Anaheim Muni-Fi and Earthlink ready for Wi-Fi Phone Beta Testing - We need details!

The New Earthlink VOIP service is now ready for beta testing. Users in Anaheim may sign up now and get a free handset with the service during the tsting period according to Glenn Fleischmann over at WiFiNetNews

Good news for those of us watching the VoWLAN emergence but I have some questions:

It appears as if the service is tied to their phone. There is this quote in the release, "'What separates our Wi-Fi phone from others is its ability to work over EarthLink's municipal Wi-Fi networks,' said Steve Howe, EarthLink's senior vice president of voice."

• Does this mean that other SIP phones or dual band phones will not work?
  
• Is this an attempt to control the hardware and service.
  
• Has anyone tried using a different phone over the Anaheim network?

If this is true and we are locked into an Earthlink/Accton solution, it could spell doom for rapid adoption and raise the hackles of the "Free the Airwaves" folks.

Later on it also mentions that the Accton system is and ATA (Analog Telephone Adapter) based system. This would be similar to Vonage's home adapter. Again, more questions, can you use the handset while roaming about the network or just when in range of your own ATA base station?

Lastly (and in my opinion, most importantly), is it encrypted or are folks going to be able to sniff my call and play it back with VoiPong or something similar? Why do people always think of security last?

There are many questions we still have no answers to. If you know any, drop me a line.

Building a Voice Capable WiFi Network

Building a wireless network that supports data traffic is hard enough but trying to support VOIP over your WLAN (also known as VoFi) can be a nightmare. To make matters worse, when you ask your vendor how to make Voice work on your WLAN they explain you will need 2X-3X as many APs as you needed for data. "Sure I do", you respond. Confident that the sales person from your vendor just wants to sell you more APs. Or, better yet, you turn to your trusted VAR and he suggests another site survey. "Right, another one", you say, with that knowing look in your eye and a sinking feeling that you are being strung along. You feel like the guy who brings his car in for a tune-up and gets told he needs a complete overhaul.

Well, I have nothing to sell you and no agenda that I will benefit from by saying this but your infrastructure vendor and your VAR are absolutely correct. You probably will need more APs and you sure as heck will need another survey. Lets find out why, shall we?

Unlike Email and web access, slight lags or delays in traffic or small losses in connectivity will completely destroy calls. A person who has access to the Internet durring a meeting in a conference room is far less likely to lose his cool for small delays than when he is on the phone with an important client.

You see, wireless handsets are much lower powered compared to the access points they talk through. A typical AP is usually set to communicate at 100milliwatts (mw) whereas the typical handset is roughly 5mw. This makes it very easy for the handset to hear the AP but very hard for the AP to hear the handset when it is far away. Also they are far less resilient to fragmented packets, retries, packet loss etc.

So what can I do? Well the simplest thing to do would be to ensure that the handset is always at the same power as the AP. That means either increasing the power on the handset or, more likely, lowering the power on the AP. This will mean, of course, that you will need more APs to cover the same area.

For example here are 4 APs at 100milliwatts:

Here are the same APs but now set to 5mw instead, notice the gaps in coverage:


In order to compensate, we must add many more APs to fill in the holes, all configured to run at 5 mw:


As you can see, much better. Now, though, our main issue is channels. APs that overlap thier signal on the same channel take away from the usable bandwidth. We want to ensure we do not trample the signal from another AP so we must adjust the channel plan.
Also, remember we only have 3 channels to work from.

Cisco, at this point recommends the following:


That explains why I limited the seen signal to -67dbM making all the other signal fall off and appear grey.

In a week or two, we will discuss debugging Voice issues and setting MOS scores.

Repost: Why Would Anyone Think Wireless Is Easy?

I really love this post over at Tech Dirt. It illustrates the fallacies that most people have about implementing Wireless VOIP. The VOIP guys think all they have to do is plop down an AP and shebang! VoWiFI. http://news.techdirt.com/news/wireless/article/6905

I especially love this sentence, "When we begun developing the mobile phone version we didn't realize the number of technical obstacles. It is challenging and is taking much longer than expected"