Let's Build Microcell Networks

I've  had an AT&T microcell sitting on my desk for two weeks now.  I've had all kinds of stupid solutions over the years to get cellular coverage in my house.  None of them have worked particularly well.  I was skeptical about the Microcell and prepared to be disappointed, but so far it seems to be working reasonably well.

My first effort was a wireless radio repeater.  With one antenna on the room and another in my office, I did indeed get five bars, but calls dropped and it didn't seem to work reliably.  I suspect that there was never enough shielding between the two antennas to avoid feedback, despite placing them at opposite ends of the house with a wire across the roof.

My second try was by using T-Mobile's UMA service.  It worked quite well, and I particularly liked being able to use it like a local phone overseas with good WiFi coverage.  The solution was not without problems.  WiFi seemed to garble the calls and switching between UMA and GSM coverage happed frequently and often resulted in dropped calls.

The AT&T solution is not perfect.  There are no good transitions from micro-cell to the main network (yet) but the quality of calls and coverage is good.  If your DSL is poor quality, you can also set the system to operate in priority mode, prioritizing your voice traffic above other traffic.  So far, I have not needed to do that.  Voice quality is good and data is not used at home because the WiFi connection.

I've also added my neighbor to my Microcell and for the first time in years, he's been able to get coverage at his house.  And I think this might just be the wave of the future.  For people who buy Microcells, why not use them to knit together extended coverage for the wireless network.

The fact is that some areas are never going to be covered well.  A recent remark that it takes up to two years (or more occasionally) to approve a new cell site in San Francisco (read) means that progress will be slow.  With user-generated coverage, we could close these gaps in many areas quickly.  The key elements of the solution could be as follows:

• Fix hand-offs between microcells and with the main network
• Validate coverage and backhaul performance for new microcells before admitting them to the main network (e.g. if you are connected to unreliable broadband, you cannot join the shared microcell network)
• Cap data transfer speeds on microcells so that multiple local users do not have a big impact on usable DSL speed.  (For people with bigger connections like Comcast's 50 megabit, it should not be material from phones, but if you have AT&T's own local DSL at 6 megabits, it could be a real issue)
• Reward users to agree to make their microcells open to all other network users with a discount on their monthly bills

This could lead to very rapid coverage improvements in urban and suburban areas.  Of course, adding a couple million cells to your network could have an impact on your switching functionality and systems.  I don't understand how it could be fully implemented, but it can't take any longer that it takes the city of San Francisco to approve a new cell site, can it?

San Francisco.  Beautiful.  Hard to cover.  Photo: Flickr cc Paraflyer