Deploying 802.11ac: Notes from the Field

A survey we’ve been conducting at Network World’s Wireless Infrastructure 2014 events confirms some key predictions on how organizations are deploying 802.11ac – but with a few surprises as well.

We just wrapped up the first four stops on the Wireless Infrastructure 2014 tour, just a few of the hundreds of events, by the way, that Network World produces every year. We've been concluding each of these with a workshop segment, exploring the different ways in which 802.11ac can be deployed within an enterprise or organizational setting. There are a good number of considerations here, most notably cost, but also including logistics, timing, and risk. As it turns out, we were able to boil down all of the possibilities into five key options. I've written an overview of these in an article for the Wi-Fi Alliance, which you can see here. Take a look, and then we'll see what the workshop participants thought. Go ahead, I'll wait.

Now that you're back, the process we used was to invite participants to self-select into groups matching each of these areas, with the exception of the ignoring-the-issue-altogether case, on the assumption that IT professionals wouldn't even consider such. Here is a summary of thoughts on each of the major directions noted in the above article:

  • 802.11ac Overlay - I had assumed that this would always be the most popular option (perhaps because it's the one I like best), but such wasn't reality. Supporters did make my case, though, nothing that .11n client performance would likely be enhanced, power-user hotspots could be deployed non-disruptively, vetting of and experience with .11ac is required before a major commitment to a given vendor can be made (pilot deployments are a great way to manage this), costs for volume deployments would likely be lower later, assurance is included in most WLAN system vendor products, and, while assurance and high-demand situations can be managed now, volume could in fact be addressed when Wave 2 reaches a suitable level of maturity. The gradual, conservative nature of this approach had universal appeal among its supporters.
  • Rip and Replace - Among the ideas justifying this approach were the need to spend budgets that will otherwise expire (yes, this still happens in some organizations), re-deploying replaced APs in locations with less demand, keeping the replaced units as spares, taking advantage of vendor financial incentives (trade-in and upgrade programs and such), and some even mentioned that they have an internal mandate to deploy the latest technology as soon as possible. Some did counter that an immediate wholesale rip-and-replace would be high risk, with some suppoorters countering that pre-determined hardware refresh cycles were a great opportunity for Rip and Replace to be implemented. And many noted that .11ac clients bases are growing rapidly, a trend that will most certainly continue, so having support for these in place was viewed as an obvious requirement by many supporters here. Finally, some mentioned that .11n is clearly approaching the end of its useful life, and, given backwards compatibility, new .11ac APs represent the best option for now and later.
  • Assurance Only - I've argued that at the very least an assurance (IDS/IPS, rogue detection, and spectral analysis, among other functions) deployment of .11ac is essential, as .11n sensors won't know what to make of .11ac signals. Participants noted that such a strategy has minimal cost (we generally assume a 1:4 to 1:6 ratio of sensors or APs deployed as sensors to APs deployed for access), the timing of Wave 2 was close enough that provisioning assurance today and waiting for Wave 2 could work, the fact that .11ac clients are scarce enough today that access can be ignored for now, the relatively lack of maturity in current .11ac products, the need to gain experience with .11ac access in relative isolation before deploying it in mission-critical settings, and the fact that that an Assurance strategy has the lowest cost of any of these options. Some also mentioned that APs deployed as sensors could be redeployed, in whole or in part, for access later.
  • Wait for Wave 2 - Some participants noted that the need for Wave 1 isn't yet established, so waiting seems to be the optimal approach. Wave 2 products, assuming they follow the traditional path of IT products, should have much improved price/performance and perhaps even lower prices overall. The ROI of current installations is not yet complete in many cases, and many existing .11n APs are most certainly not at end of life yet. Despite growing demand, many noted that their .11n networks are far from saturated, and some also worried about increased operating expense (OpEx) - training, support, spares, and related costs. And some mentioned that just don't have budget for any additional equipment at present, and have to be very careful regardless with every dollar in their CapEx budget.

To be fair, most shops will likely use a customized hybrid of the above - while I'm suggesting that 802.11ac access deployments are a good idea today based solely on improved 802.11n client performance and improved overall price/performance, assurance is also required, with greenfield deployments most certainly going with 802.11ac or .11n APs being replaced by 802.11ac elsewhere. I can't see waiting for Wave 2, which implies doing nothing in the interim, because we're not sure when Wave 2 will arrive in a form that meets the rather lofty expectations that have been set here. My personal preference, though, is to deploy .11ac in addition to existing APs - the Overlay model - thus leaving existing 2.4 GHz. and .11n deployed at 5 GHz., if any, in place. I thus prefer APs with multiple .11ac radios, and these are not at all common today. I mean, why deploy more 2.4 GHz. capacity, or replace it with the exact same level of function? 2.4 Ghz. subscribed, provisioned, and not applicable to .11ac - and thus clearly not the future of WLAN technologies or products. While I don't expect 2.4 GHz. to be abandoned, as was the case with 900 MHz., it's evolving to legacy use only.

Deriving the precise 802.11ac strategy for any given organization, as you can see, can be complex. The bottom line for me here, regardless, is to get on the .11ac bandwagon now. From my perspective, waiting is not an option.

To comment on this article and other Network World content, visit our Facebook page or our Twitter stream.
Must read: Hidden Cause of Slow Internet and how to fix it
Notice to our Readers
We're now using social media to take your comments and feedback. Learn more about this here.