Alarming plans for the NBN

Alarming plans for the NBN

Summary: Will medical alarm systems — or other devices attached to your phone line — still work when you switch to the NBN? Not necessarily.

TOPICS: NBN, Telstra, Australia

The Personal Emergency Response Services Association (PERSA) is worried about the impact that the National Broadband Network (NBN) rollout will have on people who are reliant on medical alarm systems. These devices currently work across the standard telephone service that Telstra has been obliged to provide. But all of that will change when we move to fibre.

In this week's Twisted Wire, Leica Ison, general manager of Product Management at NBN Co, says that there has been a lot of work testing devices across the NBN's analog UNI-V port. But as PERSA pointed out in its submission (PDF) to the Joint Committee on the NBN, retail service providers (RSPs) are not obliged to offer that port. In fact, they say, the indication is that many will choose to offer services entirely through a data port.

Does this mean that the responsibility for offering a working service for these devices rests with Telstra? After all, they have an ongoing commitment to provide a standard telephone service across the country. "Not necessarily," said Paul Brook, founder of consultancy firm Layer 10 Advisory. He points out that their obligation does not extend to the quality of the service or the requirement to support devices that hang off the existing phone service.

Is this issue big enough to require some sort of regulatory change? It's a question I put to Peter Harris, secretary of the Department of Broadband, Communications and the Digital Economy and acting chair of the Telecommunications Universal Service Management Agency (TUSMA). You be the judge of whether this issue is in hand, a storm in a teacup, or waiting to explode.

As always, we welcome your comments on the Twisted Wire feedback line (if you'd like to be featured in the program) on 02 9304 5198. Or, of course, leave your remarks below.

Running time: 30 minutes, 04 seconds

Topics: NBN, Telstra, Australia


Phil Dobbie has a wealth of radio and business experience. He started his career in commercial radio in the UK and, since coming to Australia in 1991, has held senior marketing and management roles with Telstra, OzEmail, the British Tourist Authority and other telecommunications, media, travel and advertising businesses.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • No Issue

    Whether voice services are offered via either of the UNI-V ports - (which are effectively ATA's) - or specifically through the UNI-D port - (with an external ATA) - there should be no issue.

    As long as an FXS port is available, these services should work.
    • No Issue

      "should" is the operative word - listen to Paul Brooks remarks about the digitisation process in relation to alarm systems
      • We know some are afected

        Already confirmed in the testing labs (raison d'etre).

        Billions of dollars worth of private PSTN equipment impacted. All to be tested, modified or replaced as the copper network is decommissioned.

        Another subsidy in the offering?;-)
        Richard Flude
        • Oh...

          The poor 1900's PSTN...!

          What about those billions of dollars worth of typewriters too?
          • Old vs new - get it right

            IP has its origins back in the 1960's. TDM didn't start its rollout until the late 1970's/early 1980's.

            If you are comparing data networks vs voice, then telegraph predates telephony.

            If you want to criticise a technology because it is old (a pointless exercise in my opinion), you are backing the wrong horse.
          • Dear oh dear

            Missed the whole facetious analogy whilst being snowed under in pedantics eh?

            BRC, there's another blast from the past *sigh*
          • yep, that's right RS, the copper should really be heritage listed. So besides the political motives for opposing the NBN we also have to consider peoples emotional attachment to the copper too. lol...
            Hubert Cumberdale
          • Perhaps both

            of you are missing Richard's point. Those services designed to operate on the POT network may need to be modified to work on the broadband network. Sooner the better, although as RS recently pointed out they have 10 years lead time.
            Blank Look
      • The issues are technical, not legal

        Most of this discussion is based on the flawed assumption that regulation can fix the issues. The issues are technical and unless someone knows how to rewrite the laws of physics, rewriting laws won't help.

        The core issue is the assumption that the UNI-V will offer a reliable service. There is a huge caveat on that - both ends have to be UNI-V. If not, there will be sync errors leading to packet loss that can potentially become so great, link timeouts and dropped connections can result.

        A TDM network uses a centralised clocking system that ensures that all devices run at precisely the same rate. One bit in, one bit out. Although the clock is synchronised to an atomic clock, even if it was a bit out, all devices would be out by exactly the same amount. Because the digitisation is performed centrally, this process is timed based on this master clock.

        The first thing an ATA must do, whether it is the one in the UNI-V port or a standalone device, is sample the analogue signal 8,000 times per second or once every 125us. How does it know how long 125us is? A TDM network determines that from the atomic clock. ONT devices such as the NTD can be synchronised from the optical fibre network which means that the UNI-V port is driven from that same clock. Unfortunately, that is the exception, not the rule. Stand-alone ATAs or those built into routers must operate off their own crystal oscillator. In fact, this crystal oscillator will determine all timing of the whole unit. So when the ATA puts together 10ms or 20ms of speech into a G.711 packet, that 10ms or 20ms is timed by that very same oscillator. By the same token, the device at the receiving end is also running on a crystal oscillator. While crystal oscillators are pretty good, no two can ever be absolutely identical.

        So what if they are out by just a little bit? It can't be that critical, can it? Inside every ATA is a jitter buffer in the receiver. These are needed to take into account the variability in packet delivery times across a typical IP network. The jitter buffer builds up a bit of a reserve of data packets so it can continue feeding out an analogue signal even if a packet arrives a bit late. After all, a break in the analogue signal would sound like a click that if it became too frequent, would come across as noise making the call unbearable. If the send clock is a bit faster than the receiver clock, the receiver's jitter buffer will eventually fill. Once this happens, new packets have nowhere to go so they get dropped. If the send clock is a bit slower that the receiver clock, the receiver's jitter buffer will eventually empty leading to a loss of signal and distortion (I'll explain that later). One way or another, the received signal will be interrupted, leading to data errors.

        Some devices have no error checking. These data errors will pass through and the receiver will do whatever it does. TTY devices will just echo characters incorrectly. If there is error detection, the receiver will request a retransmission. If the retransmission works, great. What if it too is corrupted? Eventually, most protocols will give up. Retransmissions are intended to address infrequent and inconsistent errors, not deal will regular packet loss.

        The longer the transmission, the more likely the jitter buffer will over- or under-flow. Short transactions such as those associated with alarms and EFTPOS transactions will probably be fine. Long transmissions such as faxes and EFTPOS firmware upgrades will more than likely encounter jitter buffer issues.

        Given that the NTD and hence the UNI-V is centrally clocked, if both ends are on UNI-V, it will enjoy the same benefit TDM has, hence the connection will most likely work. If one end is UNI-V and the other is in the PSTN, then there is the equivalent of an ATA in the PSTN gateway. The PSTN side is synchronised with the TDM network, but the IP side is essentially free running. There is no point syncing only one side of the connection, so having the UNI-V synced to the central clock at other devices not does not help. If the IP and TDP sides are running at different speeds, the jointer buffer problem will occur inside it. There is no way around it - if you have two clock sources, you can never assume that they will run perfectly in sync, so the jitter buffer problem is inevitable.

        Finally, to the distortion. VoIP is designed to be tolerant of packet loss. Packet loss is a reality of IP networks and for VoIP to work, it has to deal with it. It achieves this by using special algorithms to attempt to cover up for missing packets. This could be as simple as replaying the last received packet. This is good enough to trick the human ear to the point of the loss being unnoticeable. The problem is that a replayed packet will be very noticeable to a data system. Even at a snail's pace 300b/s, one 20ms packet equates to 6 bits. This is enough to corrupt two bytes. Two bytes is enough to corrupt the transaction.

        In summary, PSTN to PSTN works. UNI-V to UNI-V should work. Anything else…good luck.

        One day, all devices will be natively IP. Unfortunately, there is a significant amount of money invested in existing technologies that cannot be changed over quickly or cheaply.
  • Apparently DevotedNBN is one of the few offering UNI-V connections

    I am not sure if there are any, or many, others.
  • Redunancy

    How many Companies are still using Business equipment or even Office facilities that collectively Billions were spent on within the last ten years alone. How many 486 let alone P3 computers are in corporate use let alone tractor feed ribbon printers or CRT Monitors. All still functional, but now redundant, do we scream Billions wasted ?
    An opportunity to upgrade and improve.
    In any enterprise taken on there will be challenges, that is normal. What matters is how these challenges are addressed
    Abel Adamski
  • I don't see the problem

    As I understand it, voice is digitised at 8bit resolution at 8kHz at the exchange. If this UNI-V port is using G.711a (which is basically the same), why wouldn't all these devices work the same?

    I have a monitored alarm and therefore the availability of the UNI-V port will be a deciding factor in the choice of ISP's for me. If iiNet (my current ISP) don't offer it, I'll move to an ISP that does. I'm sure there will be some.
    • Security Systems

      I can see an evolutionary jump in security systems. I have a video security with H264 compression and HD resolution, the compression does degrade the resolution. The data feed onto the home LAN is 2Mb. It incorporates motion detection, recording and interfaces with PIR's and can interface with a standard keypad activated security system.

      With the upload available not only can the alarmed sector video be live streamed, but also lower res feeds of all the other cams, Pan Zoom Tilt cams can be remote controlled. Not only applicable for security but also for the elderly and infirm for remote assistance , alarm and monitoring with HD Video and Audio, plus transponders monitoring vital signs such as used by health and excercise fanatics can be incorporated . Possibilities are boundless
      Abel Adamski
      • hooray,

        I agree and am looking forward to being able to transition the house alarm and PTV across to broadband. Then gone is the fixed telephone line.
        Blank Look
  • Missing the point

    I think some people, much like the Govt and the NBN SME's who thought up this great UNI-V idea have not considered the customers and services that are built on the existing PSTN.

    Couple things to consider:

    - Our aged population as outlined in this discussion does not comprehend the impacts of migration from PSTN to NBN nor probably can afford any technology changes needed if there was NBN solution. Who will be there to support the aged population when their services just stop working, after the NBN Contractor leaves the premises?

    NOTE: Come May 2014 there will be 15 regions that are turning off the copper network.,nbn-co-prepares-first-copper-disconnects.aspx

    - Dial up EFTPOS and HICAPS systems are yet to get the green light from the NBN test bed that they will work. What does this mean for small business? Well there is the potential your local petrol station or Chiropractor will not be able to process transactions due to UNI-V technical limitations.

    Scenario: Petrol station can’t sell fuel, truck driver can't by fuel, truck driver can't deliver Woolworths goods, Mum and Dad can’t buy food for dinner. The impact is extensive.

    So the situation arises what then?
    Can the customer cut back to copper?
    What if it’s at the 18 month forced disconnection date?

    The responsibility in my view does not lay with any RSP (including Telstra) but rather TUSMA, NBN and the Govt.
    • Hmmm

      With respect you will find these problems have been encountered elsewhere in the world, solutions have been worked out, so by the time the copper is turned off the solutions will be available
      Abel Adamski
  • Ah...

    Y2K yes...