The Ruckus Room

The Death of Distance: Voice at Home

You don't hear a lot about the "voice" part of the proverbial "triple play."  But the voice part is perhaps the biggest part.

Much more hype has obviously been created around IPTV, but operators also have big ideas and aspirations surrounding voice over broadband (throughout the home). But the motivation to deliver voice over broadband differs from carrier to carrier.

Cellular carriers don't want to lose minutes and are keenly interested in metro Wi-Fi to eliminate coverage holes, maximize their footprint and reduce dependencies on roaming. Traditional terrestrial providers want to preserve their core business and recapture voice minutes. And cable companies see voice as a massive opportunity to steal real revenue away from traditional voice and even cellular carriers.

And why not? There's already enough broadband capacity coming into the home to support a voice service over IP.

To be clear, we're not really talking about Skype or Yahoo! voice over the Internet, we're talking about a managed digital voice over IP service that operators want to provide over the DSL connection coming into the home. But there are problems: cost, performance and user expectations.

In the home, most people use lower frequency (eg. 900 Mhz) cordless phones that provide good range and tend have stronger radio propagation characteristics (they go through walls and stuff better). So there's been a high bar established. And if you're going to provide an IP-based voice services, it will HAVE to work over Wi-Fi.  So how many Wi-Fi phones are out there?  Not many.  But that's changing.

So called "dual-mode" phones (cellular/Wi-Fi), like Samsung's SGH-P200 or ZyXEL's P2000W VoIP Wi-Fi phone are hitting the market, and fast. So are new phone gadget thingys like Sony's new Mylo and Belkin's Skype Wi-Fi phone.

Over the next three years, In-Stat says that there will be over 66 million cellular/Wi-Fi phones in operation.  So get ready to say goodbye to your portable phone. You won't need it. What you WILL need is a Wi-Fi network in your home that can actually support voice.

Even though voice isn't as bandwidth intensive as IP-based video (IPTV), voice "timescales" are must more stringent than video.  Voice is typically a two-way, interactive conversation while video isn't.  And video buffers are typically bigger (in the seconds range) while voice buffers are smaller (in milliseconds) giving voice applications less latitude in the event of problems.

For voice to work in the home, Wi-Fi systems need to be able to adapt in milliseconds to changes in the RF environment.  And interference becomes a big problem. Presenting a stronger and more reliable Wi-Fi link to handsets, one that is less likely to be negatively effected by interference, becomes super important.  It also helps with battery life, another big issue.

The lower signal-to-noise ratios and fluctuating signal strength requires voice handset power to stay on longer and at higher levels.  And handsets will typically select a lower PHY data rate, taking longer to transmit "talkspurt" frames.  This results in shorter battery life and low user satisfaction.

Being able to maximize both the transmit SNR and receive sensitivity, handsets can transmit the same amount of information in shorter times, at lower power - producing longer battery life and higher user satisfaction

In other words, Wi-Fi devices constantly "probe" the air looking for better access points. The ability to reduce transmit power and retransmissions of handsets helps fix this. With high gain directional Wi-Fi antennas that also have automatic "directionality," handsets don't need to use so much power to transmit. And retransmissions, due to link errors, effectively go away.

Periodic, low-power synchronization between APs and handsets – even when a handset is not in use – also helps ensure no missed calls.

Finally Wi-Fi systems also need to be able to distinguish between voice and other traffic so they can give voice the attention it needs.  Conventional Wi-Fi QoS only specifies four queues (voice, video, data and background) but doesn't specify how to prioritize WITHIN each queue. Wi-Fi systems need to be smart enough to be able to prioritize multiple voice calls at the same time - in the presence of video and data. Per-station queuing is useful to guarantee that multiple handsets don’t interfere.

In a couple years, you will have a single phone that acts like a cell phone when you're not at home and a portable phone when you are.  It's inevitable.

Home Networking Factoids and Women in Wyoming

S2 Data Corporation recently issued an IPTV report focused on customer premise equipment.  It's the only we've seen that really delves into what's going on at home.

They assert that among the more critical challenges facing Telcos is the distribution (and management) of new digital services to multiple devices and locations within the home.  Robust home networking, predictable quality of service, high performance and ease of use unsurprisingly top the list of "key requirements."

According to their research (and we can only share so much), in 2005 the total of broadband subscribers reach more than  141 million worldwide and is expected to reach over 300 million by 2010.  But the U.S. remains behind relative to broadband penetration and ranks third out of five regions (after Europe and Asia) on the DSL subscriber forecast list.

A key driver for triple play services, like IPTV, is fiber-to-the-home (FTTH).  S2 Data forces 7.8 million FTTH worldwide subscribers at the end of 2005 but expects subscriber uptake to balloon to over 36 million by 2010.

The report is one of the most comprehensive looks into the home from both a qualitative and quantitative standpoint - differing from others in that the forecasts provided range from IPTV subscribers to the number of set top boxes with advanced networking that will be deployed - and everything in between.

Here were some of the more interesting findings:

• The U.S. is second to last in the IPTV subscriber forecast count, behind every region but "the rest of the world"
• IP-enabled set top boxes will explode from 1.6 million in 2005 to over 20 million by 2010
• Intelligent network interface devices (iNIDs) that combine the functions of a traditional NID with residential gateways will blur the demarcation between the home and operator networks but will give operators a chance to more easily manage and offer new managed services
• The cost of installing IPTV in a "typical" (if there is such a thing) three-room deployment takes approximately 3.5 hours with two installers (using CAT-5 cable) and costs US $421 (for a million subscribers that is $421 million)

So whether or not you believe the IPTV or triple play market has legs (until now it's been mostly over-hyped and under-developed), operators everywhere are placing massive bets and putting hordes of investment into it.

Even in, of all places, Wyoming where it's illegal for woman to actually come within five feet of a bar (it's true, see for yourself) IPTV is happening.