When telephone services first came in to being it was one of the first times that we had a communications medium that would allow people to summon help in a timely fashion.
It made a lot of sense for the state monopoly telcos (such as GPO in the UK) to provide a means to summon help by calling 999, free, from a payphone or even (if you had one) your home phone.
But times have moved one, and the whole concept of a "phone line" and "traditional voice telephony" is getting somewhat blurred.
OFCOM have insisted that VoIP providers provide access to 999, which, in itself, may have some merit. The issue I have is that VoIP is really no different to any other protocol using IP (Internet Protocol). We don't expect everyone that runs a web site to have a "999" button on the site in case you have an emergency, for example. VoIP is not really much different.
There are a lot of things in progress to try and get VoIP and emergency services working well together. It largely revolves around "location information" so that when you have an emergency it is possible to locate you quickly and easily.
With landlines the records for a phone number had details of where that phone was located, and if you were on the phone you were there.
With VoIP there is no location data in the network. If there was, then how would it get there and how accurate would it be? Unlike the old days of an analogue phone line, not only is it possible that my UK VoIP phone number is in use from a device that is not even in the UK, but increasingly likely. Tunnels for IP level data, and VoIP platforms being simple software that anyone can, and do, run, means that you really have no way to know by "the network" where a VoIP user is located.
So, my initial views on this are that the end equipment being used for VoIP should be trusted to report where it is located, in VoIP headers. It would not be rocket science for location data to be default in PPP and DHCP and where handsets have location information (smart phones and the like). Technology could all work together to provide that.
However, the more I think about this, the more I think the link between "summoning help" and "telephony" needs to be broken once and for all.
Telephony is increasingly becoming just one of many types of communication that we use. I use irc, iMessage, SMS, email, FaceTime and such things more than I use phone calls, and whilst I may be a techie, this is a changing trend.
How long before the concept of a "VoIP provider" is moot - where people call their friends using a link on FaceBook, and businesses using a link on the businesses web site using webRTC. There won't be a "VoIP provider" to provide access to 999. Even now, you can call me using sip:firstname.lastname@example.org without any intervening "voice telco".
What we need is standards that allow an "emergency app" on smart phones. That can use the GPS and location services in the phone, contact the local emergency services operators by voice, video, whatever, and provide location. It does not make use of any "voice provider" in any sense, it uses IP.
Indeed, an emergency app can not only provide location, but video, stills, and other monitoring of the situation - much more data to help emergency services and allow assistance to be offered. Way more than just the old style voice call.
Maybe the few remaining landlines need to carry on supporting 999 and perhaps even conventional mobiles with cell location, but why pick on one IP application "VoIP" to jump through unnecessary hoops here when they will not really work anyway and just add cost and hassle.
The world needs to change, and not in the way the various standards bodies are heading. By the time they have the standards, and the legal framework to enforce those standards, for VoIP location data, the idea of a "VoIP provider" will be long dead. Lets make new standards that move forward.