Correct, failing over from one carrier to another for inbound calls is a provisioning action, not a routing action. Now, there's nothing to say that market pressures could not force the adoption of some administrative protocol that would permit a pre-configured, hair-trigger port or RCF transaction that would be fired off by a blurry-eyed customer engineer at 2am. Inter-carrier, inbound failover doesn't exist in the PSTN either, so architecturally the SIP network not lacking. The PSTN simply assumes that carriers as a whole do not fail, so in actual realization parts of the SIP network might be a bit behind the PSTN. Intra-carrier, inbound failover is a different matter. If your carrier has multiple proxies in multiple locations, its more likely that you'll be able to cut a deal to get your inbound calls from NY when LA slides into the sea. David On Mon, Aug 24, 2009 at 5:40 AM, Alex Balashov<abalashov at evaristesys.com> wrote:

Insofar as numbers can be ported only to one LRN per OCN per LATA at a time, there is no "provider redundancy" -- SS7 is not like BGP. Whatever redundancy exists must exist within your carrier's network architecture.

J. Oquendo wrote:

...

Thanks to all who've answered I guess I could have been more specific so apologies.

What are some options for inbound redundancy. Outbound is not an issue as I can swap off from my end on my routes out however, what does one do for inbound failures. I *heard* from my carrier that an option would be designed (sales talk mainly) for inbound redundancy. I would think a top level carrier (don't want to get into definitions of Tier1, etc.) - I would think they'd have a back-out change management plan on hand, but that to me has proven to be non-existent. (Replacing an entire switch to notice it failed 5 hours later at the start of a business day is not cool).

My thoughts, review the SLA's and come up with a MTD (Maximum Tolerable Downtime) and the whole shtick of wording, payments, etc in which they'd likely shrug at - at the end of the day. From my standpoint (engineered):

Redundant connectivity (check) Redundant equipment on hand (check) Redundant outbound links (check) Redundant inbound links (sort of checked)

On my inbound links, I've DID's across carriers, but this does not (as we know) provide redundancy for me when one inbound carrier does fail.

"Hi Global?, can you take these L3 DID's for me. I have their engineers ready to shoot you 4 million minutes in traffic until they get their act together. k thanks!"

Wish it worked that way.

-- Alex Balashov - Principal Evariste Systems Web ? ? : http://www.evaristesys.com/ Tel ? ? : (+1) (678) 954-0670 Direct ?: (+1) (678) 954-0671 _______________________________________________ VoiceOps mailing list VoiceOps at voiceops.org https://puck.nether.net/mailman/listinfo/voiceops

Another thing to consider is that in the BGP world, the IP block(s) announced by the customer are actually "owned" by the customer, in an administrative and a technical sense. This is true whether they are subnets of a provider's aggregate, or provider-independent blocks obtained directly by the customer through a RIR like ARIN. The ISPs really do just provide the transit. You can announce whatever blocks you want to them (that they let you/don't filter), and they can announce those blocks on behalf of their subtended networks (customers). PSTN numbers don't work that way, conceptually. You may, as a customer, "own" the numbers in some legally significant way, but as implemented in the underlying technology, they are homed (or ported) to a carrier switch. That's just where they go, and that's what makes them functional. When it comes to multi-provider redundancy, I think the eventual solution - if there's any interest and business case for one - will be less to hair-trigger NPAC subscriptions to another OCN/LRN, protection switch-style, and more the use of some sort of "logical" number or identifier that is a wrapper for the underlying numbers that are still statically assigned to a particular carrier. Sort of like a hunt group or find-me-follow-me functionality, but on a lower level of abstraction. :) -- Alex -- Alex Balashov - Principal Evariste Systems Web : http://www.evaristesys.com/ Tel : (+1) (678) 954-0670 Direct : (+1) (678) 954-0671

David Hiers wrote:

Since there is no problem that cannot be solved by an additional layer of abstraction, I propose a Universal Title Translation that would overlay the Global Title Translation.

I leave it to some physics wonk to pencil up a Multiverse Title Translation...

Oh, I think there's no question that it's absurd. I just see it as a more likely solution for the industry to pursue. I don't see Neustar countenancing OSPF-speed re-convergence in the NPAC. :-) In general, I think this concern about inbound redundancy is overblown and unnecessary, although all sorts of fantastical "total redundancy" schemes are a frequent topic in meetings with delusional Pointy Haired Bosses. If the voice portion of Level3 as a whole blows up, you've *probably* going to have way, way bigger problems to worry about than your DIDs not working. -- Alex Balashov - Principal Evariste Systems Web : http://www.evaristesys.com/ Tel : (+1) (678) 954-0670 Direct : (+1) (678) 954-0671