Brief summary for people with too little time: slap an ADSL modem and wifi hub into every landline phone and allow them to network automatically with each other and you suddenly have a simple way to bathe an entire home in net-enabled connectivity without needing a computer. A more detailed investigation of this concept (with pictures) follows:
So I’ve been thinking a lot about ubiquitous home networks recently, and the ways in which various appliances might start hooking up to the internet and through the internet to other people – social hardware if you will – and the problem keeps coming back to how you introduce the network into the home in the first place. There needs to be a way of wrapping all the core parts of a home in a network without it being something that requires complex set-up and specialised hardware. It also seems to me that the key to true ubiquity is to detach the networking completely from a its current reliance on a computer. Your home network of the future should not require a perpetually-on computer in a cupboard. Your gran should be able to have the benefits of internet enabled appliances without having to figure out the configuration of modems and puzzle their way through a complex OS-based interface.
And if – as I assume – we’re talking about wrapping the home in a wireless network, then it also seems to me that we should be looking for a way to do all this without introducing lots more widgets and boxes and cables around the place. Ideally – we would also try and avoid having little appliances stuck into random power supplies around the house (unless of course we can take them in a different direction and use them as control nodes as well as bridges cf. Airport Express – but more on that kind of paradigm another time). Essentially, we need a model in which home, net-enabled networks are treated more like a utility than a technology – more like water or electricity provision than …
Okay – so now we’ve got the criteria in place, how should we go about making this wifi-enabled network space? Probably the place to start is at the bridge between the appliance (including potentially a computer) and the network. Since these appliance could be in pretty much every room, then the first thing we’re going to need is a series of wifi points littered around the premises. These ideally would cover the entire home, but if they couldn’t cover it completely they’d have to be in key areas like kitchens, studies, sitting rooms, bedrooms and the like. They would not be as useful initially in storage areas, hallways, lavatories, bathrooms or on stairs – although clearly it would be an advantage if the bled into those areas. These points need to be powered in some way and they’d presumably need to connect with one another as wifi bridges. One of these appliances has to be able to connect to the internet. More than likely they’ll do this via the telecommunications grid through a phone socket. And then there will have to be some kind of interface for setting up the connection and protecting it with some kind of password, encrypted and connectable to by some kind of industry standard protocol. This interface would not need to do anything else, but conceivably could do…
So here’s my contention. Given that it would seem to be a good thing to split the provision of wireless network access from computers, and given that we’ll still need an interface and given that we need a point in all the core rooms of a home and given that we need to connect this network to the telephone network in some way – isn’t the telephone itself the ideal appliance to be the heart of the home network? Unlike the television or the radio or the stereo, any place in a home where people are likely to spend a lot of time is likely to have a telephone point in or near it. They have small interfaces on them already – a numeric keypad for one and often a small LCD screen for recording input, and they’re already connected physically to the telephone network.
So here’s what I’m thinking – and forgive the slightly ugly 80s styling of the phone itself. I tried to do something beautiful and isometric but it came out looking really nasty. So we make do with gradient fills and basic Illustrator shapes…
So the ADSL modem and wifi antenna/bridge/hub are both included within the device. This means that in terms of buying a wifi network for your house, all you have to do is purchase the phone and plug it into a phone socket. By sticking an Ethernet port into the base of the phone you could immediately use it to connect to printers or any non-wifi enabled networkable device. If you bought a second phone, however, it would operate like a wifi bridge (there’s already considerable precedent for hubs also acting as bridges – with the Airport Extreme being the most recent example), extending the network around the home. If ADSL modems did not reduce significantly in cost, then perhaps you could remove that from the additional phone units, creating master and slave phones, each of which could be strung together to extend the network still further. If ADSL modems came down in price, however, it might be useful to build them into all the devices – allowing each phone unit to negotiate with the other phones as to when it should become the dominant provider of access to the internet (ie. if the connection broke down or if it became clear that one phone could provide more throughput because of the local quality of the line or intra-phone connectivity). Either way, you’d expect the network to self-organise purely by bringing a new phone home and plugging it into a socket. The blue-lines in the following image would be self-organising connections between phones based upon proximity and strength of signal:
So now we have a wifi network in the home, where all you’d need to do to extend the network is purchase a phone and plug it in. And we have a number of devices capable of connecting to the web. Except we’ve left out questions of user names / passwords / encryptions and the like. Since we’re talking of this service as a utility, then the most obvious way of handling it would seem to me to be to get your ADSL along with your telephony from the same operator. Since the operators already know the telephone number that the phone is plugged into (and will know this whenever you use a phone on that network) it seems most obvious to consider that telephone number to be your user name for connectivity and the name of the local network. This would mean that when the phone was initially connected it could attempt to connect immediately to the operator. At this stage the operator (or the phone) could generate a numeric key with which to access the network. All you’d have to do is plug the phone in and then ring up your operator. Since they already have security provisions in place to help identify a caller, they could easily determine that a user was legitimate and give out an initial code which said user could then use to login to the network.
In practice this would mean the entire process to set up the network was to plug in the phone, ring an activation number and get your code, hang up and type in the number. Any other phones you wanted to connect would just require you to plug them into the mains and type in the activation number. And then to login from any device all you’d have to do is connect to the network which was called your home phone number (Network Name: 020 7286 ####) using (again) the activation number. Piece of cake!
The process would have other possibilities too. By using a numeric key rather than an alphanumeric key you immediately open up the number of devices that can be easily set up to use the network. Numeric keypads are far more common than full text input devices and faster to use. It would take no time at all to connect your mobile phone, television, DVD player, Tivo, Radio, CD player, tape deck and computer to such a network. But that’s just the beginning. Radio Alarm clocks have keypads, Microwave ovens have key pads. In fact the only electrical things that I can see around me in my flat that don’t immediately present some kind of numeric interface are my lights, iPod, digital camera, kettle, X-box, toaster and oven – and four of those have an interface that would allow you to choose numerals in different ways.
So that’s the concept in a nutshell. I can see some problems with it with regard to the separation of telecommunications services and the necessary connections that you might need to make between hardware and service providers that might make the whole thing unfeasible. I’m also more than aware that there have been explorations about ways of connecting telephones and connectivity elsewhere – some of which no doubt overlaps, encompasses or surpasses my thoughts – and no doubt I’ve made a few errors through the piece as well, but nonetheless I thought it was an interesting enough idea to push out into the real world and to receive feedback around. And that’s what I’m after now – please feel free to leave any thoughts, fixes, suggestions or extensions below or write a post and trackback to this one, so any interested parties can follow the discussion (if there is any) more easily…
17 replies on “A proposal for Wifi-hubs to be built into landlines…”
It’s a nice idea, but…
The way telecommunications is heading is away from the landline as it stands – I suspect a device as you envisage is probably on the cards, but I suspect the connections might be in a slightly different order.
There’s three contributing factors to my opinion here: firstly, you can now get ADSL in the US without an associated telephone service; secondly, BT is slowly (but surely) moving towards IP across the PSTN (Cable & Wireless have had an IP voice network for quite some time now, other operators have similar); finally, an awful lot of people are abandoning their land-lines in favour of mobile telephones.
So where does this leave us?
In my opinion, this means that rather than an RJ-11 telephone socket in every home, it’ll be an RJ-45 ethernet port. The great British Telecommunications company will present you, one way or another, with IP connectivity – whether it’s a small box in your home, some gizmo in the green box in the street, or IP direct from the exchange.
As for phone service, this will increasingly be carried over IP. This may or may not be via a telephone plugged into an ethernet socket, however. Devices which can roam transparently between IP, PSTN (via DECT), GSM and UTMS already exist in the labs – though we largely don’t have the infrastructure to support their use in the real world, yet. It’s only a matter of time, though.
After all, almost everybody who has a landline has a cordless telephone nowadays, and BT are trialling DECT/GSM roaming (although typically they haven’t figured out how charging will work).
As for network addressing: associating additional information with telephone numbers is precisely what ENUM is for. Unfortunately, ENUM in the UK isn’t really going anywhere – and it’s going there very slowly. If you’re not familiar with ENUM: it’s a mechanism by which DNS records can be associated with telephone numbers. The current public ENUM tree is stored under the ‘e164.arpa’ domain. As an example, the telephone number +290-510-2000002 has the domain name 2.0.0.0.0.0.2.0.1.5.0.9.2.e164.arpa. A couple of seconds with ‘dig’ and all should become clear. There’s absolutely no reason why an ENUM domain couldn’t have an A record associated with it — for example, the IP address of the network gateway associated with that telephone number.
Either way, there are going to be a lot of changes in this arena over the next couple of years.
Ah excellent! Technical feedback! So first things first – the VoIP stuff I knew about – or at least had a sense of – particularly as BT recently announced they would be handling their exchanges with TCP/IP rather than switches. This whole area was one of the things that suggested phone-based hubs to me. I’m not convinced that everyone’s going to have an ethernet connection instead of a phone socket in short order though – that seems to me to be much less likely. The cordless / roaming stuff doesn’t seem to me to be strictly relevant (or at least I don’t think it breaks anything) since you could easily replace the phone terminal itself with the hub of a cordless or with a basestation for roaming mobiles that switch to local landline-alikes when they’re within reach of their home terminal.
I think you’re quite right about the DSL connections without phones though – that’s where I’m slightly stumped. And the stuff on ENUM and the like is really interesting indeed. Slightly terrifying telephone domain, but I guess that doesn’t need to be seen by humans.
You could well be right about the Ethernet thing – maybe I’m thinking slightly too long-term (VoIP is my day job, so I get pretty emersed in this stuff!)
Yes – ENUM domains are horrible to type in. Fortunately, they’re something humans never need to see. On our VoIP platform, we use real telephone numbers (though they can’t be dialled from the PSTN). When somebody dials the number of someone who isn’t one of our users, we use ENUM to try and get a SIP or H.323 address for them, so we can route the call via IP. If that doesn’t work, we can punt the call off to the PSTN (preferably in the destination country, rather than the source). ENUM has the potential to do a whole lot more, but committees and working groups and the ITU-T and DTI and Ofcom and everybody else are conspiring to make it as painful as possible 🙂
two problems i see:
1) i have a cell phone and cable modem – no land line. an american land line costs about $30 a month even if you never use it. that plus the actual internet access fees is too high a base rate for enough people to use this to make the products reasonably priced. that’s my guess, anyway.
2) the security sounds incredibly weak. if the access number is the phone number (and maybe i’m not really understanding this part?), all i need is a wifi-enabled laptop and your phone number to access your network?
With regards to the security – no, that isn’t the model I’ve proposed. What I’ve described would mean that your home would have a network which had as its name the number of your telephone, that you would need a password to log into that network and that the initial password you would use to log into that service would be something that was negotiated between your ISP/service-provider and your phone and which you would ring up to activate. My thought would be that a user would be able to change that password via a web-form interface on the website of their ISP. It does remove one level of password complexity / security in that it conflates first network name and ISP login name and secondly ISP password/network access password – but realistically I only did this because it seemed to me to reduce set-up complexity. There doesn’t seem to me to be any reason why they couldn’t be different if you were supplied with a suitably rich interface through which to change them.
With regards to the landline – there’s a few things there – landline companies are trying to sort out the relationship between the mobile phone and the landline, including possibilities that mean that while you’re at home your phone works like a landline and while you’re out it operates as a mobile. And appliances like this clearly require some kind of local hub plugged into phone sockets, so that would still work.
Also, clearly we’re not talking about a model here that everyone would have to go with. I’m fairly sure that there is still an enormous proportion of people with landlines in the States and that this market is unlikely to collapse completely over the next ten years.
With regards to cost – I’m basically proposing that the phone companies move closer towards a TCP/IP based business – with wifi internet access and telephony increasingly starting to be merged together as a service (as some thing like “home connectivity”). The assumption would be that telephone calls would start to move towards Voice over IP-based models and be included within one rate. I think it’s difficult for us today to state conclusively the economic models around future technology. Clearly bandwidth is on the increase, clearly prices will drop further and clearly what starts as expensive technology becomes almost trivial in reasonably short order. Whether these prices drop fast enough to make the phone I propose practical before the technology is obselete is, of course, far from a certainty…
I think the sentiment is correct, but I think the devil is in the details. Random thoughts:
The biggest problem is that companies that make phones don’t tend to be those that understand IP and wi-fi. They’re used to people buying a phone once or twice in 5-10 years, and the cost of the phone has to be as low as possible. Sure, there are lots of companies making business phones with VOIP, but they are stupidly expensive.
Next problem is that most houses aren’t cabled well for phones – they generally have one or two sockets at most. Also, the phone socket tends to be in some stupid place – normally near the front door, and generally somewhere without power (rather than an old-fashioned telephone table to go in my hallway, I’m going to get a little ADSL table). Makes more sense to use the wifi for other phone connections, not the copper phone system. Or do IP over the power network (useless on a large scale, can work ok in the home environment).
Setting up ADSL isn’t easy – there are lots of variables, mainly at the network level, but then there’s also lots of different ways to authenticate. Many networks rely on MAC address authentication. Setting up this sort of thing is going to take more than typing in a number. One solution (that contracts other points) is to add a POTS modem in there too, so after ringing up to subscribe, they can dial in and set it up for you.
Using phone number for a wi-fi network name is pretty bad, security wise. Unless you want tech support calls at 3am in the morning from someone wardriving outside your window. Wardialling is back.
Wi-fi bridging is pretty untested – what happens when you have large meshes of networks? What happens when all your neighbours set up similar networks? What happens when you have 100 bits of consumer electronics trying to get on the network? We’ve all seen the problems of wi-fi at conferences with 100+ wireless users.
Must admit, I doubt BT will ever be able to give you an Ethernet jack. Their POTS system is just too old and tired – until ADSL, they were busy trying to double up phone lines onto one piece of copper (with noticable drop in voice sound quality). Maybe it’ll happen in other countries, especially those with good cable TV networks, or some of the LLU companies in cities.
But anyway, great post, ridiculously easy home networking is a really interesting topic.
Hi Tom,
I know a lot of carriers are interested in the 802.16d standard (WIMAX)which aims to be a last mile broadband solution.
Have a look at http://www.wimaxforum.org/
I believe BT and Eircom (here in Ireland) are keeping a close eye on this.
“Wi-Fi and WiMAX are complementary technologies. As WiMAX is a “last mile” technology, meaning that it connects businesses and homes to the high-speed Internet. Wi-Fi provides the wireless LAN connectivity within a building or a home. In the notebook computer of the future, you may have both WiMAX and Wi-Fi technology to make connections to the broadband Internet. These two technologies have been architected as close cousins, and will work together to provide the best connection for your needs.” – http://www.intel.com/netcomms/technologies/wimax/index.htm
Digressing a little -> What would be cool is if we replace mobile phones with little devices like the Vocera pagers (http://www.vocera.com/products/products.shtm) and you as individual is reachable via a DNS entry / URL instead of a phone number 🙂
Complete IP solution (threat to GSM from landline data carriers ?)
Derek
and you as individual is reachable via a DNS entry / URL instead of a phone number 🙂
The only problem with that one of practicality. Ever tried typing a URL on a telephone keypad? It sucks 🙂
(Which is part of why ENUM was created, of course)
Chris’s comment about the phone socket/power socket problem is an important one. Whilst many people now use Dect/cordless type phones, which require a mains socket to charge from, and other devices such as answerphones also require power, the standard corded phone in this country doesn’t require any form of power socket to run. It gets all its power from the phone socket.
To run the kind of device you’re proposing means that every comms socket (be it RJ11 or RJ45) must have a power socket next to it. Even if BT go onto ethernet, what’s the chance of them providing power-over-ethernet (or even if it’s enough power).
Of course, using power cables for connectivity gets around this problem, but it’s certainly a unique little problem brought on by houses designed a long while ago around phones which didn’t need a mains plug.
Yeah – i agree. The problem with small form factors is that our hands don’t scale that well 🙂
What about storing a voice recog label on a server somewhere that maps to the ENUM address. That way you don’t have to type. Still have the problem of calling someone new…
Wireless broadband is the way to go and it would be great if technologies like 802.16d become commercially available (soon!). There are lots of stuff we can pipe down IP, voice being just one.
Somewhat related to the house wireless network is Zigbee (http://www.zigbee.com/) which is a proposed standard for home control and building automation using cost-effective, low-power, wirelessly networked monitoring and control products.
The industry big players seem keen to get some WIFI broadband solutions out there. The major obstacle seems to be getting agreement from local regulators managing the RF spectrum.
I thought you said landMines as I scrolled past. Which is an interesting but horrifying concept. I’ve read a few thoughts about smart mines and they all require wifi like communication. Brings a whole new meaning to war driving.
A bit late in the game on this one, I’m catching up on my newsreading… Anyway, this kind of device is already available in France, I’ve got one in my living room. See http://www.freebox.fr for details (in French only).
The general idea is that you subscribe to an ADSL service, with the option of disconnecting from your usual telecom supplier (the equivalent of BT) so you only pay your ADSL service. You are lent a small box, called a Freebox in my case, which plugs in to your land line and provides a SCART socket for digital TV, a phone line (with your old land line number if you disconnect from your old supplier), an Ethernet jack and a USB socket to connect a PC to the Internet, and optionally Wifi (802.11b only for now). The device can also act as a NAT router… Hey presto! One box anyone can plug in (literally – there are no passwords to provide, the box you are sent is keyed to your subscription), buy the Wifi card and activate the Wifi and router functionalities (easy enough as long as you don’t want to forward ports, which most users won’t), and all your Wifi-equipped friends can surf without there being a “master” PC to share the connection. There are other advantages too, the main one being free phone calls within the whole of France…
Link Harvest
The 10 greatest Scots of all time: according to the Adam Smith Institute; A feature, not a bug: Telepocalypse point out smsBug, with linklove for my recent mobile search article; Ex-Symbian Chief Predicts the Future of Mobile Software: Services…
Link Harvest
The 10 greatest Scots of all time: according to the Adam Smith Institute; A feature, not a bug: Telepocalypse point out smsBug, with linklove for my recent mobile search article; Ex-Symbian Chief Predicts the Future of Mobile Software: Services…
Hello, I wish to set up a wifi service for a small city, how do I do that ?
well done i really was looking for this
I wish to set up a wifi service for a small city, how do I do that ?