No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

SIP Trunks: Goodbye to an Old Era?

Cutting the Verizon cord was a long mulled-over decision. After weighing the pros and cons, I concluded that the time had come for change. This doesn't necessarily mean that I endorse SIP trunks as the solution for every SMB, because not all SMBs have access to qualified services.

Our Panasonic IP-PBX is certified with just three providers: Broadvox, cBeyond and XO Communications. We are using Broadvox for our needs because they are the only fit, since we don't have T1s or PRIs; we have FIOS. SMBs must ante up good reliable links to take advantage of SIP trunking and I don’t think anything beats FIOS (except MPLS), and I mean that with the highest compliment to Verizon.

Saying goodbye to Verizon voice is not because the CustoPak (Centrex) service wasn’t good or didn't offer value--it was/is good reliable service that offers basic tried and true features. It’s simply too expensive and the savings with SIP trunks are hard to ignore. Maybe it's "The Allure!" Verizon never delivered on their rumored services for FIOS customers in the form of IP or SIP service offerings, and when the Verizon customer service rep asked why I was disconnecting service I didn't bother to tell her the same story again. Since we had FIOS and our CustoPak (Centrex) changed over to fiber (FIOS) in 2005, we had no significant issues other than Verizon made a change from PPP to DHCP and we weren't notified, and then when we replaced our router we found on Verizon's side that the ARP cache does not refresh for two hours and no matter what, our new gear would not work for two hours. It's still fair to say because we have FIOS that we enjoy better than average service that most SMBs don't enjoy.

In It's the Economy, Stupid: SIP Trunks & Hosted vs. Premise, I noted some elements of fear. I'm not afraid to go after the provider, the factory guys or any vendor. Some exposure to SIP trunks helped me decide, as did time spent logging/discovering the reasons for past outages during the last 12 months and then pestering the provider about the configuration of the services and RFOs (reasons for outage). This all led me to believe the same rules of successful service apply--don't let them wear you down; instead you must wear them out. My decision isn't what I'd call popular among Interconnects and as Sandman of Chicago is quick to point out, the past misgivings and failures of ITSPs maybe indicative of the future. Even so, the cost benefits are too hard to ignore and I really think SIP trunks will take hold, but it's going to take much longer than what industry moguls think.

I also mentioned in The Dealer/VAR World of SIP Trunking that, "the savings are too substantial to ignore." Maybe I'm a miser but I've been paying to have cheap LD with the SIP trunks I originally wrote about, and we did save money. The real savings is in the bundle--and that amounted to 33% for us in this new implementation. After install costs our payback is less than two months.

Once the paperwork was submitted to Broadvox, a steady flow of communications ensued. Shortly afterwards Broadvox emailed a letter stating, "In accordance with the FCC requirements that were listed in the email you received when your SIP Trunking service was deployed, please find 911 labels enclosed with this letter. The 911 labels must be placed on or near equipment used in conjunction with this service."

No stickers are going on my phones--my box will route 911 or 9+911 calls to the analog line--yet another reason why LCR/ARS (least cost routing/automatic route selection) is not obsolete. Broadvox also noted in their letter, "Calls to 911 for testing purposes prior to trunk deployment will be assessed a non-negotiable minimum $50 fee per call if the number placing the call is not registered." So I called customer service to verify that registration was completed before testing 911 using the SIP trunks. There were two additional pages discussing the limitations of E911 using SIP trunks.

When an email arrived with a FOC (firm order confirmation), I wondered if the cut would happen on the date, and according to the email the estimated time to port the Verizon numbers was 11:00 am and the actual porting took place around 1:45 pm, and I didn't notice any disruption in calls. Later, I called Verizon and disconnect orders were in the system for our CustoPak service. By the next morning the Verizon lines were disconnected. Granted, all I did was make minor changes to our IP-PBX and the new SIP trunks were up and running in a few minutes. But I'd like for all the die-hards and anyone else to stop and think about this. My order was received with all necessary paperwork on a Monday after 4:30 pm and the trunks were ready for testing 48 hours later. Next, we were up and running in a few minutes using our new SIP trunks –not too many Telco customers can claim this unless it's getting your bandwidth increased (usually instantly for FIOS) or simple feature activations but most other changes to your services still takes days and weeks, or longer. Still, I'm not going to ignore 33% savings and I’m not one that usually gets lured by cost savings alone.

Failover for us is pretty simple--because our calls hit the cell phones. With any failures from our IP-PBX (Panasonic NCP1000), the Internet (Verizon FIOS), the ADTRAN router (1335) or the SIP provider (Broadvox) re-routes our calls to the cell network (AT&T). Normally I would use POTS or analog phone lines as the backup but we are usually in the field. I still have an analog line just in case, namely to make outbound calls and for 911 but I don’t look at hanging on to that forever. So for now a few things ring true for me –the PSTN remains as a backup. Wireless is another backup mechanism in case you missed some neat tricks using wireless gateways on your IP-PBX that I discussed in SIP Trunking: When Disaster Strikes.

In testing the failover, I started at the IP-PBX by putting the gateway out of service. In this instance the gateway still sends packets to the SIP server in a busy out condition so failover will not invoke and callers will instead hear a busy tone. The easiest test is to disconnect the LAN connection to the IP-PBX gateway--then failover occurs very quickly. A NOC technician at Broadvox also told me that in disaster conditions they would put a hard forward to a provided number, which is good to know. The important difference over the old world telephony is that I have easy access to the guys and gals running the show in the Broadvox NOC and I will pay homage to them too because they've been pretty responsive with my order (thus far).

The biggest changes were printing new desi's (labels) for all the phones. With the DID numbers and some ACD Incoming Call Distribution group tricks I ended up getting all our bases covered and emulating what we used CustoPak features for; what changed is how--but the end result is all that counts. So I'll add my Interconnect comment to the IT devils that forsake this golden rule--if you can't do what the customer wants, rethink what you're doing. As Interconnects we are always bending over backwards to recreate what customers have and the bad part of this is that we sometimes fail to see new processes or possibilities of doing something different that achieves the same end result.

The second change was creating phantom numbers in our ACD call groups to delay-ring cell phones and this is pretty cool. We get ACD reports on all calls including calls to the cell phones. The offsite ringing capability requires an enhanced feature license that we purchased with the system since it was originally in our test bed. Unlike the old days, bridging two POTS lines was never a good idea. TBCT (Two Bearer Channel Transfer) or this same concept in IP is seemingly a dead horse and the providers are anti-competitive when it comes to providing an IP-PBX user/customer with any basic features. A few days after this implementation I received the news release from Broadvox that they are now offering hosted IP-PBX services.

When calls arrive on the new SIP trunks, after a brief delay of 1 ring cycle, the IP-PBX dials out and also connects to the cell phones unless the calls are answered. Audio quality is good and it's been transparent to the calling party. So, I confess that this works great but I still don't like using two call sessions when the industry needs to use just one for this purpose. We also have another trap using a timer and that is to intercept calls after so many ring cycles and answer with the auto attendant as backup. The mailboxes are also set to forward our voice mail messages to our iPhones. That pretty much sums up our call coverage.

While we've given up our analog lines it doesn't mean we've given up our devices. Our internal gear using 2500 ports works great with the SIP trunks as it did before. Our outbound faxing always worked well and I attribute this to Panasonic. I think there should be excitement for the potential that SIP trunking offers but reality is going to dampen many efforts by the SMB because the transport still remains key and not having a rock solid link is going to cause more grief than reward.

Other than the 33% savings, we gained what I think are new benefits. For one, when and if we were ever to move, unlikely as it is, the move is limited by availability of reliable bandwidth--not necessarily by constraints of an address, proximity to the Central Office or Telco LATAs to preserve old telephone numbers. I've wanted a converged pipe because I do think it makes sense for SMBs and large enterprise but having the best transport is key. I'm not overly confident in cable or DSL so I'm likely to pass on these deals because I think they will generate too much service work for too little reward, garnering bad image. Implementing SIP trunks doesn’t mean businesses must sacrifice diverse routing or call quality. In fact, I've noted before that call quality is very good and even when there are noted issues they seem to be at magical hours from 4-5 pm EST. One of the real values of using SIP trunks is the routing diversity over traditional telephony.

Another consideration is a nagging issue for those that are stuck on copper facilities from their Telco. Larger customers with greater mixes of analog copper lines have more service issues and maintenance mainly due to weather, traffic accidents and other terrestrial influences. We've billed numerous hours each year for service calls from these failures. It's a well-known fire drill that the Telcos put every business through--"if we dispatch and find the issue with your gear or wiring then we charge X-dollars per minute." We don’t normally work for free either.

I've pondered this issue too for a long time. While SIP and IP telephony generates more service calls that Eric discusses in The Long SIP Trunking Road, analog services aren’t immune from troubles. The real difference is with IP and or SIP, I don't need to Roll the Truck to the site with the same frequency as I do with analog, POTS and Centrex services. In the analog PSTN world, the trouble must be proven at the demarc and this almost always requires a site visit. Again, as an Interconnect I know that I can make more money by not dispatching to the customer demarc and end up with happier customers by servicing more issues remotely and providing resolutions faster. Large enterprise is motivated by using the data center and one pipe or diverse pipes as it should, but for the SMB SIP trunks offers freedom from an invisible leash and the hassle of changing phone numbers, paying extra fees to maintain work-arounds, and the business is always disrupted more than what anyone wants to admit when an SMB moves to a new location.

We have a good tool in place using ADTRAN’s Voice Quality Monitoring (VQM) to report on call quality. Our past record from a year earlier shows that 70% of our SIP calls are rated excellent and the remaining 30% good. I spoke with engineers at the NOC and they prefer codecs set with first choice G.711 and second choice G.729 and our IP-PBX is setup the same way. In our IAD we have reserved one switch port setup with monitoring to run packet traces whenever the need arises.

Should we ever decide to sell the business, SIP trunking offers cheap portability unlike any of the Telco offerings. We also gained virtual DID numbers not just from the local area but from anywhere we need them and at a price the Telcos can’t touch. Broadvox provides online account data showing trouble tickets in the system and allows customers to add updates, attach files or comments. Why for such a service oriented industry that we are in, have the Telcos never understood and deployed to the masses tools like these?

I have other reasons for my decision and it's not just because I'm a penny-pinching miser that won't pay interest to a bunch of wolves. A year ago after writing SIP Trunks: A Call To the Wild, some more thinking led me down this path. Also, because our phone service is now "IP," when calls ring in there is virtually no delay in system processing and when either party hangs up--the call session is cleared. SIP trunk calls are notably processed faster than analog so call setup and tear down rocks. I've noted before that if you're not counting time then you’re losing sight of a key benefit. You see it's very simple, analog AC ring current traversing DC copper loops takes much longer than a couple of packets to make the phones ring. The FIOS fiber ONT (optical network terminal) is using anlog ports emulating the AC ring current and DC loop and it's still slower than SIP trunks. One of the thoughts that rattles my brain awake at night is whether or not IPT will save significant amounts of energy over old tip and ring. Still, I've a long road ahead but refining and boiler plating what we've done for adoption by some of our customers will be our reward.

I'm adding two additional comments about SIP trunks, again only from my limited experience because the mileage may vary. Regarding call quality, overall I've been very pleased. Unlike the cellular networks, SIP trunks call quality goes bad but doesn't stay bad. The next important issue is service. I've thought about this concept long and hard too and it's the soft virtual demarcation. On a positive note I have more gurus that I can call on to fight one another to resolve an issue than with a traditional hard wire demarc.

What would really make my year is a solution to beat back the cell bills like we have with the phone bills. I know that where we are today with one pipe (it's a very good pipe) that we won't remain static. Bandwidth and time will eventually push us into upgrading the pipe and that is simply a phone call away. But that's minor considering that I know I can provide the IP-PBX in an IAD from ADTRAN or I can fire up another Mac Mini Server and load it with software to act as an IP-PBX to further reduce our footprint. Hosted I could easily do but it's simply not a cost effective change for my company. Is it time to say goodbye to an old era?

Matt Brunk is president of Telecomworx and a No Jitter blogger.