How can we make Voice over Wi-Fi work?
One approach to getting VoIP distrubted round the building.
By Daniel Briere and Patrick Hurley, Network World | Network World US | Published: 01:00, 15 December 2005
Broadband service providers of all types have begun to roll their VoIP services out to a wider audience within their customer base providing services to both residential and business customers. This isnt big news VoIP, of course, has been happening for several years but the big players (RBOCs, cable MSOs and other national service providers) are moving ahead at a rapid pace these days.
As consolidations and acquisitions (like SBC's integration of AT&T) are finalised, expect this trend to continue and even quicken. And triple (quadruple? Sextuple?) play services, of course, are increasingly built around VoIP with architectural decisions (like the forthcoming Gigabit PON (GPON) RFP) assuming that VoIP will be a primary means of provisioning voice services.
Theres still a lot to figure out in the VoIP arena though dealing with emergency calls, providing lifeline services, dealing with power issues, etc. One area that is perhaps the greatest sticking point is dealing with in-home distribution of VoIP. Getting a multi-line VoIP service from the network interface device (NID), or the optical network terminal (ONT), and around the home shouldnt have to require truck rolls and inside wiring changes, but it often does.
Use a cordless phone?
Most existing VoIP services recommend that their customers use a cordless phone system to handle this distribution. Theres a lot to be said for that approach, and weve tested numerous multi-line, multi-handset POTS phones that would do the job well.
But many (and soon most) of the homes that will be served by VoIP already have a wireless system that can do the job even better Wi-Fi. Why create a secondary wireless infrastructure in the home for voice when an existing Wi-Fi network can work?
The customer enterprise and networking vendors are already making moves here. Companies like UTStarcom and ZyXEL have been shipping feature-rich and well-performing VoWi-Fi handsets for a year or more. But there are a few issues that need to be ironed out before we can all ditch our existing phone systems.
Wi-Fi is too fiddly
Todays systems require more configuration and integration than we think most consumers (and many small businesses) will want to put up with. They work, and work well, but once you start requiring users to deal with Service Set Identifiers (SSID) and encryption keys to get their phone working, youve lost them. Networking vendors have a hard enough time getting their customers up and running with laptop PCs and Wi-Fi, but mixing in a (relatively) tiny phone keypad with this process just makes things too hard.
Security is an issue too. Consumers and small businesses have compelling reasons to lock down their Wi-Fi networks. A recent case in Florida involves theft of Wi-Fi services, and could result in a felony conviction. And Westchester County New York officials have recently floated an ill-advised but not entirely ludicrous proposal to require security measures for any Wi-Fi users in the county.
Security isnt just good for the end user either, service providers dont want every Tom, Dick and Harry in a neighborhood jumping onto a neighbors unsecured access point and using their facilities to carry VoIP phone calls. And security of the calls themselves both the user authentication/provisioning and the security of the actual call data is something that carriers need to worry about.
Finally, performance issues can crop up for VoWi-Fi, just as they do with any of todays Wi-Fi applications. While 802.11e gear is going to be available soon, todays products simply dont offer any true QoS mechanisms.
As a service provider, you can do one of two things: you can wait it out, and let the access points, handsets and routers used by your small business and residential customers catch up to emerging technical standards; or you can take a more proactive approach.
Here's one answer
We recently spoke with the folks at Intoto they build the software that runs many popular Wi-Fi routers and residential gateways about how they fit into this latter approach. Theyve put together software within their residential gateway platforms that can solve all three of these VoWi-Fi problems without any huge changes in the hardware itself.
There are a couple of elements to this approach. First the residential gateway or wireless router uses two SSIDs one that is set up for traditional data networking, and a second, hidden, SSID that is used just for the Wi-Fi handsets. The phones themselves are shipped pre-configured for this secondary SSID, providing an it works out of the box solution for end users.
Security is handled with a combination of Wi-Fi airlink security (WEP or WPA ) and also by using IPSec VPN connections between the phones and the service providers Softswitch or gateway providing security for both the provider (authentication and call control data) and the user (voice packet security).
Finally, packet prioritisation is applied in the residential gateway or wireless router. This, combined with the virtually separate Wi-Fi network for the handsets, keeps the latest Microsoft operating system security patch download from intruding on that all-important phone call.
Could it solve the problem?
We think that this kind of approach can make VoWi-Fi a workable and attractive solution for a service provider making it easy for a provider to offer something truly differentiated while reducing the need to spend hours and hours inside a customers home or office reconfiguring POTS wiring. Intoto has even been working with Intel on a reference design for a small office in a box communications device which incorporates these concepts and more something we think will be truly compelling when it hits the streets (and which well discuss more when its closer to shipping).