Powerline: More Reliable Than Wireless? You've Got To Be Kidding...Yes?

I approach this particular writeup with no shortage of trepidation, because I don’t want to discourage any of you from commenting on my posts in the future. However, every once in a while someone says something so notable that it begs for follow-up commentary from me. And anyway, I assume you all realize that even if I don’t explicitly showcase your prose, it’s available for the world to peruse and derive an opinion of…a private email is always an alternate option if you want to keep the correspondence just between the two of us.

2.5 weeks back, I reported on my inability to get either a fully HomePlug AV powerline or fully 802.11n wireless source-to-destination network span to reliably transport high-definition video, and the consequent need for a hybrid-technology approach. The blog series generated a fair bit of comments traffic, although considering that all of it was concentrated at the part 1 post point, I can’t help but wonder how many folks read all the way to the end of part 2 Anyway, for today I’d like to focus your attention on the thoughts of ‘cgorange’:

Powerline technologies are improving significantly in terms of performance, reliability and ease of use. Compare the customer ratings for Powerline products to wireless products on Amazon and you’ll see what I mean — latest generation Powerline products typically have a 4.5 star rating and above, while even high-end routers from top-name manufacturers get a 3 to 3.5 star rating. Why? The time spent going from having a problem to resolving the problem is significantly faster with Powerline than wireless. While people with tricky wiring or extreme bandwidth requirements (e.g. multiple simultaneous HD streams) may not find Powerline useful — for the other 95% of users, it works great! And, if you’re one of those 5% people — you should probably look at MoCA or CAT5.

First and foremost, although I can’t say with absolute certainty because ‘cgorange’ frustratingly didn’t identify him/herself, given the ‘hype’ tone of the discourse it seems pretty clear to me that this person has a vested interest in powerline’s success. I don’t know about you, but whenever I read something with such obvious bias, my reaction is counter to the writer’s desired end result; my ’skeptical gene’ kicks in and I start picking apart the posited arguments in an unbalanced fashion without also considering their merits as I would with a more pragmatic proposal.

I actually haven’t picked up the Amazon gauntlet that ’cgorange’ offered, because it’s emblematic of the ‘Data can be selected, sorted and interpreted to justify any possible desired outcome’ point that I made in mid-January. In other words, even if the powerline-to-wireless ratings comparison is quantifiably accurate, it’s irrelevant. For one thing, the relative market sizes are vastly disproportional, making any comparison attempt statistically meaningless. Two weeks back Intellon announced it’d shipped 26 million HomePlug ICs to date. Considering that this claim encompasses HomePlug 1.0, HomePlug 1.0 Turbo and HomePlug AV (where, in all cases, Intellon has to date been the overwhelmingly dominant IC supplier), I’m going to estimate that total to-date powerline IC shipments are on the order of 50 million units (thereby also encompassing DS2’s UPA technology and Panasonic’s HD-PLC products).

How does this stack up against Wi-Fi? Stephen Palm from Broadcom told me two Fridays ago that his company alone has to date shipped more than 350 million 802.11 chips. And one day earlier, Atheros PR representative Greg Wood reported:

Atheros publicly announced, last June at COMPUTEX, that the company had shipped more than 200 million WLAN chipsets to date…It’s safe to assume that we are well past that mark. Last year, 2008, was the biggest revenue year of the company’s history.

Keep in mind that these Wi-Fi shipment statistics don’t include Intel’s Centrino wireless chipsets, nor do they encompass the supply contributions of other (smaller) 802.11 vendors.

And where are the powerline and wireless ICs respectively being used? The answer to this question points to another flaw in the logic ‘cgorange’ attempted to employ. As ‘cgorange’ rightly points out, the preponderance of Wi-Fi gear sold on Amazon comes in the form of a wireless-inclusive router. A router, for any of you who’ve either designed or implemented one, is a fairly complex piece of equipment, handling both wired and wireless LAN traffic (along with hand-offs between the two), as well as the connection to the WAN. Conversely, the bulk of powerline shipments end up in comparatively straightforward CAT5-to-powerline bridge adapters, whose simplicity would presumably reflect in higher user ratings (although not so in my particular case…more on that in a bit).

Next question: how are the powerline and wireless ICs being used? In answering this one, I’ll first point out that in the referenced writeup, I found shortcomings with both HomePlug AV and 802.11n, which ultimately necessitated my adoption of a hybrid LAN topology that incorporated both of them. That ‘cgorange’ incorrectly discerned a ‘powerline is bad, wireless is good’ conclusion is again suggestive of the pre-existing bias with which he or she interpreted my words. However, whether the networking adapter employed by any particular LAN client is powerline- or wireless-based, my contention is that the vast majority of consumers are today using their networks for nothing more complex than latency-tolerant and protocol-limited casual web browsing, email, file access and transfer, and (perhaps) shared printing. Conversely, my study is suggestive of where the bulk of consumers will (hopefully) be in the not-too-distant future, desiring to route isochronous (therefore latency-intolerant) and protocol-diverse audio (including VoIP) and video streams around their homes and offices. That I’ve stumbled across problems that most users haven’t yet experienced is therefore no surprise.

Finally, I’ll take issue with the suggestions of ‘cgorange’ that ‘Powerline technologies are improving significantly in terms of performance, reliability and ease of use’, that ‘The time spent going from having a problem to resolving the problem is significantly faster with Powerline than wireless’, and that ‘people with tricky wiring or extreme bandwidth requirements (e.g. multiple simultaneous HD streams)’ only represent 5% of the potential customer base. Let’s tackle that last point first. As my earlier article points out, there is no such thing as a single HD stream between one LAN client and another, unless the path from the source to router and the path from the router to destination employ different (therefore non-bandwidth-contending) networking technologies. Also, to be clear, Wi-Fi isn’t blameless in this regard; I’ve dealt with plenty of past problems, including RF interference and attenuation caused by reflective tiling and metallic objects, Faraday Cage-ish chicken wire in walls, moisture-rich construction materials, and other factors.

And if I’d only wrestled with one particular powerline technology, with one particular protocol, and/or with one particular installation setting, I might be more inclined to accept the rosy picture that ‘cgorange’ is attempting to paint. However, consider that I’ve been dabbling with numerous powerline technologies for approximately five years at this point, at three different home-based offices of various ages and power grid characteristics. And consider, too, that as a ‘power user’ with an in-depth knowledge of electronics, I’m both far more determined-to-succeed and far more able to debug and resolve problems than is the typical consumer. Over this lengthy time span, I’ve stumbled across innumerable issues, such as:

1. Susceptibility to power grid noise: motor-inclusive devices such as heaters, air conditioners and stand-alone fans, power tools, hair dryers, vacuum cleaners, flourescent lights, and even the switching power supplies inside battery chargers and other AC/DC adapters.
2. An inability to operate behind a data-filtering surge protector (a function that’s increasingly being built directly into wall outlets) or UPS.
3. Susceptibility to impedance and orientation effects: an adapter might work well when directly plugged into a wall outlet, but not through an extension cord (or visa versa), or work when plugged in ‘right side up’ but not ‘upside down’ (or, gain, visa versa), or not work in one outlet but work in a nearby neighbor on the very same circuit due to varying wiring lengths to the circuit breaker box. Speaking of which..
4. Either complete loss of connectivity or greatly diminished bandwidth when two adapters end up on different 110V phases of the incoming 220V premises feed from the street.
5. An inability to pass various protocols, both obscure and common, across the CAT5<->powerline bridge.
6. Incompatibilities with various routers, switches and other legacy networking equipment.
7. Erratic and frail powerline network recovery from premises power loss (I experienced this again just a few days ago), and
8. Varying behavior in all of the above regards, both across time-of-day and time-of-year

I wonder how many of the folks who gave the powerline equipment they bought on Amazon a 4.5 star rating upfront are still singing the same tune six months down the road?

The ability to plug a piece of equipment into a power outlet and have it automatically also join the network (in a performance- and protocol-robust fashion, mind you) is a mighty compelling proposition. As such, I believe I’ve been one of the most (if not the most) enthusiastic followers in the technical press of powerline networking’s evolution over the years. But all of this attention, I realize, is a mixed blessing for the technology’s promoters. If something works well, I’m happy to pass the word along. Conversely, if it doesn’t, I’m not shy about telling my readers about the problems.

While I appreciate your enthusiasm, ‘cgorange’, don’t try to sugarcoat the powerline network situation. Your credibility, and ultimately your advocated technology’s viability, will benefit from a more pragmatic approach.