Wi-Fi law of reciprocity

So I was thinking about the Law of Reciprocity … It’s not like I and my barely-average IQ normally sit around and ponder things like that, but when I saw Connect802’s recent tweet on the reciprocity law, it was stuck in my head like a bad song.

This law, paraphrased, says, “what you give is what you get” — as it relates to antennas. Basically, Tx and Rx gain are the same. An antenna’s mouth is equal to its ears.

Since my brain is stuck in a continual game of Tetris, I eventually got around to contemplating how that relates to Ruckus’ funkadelic antenna array. The 802.11 protocol puts the kibosh on using a switched or dynamic directional array for listening to uplink data frames because an AP never knows from whence the next transmission cometh. This means that while uplink ACKs may be received using the array (immediately following a data Tx), omni antennas must be used for uplink data frames.

In a mental exercise that would make Forrest Gump proud, I figured if I’d come this far, I might as well keep on goin’. So that’s what I did. My mind ran all the way to the conclusion that their uplink range can’t possibly be as good as their downlink range. I highly doubt Ruckus will argue against this point. So, I tested it … and that reciprocity thing is broken, given that Ruckus uses two separate antennas for Tx and Rx (a set of omni antennas for uplink and a static switched array for downlink). You might call it link asymmetry in Ruckus’ case.

The issue then is that when you survey, you have to survey for the uplink range, and given that they use the same “normal gain” omni antennas, the same Atheros chipset, and the same radios (as it relates to radio sensitivity) as most everyone else on the planet, their claim of less APs to cover a given area cannot be altogether accurate unless they are surveying in such a way that uplink throughput isn’t as good as or stable as downlink throughput.

Their array, per their spec sheet, yields about 7dBi. I believe that to be accurate due to comparisons with other antennas. That gain is due to the downlink directionality of the array. The omni antennas wouldn’t be better than 5dBi @ 2.4GHz or 5GHz … likely less.

On the flip side of this argument (this is Devinator trying to be vendor-neutral), their downlink efficiency must be better than with omnis (IF the client is outside the max data rate coverage) because downlink signal strength would be higher due to the higher-gain (~7dBi) of the array. That means faster data rates and less airtime usage.

So, it seems to be an “improved downlink at the expense of the uplink” situation. I don’t believe that the improved downlink can make up for the suffering uplink when using bi-dir traffic because some apps require symmetry.

Granted, some networks have more downlink flow than uplink (like small offices and homes), and in those cases, Ruckus could no doubt perform just fine with the everyone-applicable, “if the network is designed properly.” I do think, however, that the type of apps plays a big role because many apps are bi-directional (voice, file transfers, etc.) or primarily uplink like backups. If these types of apps are in use, then the network would typically need symmetrical throughput capability.

Now please don’t misunderstand the purpose of this post. I’m not attacking Ruckus. They have good people and good gear. They are a good company, and I believe they will be successful in specific markets. I’ll also say that they’re smart enough to primarily focus on markets where their array works best.

I even use a Ruckus AP at home from time-to-time (to see where they’ve improved with each code release). Please notice that this isn’t a technical marketing monologue on the 50 things I think are wrong with Ruckus’ technology (because there’s clearly no more than 10 things wrong with it!) That’s a joke. Laugh with me!

The purpose of this post is to point out the only known attempt at violation of The Law of Reciprocity, which has seemingly been stymied by the 802.11 protocol. I just poked fun at Ruckus because they have a good sense of humor and will likely bury me in clever retorts as soon as I post this.

If I weighed everything into the equation, I still seem to come back to a static switched array borrows from uplink to deliver better downlink in order that Ruckus can better serve specific markets where that tradeoff is useful (specifically, in homes, where carriers are selling into).

What do you think? Is this line of thinking much ado about nothing? Is my mind wandering to places that don’t really matter? I submit that it’s entirely possible. I’ll also invite those who might know more than I know about Ruckus’ system and The Law of Reciprocity to weigh in on this topic so that I can learn if I’m off-base on this.

Devin Akin is Chief Wi-Fi Architect at Aerohive. Devin has over 10 years in the wireless LAN market and over 15 years in information technology. Devin's background includes working as a network design engineer for EarthLink, AT&T/BellSouth, Foundry Networks, and Sentinel Technologies as well as working as an RF engineer in the US Army. He has authored and edited several books with Wiley-Sybex and McGraw-Hill and holds some of the industry's most esteemed certifications, including CWNE, MCNE, MCSE, CCNP, CCDP, CCSP, and INFOSEC. He is considered an authority on Wi-Fi.