That is all interesting, Ken and Jim.
Thanks
Chris
> On 18 Jan 2017, at 17:49, Jim Nichols <jhnichols@xxxxxxxxxxxxx> wrote:
>
> Hi Ken,
>
> Thanks for the details on the telephone business. As a hearing-challenged
> old guy with a damaged left ear, I am sensitive to phone call quality.
> Lately, I have found that call quality is much better with my Verizon iPhone
> 6 than with a fiber-connected land line. Just my 2 cents.
>
> Jim Nichols
> Tullahoma, TN USA
>
> On 1/18/2017 11:18 AM, Ken Norton wrote:
>>> VOIP is less reliable than traditional public switched telephone network
>>> (PSTN).
>> So, let me tell you a story...
>>
>> Last Friday, I was giving a "run for your life" introduction tour of a
>> telephone switching office to a new employee in our group. This
>> particular telephone switching office had a massive Nortel DMS 100/200
>> switch in it and was the host switch for many telephone exchanges
>> across the state. Well, "massive" is relative, but we still have a
>> dozen Nortel racks installed, but the main brain has been replaced. We
>> replaced about 30 racks of equipment with one, and this new switch is
>> capable of about 10x the load of the old one.
>>
>> In another location in the same state, we decommissioned an entire
>> switching office and moved everything to another location 90 miles
>> away. In that case, we went from about 120 racks to two racks and at
>> that other location, we also eliminated about 60 racks. So 180 racks
>> to 2. This is repeated over and over again across the country. Sadly,
>> it doesn't actually change our power load, but it does change our
>> physical load. The heat coming out of this rack (chilled cabinet) is
>> equal to what was generated by 300 linear feet of equipment. It's not
>> uncommon to have 12000 watts of power consumption in a single
>> home-refrigerator sized cabinet.
>>
>> These are all VoIP switches.
>>
>> But here is the deal. That is NOT the same as the Internet. The
>> connectivity between switches in the PSTN system is still over
>> dedicated trunks. Most of the trunks are still TDM (T1, DS3, OC3)
>> based, but more and more are being converted to Ethernet 1G or 10G
>> pipes. The basic architecture of the PSTN system is not changing, only
>> the equipment itself. But that's only on the switch level.
>>
>> What IS changing is on the access level. Instead of TDM muxing "at the
>> edge", the industry is converting over to IP transport to the edge. In
>> the case of a business with multiple telephone lines, we'll run a
>> single (or dual-diverse, depending on how suicidal the customer is)
>> Ethernet connection to the customer over fiber, copper or even
>> third-party service provider. This connection can be direct connected
>> to the switch, but in nearly all cases, it is transported and
>> terminated to the switch through MPLS routers. Again, no public
>> internet involved yet, except for low-grade redundancy which doesn't
>> really work. I don't want the Internet flying my airplane.
>>
>> On the consumer level, things get even screwier. FttH (Fiber to the
>> Home) seems like a great advancement, but it really isn't as the
>> current strength is only through the excess bandwidth available. It is
>> a shared environment and requires robust traffic management in the
>> switched Ethernet environment as well as the routers involved. To give
>> you an idea of how hard it is to stay on top of the growth and loads,
>> we have to replace our routers about every 4.5 years. (Moore's Law,
>> three cycles worth of aggregate bandwidth growth). It doesn't matter
>> what I put in, my absolute state-of-the-art transport network has to
>> be replaced or overlayed, in entirety, before six years, with the
>> first additions at 3 years.
>>
>> So, VoIP itself is not the evil. How it is used is the evil. And then
>> you have consumer-grade VoIP products that use the public Internet for
>> communication. That's pure garbage. More to that in a moment.
>>
>> All that said, I'm going to opinionate for a moment. Throughout the
>> history of PSTN, each new generation of telephone switch has brought
>> on major enhancements, features and capabilities. Each generation of
>> switch or switching technology has about a 20-25 year lifespan. This
>> is the first time that we've actually gone BACKWARDS in the PSTN. VoIP
>> has brought advancement in terms of options on the access circuit side
>> (as well as the ability to consolidate assets), but in terms of raw
>> features, voice quality, etc., it is either the same as or worse.
>> Voice quality is a massive step backwards. Echos, delays, buzzes, and
>> other evils are common issues, as is the ability to handle dial-up
>> modems and fax machines.
>>
>> Ok, here is my antidote on when belt+suspenders+rope+staples will end
>> up strangling you: Somewhere in the USA, there is a state-wide
>> emergency communications network that needed absolute redundancy and
>> protection. The tried-and-true (relatively speaking, as there is
>> always issues with fiber cuts, and so forth) system worked for them
>> almost glitch-free for 15 years. This was replaced with a robust VoIP
>> network that was highly advanced in the redundancy and protection
>> environment. Not only are there parallel fiber and copper connections
>> to every location, but there is also a cellular connection too. OK,
>> fine. Three physical connections to each site. Everything feeds back
>> into an MPLS routed network for termination to the head-end location
>> in the state. Incidentally the majority of the cell towers ALSO
>> physically backhaul over the same fiber network as terrestrial system
>> and even passes through the exact same equipment. Oh well... But the
>> big issue is latency. The cellular network ends up routing about 200
>> miles out of the way to an entirely different VoIP switch, and the
>> router-to-router connections have even more additional delay. The
>> fiber-network itself also has redundant paths with auto-switching on
>> failures, but can result in swings of 100-200 miles of additional
>> latency. (The biggest difference between working and protect paths
>> I've ever seen is about 2000 miles, but that's another story). So, at
>> the end of the day, what happens is that all this redundancy and
>> protection is for naught because even in the best of conditions, a
>> fault-recovery will result in the loss of ALL calls from affected
>> portions of the network. One such outage literally dropped every
>> in-progress 911 call and every other call out of the offices state
>> wide. I cannot say which state this happened or what companies
>> (multiple) were involved, but I can say that steps (very very
>> expensive steps) are being done to make sure that this doesn't happen
>> again and this will become a case-study for the industry.
>>
>> AG Schnozz
>
> --
> _________________________________________________________________
> Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
> Archives: http://lists.thomasclausen.net/mailman/private/olympus/
> Themed Olympus Photo Exhibition: http://www.tope.nl/
>
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|