802.11ac Draft 2 Ballot Update
Last Thursday, the second draft of 802.11ac went out for what’s called a letter ballot. (I think the term “letter ballot” is an anachronism from back when ballots were actually mailed, with stamps!) To submit the draft to letter ballot, all the comments from the draft 1.0 ballot had to be resolved.
Just as with the first draft of any standard, a significant focus of comments is resolving small ambiguities. With complex specifications, the number of places for ambiguity to creep in can be quite large. As a result of the comment resolution efforts of the past half-year, 802.11ac draft 2.0 has a page count that’s about 40% higher than the first draft. (This increase is very similar in percentage terms to the length increase between the first and second drafts of 802.11n.)
One small but important change was made regarding encryption protocols. In the first draft, 802.11ac specified that the sole encryption protocol for use with the multi-gigabit was CCMP, the protocol that is the cryptographic core of today’s WPA2 certification program. It is a secure protocol, but its design imposed limits on throughput that were not consistent with multi-gigabit speeds. In response to several comments in the first ballot, the new 802.11ac draft added GCMP (the Galois/Counter Mode Protocol), an efficient cryptographic system whose parallelism and efficiency are well-suited for high-speeds.
Or, if I can put it more simply:
1. 802.11ac draft 1.0 prohibited the use of GCMP
2. Due to the performance limitations of earlier protocols, 802.11ac draft 2.0 allows the use of GCMP
3. GCMP is needed for very high speeds, so make sure you know how you’re going to get there as your 802.11ac network evolves.
As a network administrator, you’ll get a choice: you can continue to use CCMP for the first wave of relatively slow 802.11ac products, but when it comes time to “push the pedal to the metal,” you’ll be glad that GCMP is there to support the burst of speed. The only downside to all this speed is that any network equipment that doesn’t support GCMP will need to be replaced – and there’s a fair bit of it out there, since many early wireless security chips (such as the Cavium Nitrox) are limited to support of CCMP only.
There was also significant text revision around beamforming, which is especially important given the focus on multi-user MIMO for future speed increases. Beamforming in the 802.11ac draft is staying simple, which is exactly what we need. Although 802.11n was the first standard to incorporate beamforming, there were a large number of options for implementers to choose from. In the resulting confusion, most vendors chose not to build any beamforming product features. By simplifying the standard, 802.11ac will enable widespread use of beamforming in radio chips. Although 802.11ac will be fast on its own, multi-user MIMO will enable a whole new level of spatial re-use.