VoIP quality -- 911 reliability

[Bill Horne]

dsr at tao.merseine.nu wrote:

>On Tue, May 24, 2005 at 10:49:14AM -0400, Josh ChaitinPollak wrote:
>  
>
>>On May 23, 2005, at 4:11 PM, Jerry Feldman wrote:
>>    
>>
>>>The FCC is being pushed pretty hard by the POTS people to regulate  
>>>VOIP.
>>>But, also remember that cell phones must all contain GPS technology  
>>>at the
>>>insistence of the FCC. Putting GPS technology in the cell phone  
>>>makes sense
>>>since on a cell, when dialing a number, it may be physically dialed  
>>>from
>>>miles away.
>>>      
>>>
>>Why not do the same with VOIP and require them to have GSM as well?
>>    
>>
>
>First, GSM is a speech codec. GPS is a satellite-based locating
>system.
>
>You can't require VOIP to have GPS because VOIP doesn't require
>any specialized hardware beyond a microphone and speaker and
>appropriate ADC/DACs. You can have a VOIP phone implemented in
>software on a laptop, or a PDA, or your mother's Windows box.
>
>Second, the price of a hardware VOIP interface is down to $40 and
>falling, whereas the cheapest GPS unit I'm aware of runs $90 or
>so. 
>
>Third, the term "VOIP" covers two or three dozen implementations
>of a dozen or more protocols. 
>
>-dsr-
>  
>
I'm going to join this debate just out of contrariness ;-J.

A discussion of VoIP and E911 must take place inside a political, not a 
technical, context. Technical solutions won't address the social and 
economic forces in competition here.

There are three or more competing organizations in this debate, and I 
hope I can shed some light on the agendas each brings to the table:

1. The public sector is seeking to minimize both the apparent and actual 
costs of
    providing emergency services -

   A. E911 is a highly automated system, which demands *UNIQUE* addresses
        be associated with every phone number, thus minimizing training 
costs,
        allowing use of less experienced personnel who don't have to be 
familiar
        with the area in which they're working, and making it possible to
        distribute emergency equipment and personnel much more effectively.
   B. The cost is distributed more to businesses than to voters, since 
businesses
        pay the majority of telephone charges, thus making voters less 
aware of the
        tax.
   C. The decreased (i.e., improved) response times made possible by the
        automation allow for wider coverage areas and lesser staffing. 
This, in
        turn, causes insurance underwriters to raise rates for 
catastrophic coverage,
        since emergency services aren't manned at the old levels and 
can't help
        each other as much when a disaster strikes. Again, the public is
        less aware of the added tax.

2. The ILECs (Incumbent Local Exchange Carriers) are heavily invested
    in the E911 infrastructure, and don't want it to change -

   A. The need to tie an individual phone number to a unique address fits
        perfectly with the ILECs existing distribution model, i.e.,
        one-pair-of-wires-per-phone. Chicken or egg debates aside, the
        ILECs methods and procedures have always been geared to serving
        fixed addresses, so E911 was much easier to implement in a
        wire line world.
   B. ILECs were burned badly by the "unbundling" of their services and
        the requirement to offer cut-rate E911 access to smaller
        competitors. Saddled with a high-cost unionized workforce, they
        nonetheless were forced to share the E911 network with the
        (mostly) non-union Competitive Local Exchange Carriers (CLECs),
        thus forcing the ILECs to bear what they feel is a disproportionate
        share of the E911 infrastructure cost. ILECs are loathe to allow 
VoIP
        competitors the same advantage, and are lobbying to require that 
VoIP
        providers make heavy investments in the wire line infrastructure
        upon which E911 is built, even as the ILECs, paradoxically, provide
        the cheap bandwidth VoIP is hoping will be its salvation.

3. The cellular providers, although separated from their ILEC mothers by
    a lot less than six degrees, are starting to realize that they are 
becoming
    serious competitors to the ILECs and will have to split from Ma Bell's
    old family in the future.

   A. There has been, and continues to be, a dramatic shift away from
        wire line service for young consumers raised on cellular and used
        to its convenience and transportability. Although exact figures are
       proprietary, there has been a very substantial drop in the "fall 
rush"
       orders that used to inundate ILECs at the start of the college term
       here in Boston. The Negroponte Switch is actually happening, at
       least in this case.
   B. Coupled with the shift to wireless, the Cellcos are trying to avoid
        - or at least delay for yet-another-fiscal-year - any added invest-
       ments in E911. However, since they're still tied to Ma Bell's old
       methods - but not its unionized workforce - they are schizophrenic
       about the best way to avoid paying for it. GPS is nice in theory,
        but terrible in fact, with all the other players pushing back - 
hard -
        to try to avoid shouldering any more of the cost. (1)

4. Into this Evening-In-Byzantium mix of politics and power come the
    VoIP providers, enjoying the fundamental advantage of having
    externalized their transport costs onto their customers, but living on
    a razor's edge between experimenters and early-adopters, all the
    while praying for a technological miracle that Shannon proved can't
    happen and keeping their fingers crossed as they shill "telephone"
    service to a country which equates "telephone" with an ultra-reliable,
    always-on network that has the best fidelity and maintenance in
    the world. (2)

   A. The public service agencies have the political clout to force
        anybody-else to bear the cost of shoehorning both Cellular
        and VoIP into the existing E911 infrastructure. Politicians
        of all stripes would rather voters pay $10 to an insurance
        company or a VoIP provider than $1 in direct taxes.
   B. The VoIP providers can't afford to rent wires into the E911 tandems,
       since their entire business model is dependent on someone else 
paying
       for transport.

VoIP can't deliver Common-Carrier grade service, because there is no way 
to force a network that VoIP providers don't own and don't control to 
allocate enough bandwidth to deliver virtual-circuit quality. Despite 
their ability to provide some novel features, they aren't able to break 
into the mainstream market unless they solve the quality-of-service 
problem, and even then the VoIP business model only works with those 
whom already have investments in IP transport and are unwilling to 
consider those costs as being part of their phone bill. (3)

IMHO, the E911 issue isn't a serious problem with VoIP, since it isn't a 
serious competitor to the existing carriers. It's only a niche player, 
and will remain so: inevitably, VoIP providers will shell out the money 
needed to have the CLECs do their E911 for them, because that's the 
least cost route.

FWIW. YMMV.

Bill Horne

1.) The E911 services have their own, separate databases to correlate 
phone numbers and addresses. The only thing they get from the telco is 
the number where the call originated. There is no room in the spec for 
position info, so a major modification is needed.
2.) America and Canada are the only countries in the world where 
customers routinely pick up a telephone and dial without bothering to 
listen for a dial tone. Everybody else has to check: we don't.
3.) The intersection of persons needing emergency services who also 
poach WiFi bandwidth is small enough to ignore.

-- 
E. William Horne
William Warren Consulting
http://william_warren.home.comcast.net
781 784-7287