/ha100-deployment-guide/


#4

Our HA goal is to remain operational whether a telco link or a UCM6510 fails to be operational. The UCM6510 does not support two SIP VoIP links nor two E1 PRI links, but it does have ports for one of each.

QUESTION:Can we exploit the one-of-each?

  • NO;
  • YES, for failover with one link active and one link passive;
  • YES, with both links concurrently active

We currently have two independent UCM6510's, with no HA100, each with its own E1. The telco balances calls across the links. If one UCM6510 powers down, the telco notices the E1 interface is down, and sends all traffic to the remaining E1 attached to the other UCM6510.

However when the failing UCM6510 is powered up but is not responsive (hung), then

  1. The E1 interface is electriclally up,
  2. so the telco doesn't notice the outage,
  3. so the link still gets half the calls,
  4. which never get answered.

QUESTION: From the Grandstream UCM6510 perspective, can both the E1 PRI and the VoIP WAN inbound links be concurrently active?

QUESTION: Does the HA100 change the answer? From the Grandstream UCM6510 HA100 perspective, can two inbound links be concurrently active. one being an E1 PRI link and the other a WAN SIP VoIP link?

Obviously, we would be transitioning from two independent UCM6510's to an HA100 + 2x UCM6510 configuration. The active UCM6510 would normally be handling all the work.

Single or dual providers: I think that for both inbound links to have the same well known phone number, that both the E1 and the WAN VoIP would have to come from one telco. We would prefer to have two independent providers, but don't think that will be possible.


#5

There is a terminology gap, but yes, the UCM can support multiple SIP providers at the same time. The SIP VoIP link that you refer to is nothing but an Ethernet connection. The 6510 has a built-in PRI gateway which is used to connect directly to the provider’s PRI device. It can also accommodate an external PRI gateway as well which converts the PRI to SIP so that it too looks like a SIP trunk.

Assuming that the providers remain up and the active UCM fails and the HA fails over to the secondary, then the providers will still be active on the secondary. The HA will accommodate the internal E1 as well as the SIP providers to include the external PRI gateway to SIP device if so equipped…

It is not clear to me about the hung part and how the follow-on question relates to one another. If it is not responsive, then I assume that the E1 port and VoIP/SIP are also not responsive. Perhaps there may be something that causes the PRI provider to think the UCM is up, but surely the UCM is not responding with any signaling to either the PRI or SIP.

The second question is answered in the document and in my original reply. Yes, the UCM is not confined to one or the other provider, but can handle both provided that both are operational.

The whole point of the HA is to provide redundancy for the PBX in case of a failure to the primary. How you handle a failure of a provider or ISP is a different matter.


#6

Thank you for the reply.

I try not to make unnecessary assumptions. The response and documentation make it clear that the UCM6510 supports both WAN Ethernet and E1 PRI, but does not say that both can be operational concurrently. In fact, the 6510 User manual pdf page 62, states the opposite, "UCM6510 can also connect to network via E1/T1/J1 data trunk instead of using the WAN/LAN ports. Please see section [DATA TRUNK] for more details."

Not finding more about exclusivity in the [DATA TRUNK] section, I asked here.

Regarding "hung": We have had experiences where the PBX was not responsive, but the PRI E1 adapter appeared to be active as far as the provider could tell.

If we do not implement the HA100, we may develop our own external health check mechanism, and if one PRI-PBX pair dysfunctional, send a command to an intelligent PDU to remove AC power from that PBX.


#7

The wording used is that The UCM “can also connect to network via E1/T1/J1 data trunk instead”. The meaning here is that it need not have a SIP trunk to function, but can “also” function using the E1/T1/J1 as standalone.

I have 6510s using both T1 and SIP trunks simultaneously.

If the provider is sending calls, and the UCM is “hung”, and the calls are not being answered, then more investigation is needed to understand the scenario as to why the provider thinks it is up and exactly what conditions of “hung” work and do not work.

I have not experienced such using either standalone 6510s or those using HA.