slagburn

What the heck are you building that you need that much redundancy? Or are Canuck COs not quite hitting seven 9s on the reliability scale? :P

Addatel

I know you can set up (for a fee) to have all the lines call forwarded in a disaster recovery situation from another CO. Limitation would be the type of CO equipment as some of the more remote units are still not upgraded. Some can't even handle T1 yet like Langley, Tsawwassen, etc... But I think you are talking Vancouver proper only right? If that is the case, I think they can set up an auto detect of the PRI failing at the far end and auto switch to another PRI in your network. In the same sort of way as the last challenge posted here. I will make a few calls tomorrow, as you have me intrigued. I know a few companies that did set up the disaster recovery forwarding, and tested it to death.

papa-bear

Well, there is 'need' and "need". It is more the issue of a single point of failure. Right now a majority of our PRIs (5) come into one site with one DID block and another batch (2 PRI) into another site with another block.

These two sites are currently tied together with an E1 through ITG linecards so we can get outgoing dial tone from either one if need be.

Problem is, if a fibre seeking backhoe comes along and takes out the conduit to either side we loose conectivity to the PSTN and between the sites. (single point of access to each facility).

As we move forward with an IP telephony solution (almost trippling out user base as we would be replacing many Centrex lines) I would like to distribute the PRIs thorugh out the network (two sites minimum, maybe 3 or 4) to provide capacity to keep calls coming in in the event of one site being disabled for some reason.

The problem with most IP designs I have seen for failover is to provide dialtone for outgoing calls only. There is no way to keep your incoming calls available if the data centre where the PRIs live gets cut off from the world. Management hears that there is dialtone available and assumes it is incoming calls as well so I need to correct them and let them know there is no (inexpensive) solution.

Okay, I am being over cautious but I want to be able to offer solutions to as many potential issues as possible and let the bean counters decide which option they want. Don't like them coming back to me saying that I did not give them all of the options, it makes life miserable

papa-bear

Yes, I know the local telco's offer a product called "ServiceSaver" and you pay through the nose for each DID you want to have a pre-plan in place for to be able to re-direct in the event of a failure.

I am feeling things out to see what others may have been able to get their carriers to provide on the CO side to re-direct DIDs between them in the same way they do it within a CO between PRIs.

If I can find someone who has been able to have the service provided I have more leverage when asking them about it.

jack

Not sure about Telco's in the Great White North but down this way the Telco's offer service via a Sonnet ring.

At last gig had a Sonnet that connected 3 sites in the city to a fourth out in the burbs -- about 50 miles out.

This ring went thru three different CO's and provided connectivity to any site from any CO providing the type of redundancy that your describing.

Although the install wasn't exactly cheap the ongoing and the future provisioning were cheaper and with quicker install times.

Again not sure if this is offered up your way but it can not hurt to ask.

My issue with the Service Saver deal is that if the CO goes so does your service as well as your backup plan, since its is based at the CO. Would really suck to have to explain why after paying for yeras the first time you needed your DR it didn't work.


BTW that single entry point stuff needs to be addressed, a true problem waiting to happen.

Addatel

Hi Rstabler,

Here is an answer to your question speaking specifically about the fraser valley lower mainland. What they mean by "switch" is in their network so it can go from CO to CO, even to the island they say.

Let me know if this is what you are looking at as a plausable solution. I can get more detailed info if you wish.

Answer from provider:

Firstly, our Sonet ring works for voice, but that is only if Fibre is cut between buildings, not if a c/o goes down.



For PRI’s we have two things, firstly, there is a feature called ‘overflow’ this means if the call coming in gets a busy signal (or circuit is dead), the call will overflow to another number, this other number has to be on our switch. So this could be the best option, have each PRI set up in a piggy back configuration so that it will always find a free pri. The only time someone would not want this is that if everything is working fine, but the customer would normally get a busy signal, this way the call would go to the next PRI. For some customers, this is actually great as it means their customers never get a busy signal.



The other option, is not as good, as it is manual and this is our re-direct product, the customer would have to call it in though.



Do you think the first option would work?

papa-bear

Thanks Addatel,

The first option would probably be the best but still has the down side of a single point of failure in the CO. Our biggest concern would be the fiber being cut from one of the feeding sites to the WAN where everyone would lose dialtone (except for the emergency 1Bs in place).

I'll bring the idea up with a few of our vendors and see what they think.

No problem. They actually mean switch is anywhere within the network. So any of the on Net CO's, just not the Telus CO's. So it can still be routed to another CO. If it gets a busy signal at CO A it would then route through CO B. It just has to be a CO they control, but the master "switch" will route to the other CO in a failure. It's Allstream if you were wondering.

Spice

The Media Gateway 1000T might be able to help here. It provides route diversity but I'm not sure how it handles DID blocks. You can put a gateway in a remote building and attach up to 20 x PRI. You also need an IP network back to your main switch. It also requires your switch to be running Succession Release 4 software.