Brian's Brain

Powerline networking, whether as a means of potentially bringing Internet access to the masses (along with power usage monitoring and control to them and their utility companies) or for connecting LAN clients to each other and to the router nexus, or both, has compelling (some might even say Siren-like) allure. How else can you explain the sustained, steady and strong interest and investment in various powerline technology candidates, in spite of innumerable issues with the resultant implementations?

There's something fundamentally compelling, I continue to say in spite of my many scars from past hands-on project frustrations, with the ability to send an Ethernet 'heartbeat' up and down the same wiring that's already in place to provide electricity to a widget. And recent news on both the WAN and LAN fronts extends the repeated boom-and-bust pattern that those of use who are veterans of this technology already know intimately.

Take, for example, a segment on rural broadband from Monday's NBC Nightly News:

It exemplifies the experiences of my mom, who lives only 5 miles (by car, more like 3 miles 'as the crow flies') from the center of a Northern Indiana town (population 46,000+) but who isn't able to obtain either cable, DSL or fiber broadband access. As such, I had to set her up with a comparatively expensive and severely metered-usage cellular data service plan.

Expanding rural broadband is a key tenet of the economic stimulus packages currently being debated in the US Government, although the extent of the problem is not clearly understood. While the above NBC News clip claims that 60% of Americans don't have access to affordable broadband (note the emphasis; NBC doesn't define the term), another recent study concluded that 2/3 of US homes subscribe to broadband of some kind.

Running copper, fiber and coax physical cable is an expensive proposition, regardless of whether it's being funded by service providers, federal, state and local government coffers, or some combination thereof. WiMAX wireless access shows potential but its rollout is slower than many would prefer, with competitive LTE falling even further behind its original over-optimistic forecasts.

Enter BPL (broadband over powerline). Scourge of the amateur radio community, many of whose members have publicly commented on my past posts on this subject, its formal approval was halted last year when an appeals court determined that the FCC had suppressed data contrary to the desired positive conclusion. Yet BPL's low-investment, high-return rollout potential remains so alluring that the FCC is trying again, vowing greater transparency and cooperation with skeptics such as the ARRL (American Radio Relay League) this time around. Detractors of the FCC's embrace of White Spaces Technology can be forgiven for any déjà vu they're feeling right about now.

Now let's switch gears and look at the latest in LAN-based powerline developments. I'm talking, of course, about Belkin's F5D4076 'Gigabit Powerline HD' adapters, introduced a month back. They're based on Gigle Semiconductor's GGL451, which supports both HomePlug AV and Gigle's proprietary mediaxtreme mode. As those of you who follow my professional Twitter account may already know, I met Gigle for dinner last month, where company officials gave me the technology lowdown along with a few adapters to try for myself.

Whereas HomePlug AV employs a 2-28 MHz power grid spectrum footprint, mediaxtreme leverages the 50-300 MHz broadcast band. As I recently wrote in the first draft of my upcoming August 20 cover story:

Like 5 GHz Wi-Fi versus 2.4 GHz 802.11, mediaxtream's higher-frequency reliance delivers higher performance potential. Indicative of this fact, the F5D4076 includes a 1 Gbps Ethernet transceiver, whereas consumer HomePlug AV adapters belie their '200 Mbps' marketing claims by only embedding 10/100 Mbps PHYs. However, again as with 5 GHz vs 2.4 GHz wireless, Mediaxtream has notably shorter usable range than HomePlug AV.

In order to increase usable range, adapters can find use not only as powerline network nodes (i.e. Ethernet bridges) but also as power grid-based repeaters. There's a store-and-forward performance degradation to such a topology, of course, but the end result may still be a net sum gain versus HomePlug AV speeds. In my initial testing, with five adapters in operation (one at my Wndows Vista Ultimate-based laptop, one each at the entertainment centers in my living room and bedroom, one at my INSTEON home automation controller, and one at the router), the Media Center UDP-based streaming experience was robust for approximately five minutes...at which point the PC-to-Xbox 360 connection dropped and refused all resurrection attempts.

As it turns out, there's an unadvertised four-node limitation on the initial production F5D4076 firmware release. Dropping down to only three adapters resulted in a stable connection and the following Network Performance Tuner analysis plot (HDTV = 22 Mbps, 'Acceptable for TV' = 8 Mbps):

Compare it against this five-node HomePlug AV configuration using NETGEAR XAV101 adapters:

and you may conclude, as I did, that not only do the Gigle-based adapters not deliver higher performance than HomePlug AV, they actually run at a bandwidth (not to mention node-count) deficit.

The initial production firmware in the Belkin adapters selects either HomePlug AV or mediaxtream mode, depending on the power grid characteristics determined at power-up by the ICs' embedded DSPs. Gigle assures me that the company is hard at work on firmware improvements, both to increase the number of supported nodes, to improve standalone HomePlug AV and mediaxtreme speeds, and to bond the HomePlug AV and mediaxtreme channels together versus a more elementary either-not-both approach. I'll report updates via posts here at the Brian's Brain blog.