Even though the IEEE 802.11ac standard for gigabit Wi-Fi has yet to be finalized, that fact doesn't seem to be holding back key players across the wireless LAN equipment value chain. Everyone from chip suppliers to vendors of access points for both residential and enterprise-class applications are ramping up.
|The Wi-Fi-enabled world|
|Hotspot 2.0 promises to redefine the Wi-Fi experience|
|Wireless LANs: Assuring optimal performance|
|11ac will be faster, but how much faster really?|
|Don't sweat 802.11ac Wi-Fi -- because 802.11ad will knock your socks off|
The controversy here resides in the basic underlying technologies of .11ac and what value they might ultimately offer organizations that have spent handsomely on 802.11n over the past few years and are hoping to continue to gain return on those investments. But an excellent case can be made, as we'll explore here, for the eventual replacement of 802.11n by .11ac, provided the vendors can deliver demonstrable benefits in 802.11ac-based products, and that IT organizations take a few simple steps to prepare their networks for the enhanced data volumes that will most certainly be realized over the next few years no matter what 802.11 technology is in place.
The enticement for 802.11ac? Well, for starters, throughput numbers like 1.3Gbps, and the standard will, in fact, specify performance all the way to 6.93Gbps. But, as has been the case since the early days of 802.11n, raw throughput should not be the key motivator for the adoption of higher-performance WLANs. Rather, it's important to think instead in terms of capacity -- making the most efficient use of radio spectrum in the service of a large, diverse and growing base of users and applications, rather than provisioning gigabit-plus speed for any single user alone. And .11ac makes significant strides here.
With BYOD driving the new normal of multiple simultaneous devices per user, and WLANs now supporting essentially every IT application, including many that require time-bounded services, enterprises everywhere are well-served to contemplate upgrades in terms of how best to handle that ever-growing traffic load.
And for many, and eventually most organizations, that vehicle will ultimately be 802.11ac. And the decision here will be easy -- not only will capacity be improved, but we expect that the price of .11ac-based products will be essentially the same as that for .11n, meaning price/performance will be substantially improved as well.
Key technology advances in 802.11ac
802.11ac is really more evolutionary than revolutionary, building upon technologies proved in 802.11n. Nonetheless, five key technical enhancements at the core of the standard stand out:
* 256-QAM: This term is an abbreviation for quadrature amplitude modulation with 256 points in the phase/amplitude constellation. Without delving too deeply here, what this means is that 8 bits of encoded channel information (which is not the same as user data bits) can be sent during each clock period of transmission. 802.11n was much less efficient, with only up to 64-QAM specified.
There is, however, controversy here: It is unclear just how often 256-QAM will be effective in any given environment. The more complex a radio signal is, the more likely it is to be damaged during transmission and thus require a retransmission, hardly the key to efficiency. Nonetheless, 256-QAM is a proven technology and we do expect to see its benefits in .11ac in many cases.
* 80MHz and 160MHz channels: The key to throughput, as anyone working in any area of communications knows, is the availability of more bandwidth, referred to as spectrum in the wireless world. Spectrum and its use is heavily regulated by governmental authorities around the world, and just because the spectrum involved in WLANs is unlicensed doesn't mean that it's unregulated -- it's not.
The most important metric here is how much transmit power is allowed. While 802.11ac introduces 80MHz and even 160MHz channels, upping the 40MHz maximum of 802.11n, no additional transmit power is likely. This means, as is the case with 256-QAM, more potential throughput per unit of time, but likely less effective range for that maximum throughput. Still, we expect to see broad benefits here, especially with denser deployments of APs in the 80MHz case. We do not, however, expect to see 160MHz channels in general use anytime soon.
* Standard beamforming: More properly called beamsteering, beamforming is the art and science of aiming a radio transmission in a particular direction. While this can be easily accomplished with directional antennas, the use of omnidirectional antennas in WLAN systems is essential because one station must be able to communicate with another that might be anywhere in the 360 degrees around it.
As it turns out, sophisticated software can be used to determine the relative location of a receiver, and then carefully bias the timing of transmissions from two or more antennas so that transmitted radio waves reinforce one another in a particular direction. This can increase both range and reliability, and thus throughput and capacity. Beamforming has been available in 802.11n implementations for some time, but it's standardized in 802.11ac.
* New modulation and coding schemes (MCS): An MCS is a combination of parameters that ultimately determine the transmit data rate, and thus the effective data rate assuming any given transmission is successfully received. MCS parameters include modulation (to 256-QAM), coding rate (data bits vs. channel bits; more channel bits improve reliability at the expense of throughput), number of spatial streams (from 1-8), channel bandwidth (20/40/80/2x80/160MHz), and guard interval (400/800 nanoseconds, the amount of latency enforced between transmissions to allow multipath reflections to fade away).
Various combinations of these yield a range of raw throughput from 6.5Mbps to 866.7Mbps per stream. And, of course, in practice the MCS applied can vary from moment to moment as vendor-specific radio resource management logic makes adjustments to these and other parameters in response to changes in both traffic and radio conditions, as well as user-specified configuration options.
* Multi-user MIMO (MU-MIMO): An absolutely new, unique and potentially very important feature of 802.11ac, and perhaps even its killer app, is multi-user MIMO. All 802.11 signals to this point have been directed to a single receiver. The Wi-Fi channel is thus serially reused, with all other stations requiring service having to wait their turn.
Since that 1.3Gbps (and potentially higher-speed) channel might be far more than required in any given transmission from an AP to a station, .11ac specifies the ability to transmit, at data rates lower than the maximum, to multiple clients simultaneously. For example, a three-stream AP might send three separate and distinct transmissions to three single-stream receivers all in one transmission cycle.
Since we expect the majority of 802.11ac-equipped subscriber devices to be single- and two-stream for some time, MU-MIMO could be a huge step forward for effective throughput. We'll have to wait, however, for the second generation of .11ac chips, known as Wave 2, to use this feature. And, yes, beamforming can be applied simultaneously with MU-MIMO.
A summary of maximum throughput for popular channel sizes and stream counts can be seen in the table below. Note that 802.11ac offers significant potential improvements over 802.11n across the board, and, depending upon implementation, 802.11ac APs operating in .11n mode may even offer better service than today's .11n radios. We expect that WLAN chipset vendors going forward will incorporate gains in their experience into .11ac chipsets, rather than issuing new .11n-only products.
Preparing your network for 802.11ac
The ultimate success of 802.11ac in any given installation will of course be a function of the quality of the radio environment, but also depends upon considerations in the remainder of the network value chain. Specifically, we recommend the following:
* Cabling: It's a good idea to pull two Cat 6 (or greater) cables to any location where a new AP is required. This isn't because .11ac will require greater than 1Gbps connections, at least not anytime soon, but rather because we believe a strategy based on dense deployments will continue to be desirable.
Density in AP deployment is analogous to overprovisioning in wired networks, and an excellent strategy to handle ever-increasing demand generated by the adoption of BYOD, streaming media and cloud-based IT strategies. It's also desirable to purchase any new switches with support for 802.3at Power over Ethernet (PoE). While not all .11ac APs will require PoE at that level, we believe many will, at least for optimal operation, and this is one challenge that's easy to address.
* Switch, router and backhaul capacity: Similarly, network management logs should be consulted to identify any bottlenecks in wired and especially core network elements. Remember, demands on networks only increase over time, and it's not possible to get the best return on an investment in .11ac if traffic is congested elsewhere in the network.
* Initial deployments: We're often recommending that initial 802.11ac deployments use 40MHz channels so as to more easily fit into existing Wi-Fi channel plans. Organizations that have not made much use of the 5GHz bands (.11ac is 5GHz-only) will have an easier time deploying 80MHz channels, but even 40MHz channels offer 33% (peak) greater throughput than 802.11n.
As many initial .11ac clients are expected to be handsets and tablets, 40MHz channels may be more than adequate in meeting service demands for the time being. We also recommend that initial .11ac deployments occur in a greenfield setting if possible (a previously un-deployed building, for example), so as to give IT staff time to become familiar with the technology in isolation from other production users.
It's also not too early to define your next-generation WLAN -- and unified wired/wireless -- management strategy. Management consoles are also becoming key product differentiators, so look for continuing and valuable innovation here.
A timeline for 802.11ac adoption
The chart at right shows our forecast for 802.11ac adoption. Many products are on the market today, and we're expecting essentially all enterprise-class WLAN system vendors to have products this year. Critical mass, where more .11ac than .11n is sold, occurs in 2015 thanks to earlier-than-anticipated availability of Wave 2 chipsets, these including MU-MIMO. Don't hesitate to continue .11n purchases, though -- we don't expect wholesale replacement of .11n by .11ac until into 2018, more than enough time to convince both users and finance departments that ROI won't be a problem.
Finally, 802.11ac, like LTE and LTE-Advanced in the cellular world, ushers in the era of sufficiency, when sufficient availability, capacity and other measures of performance remove any vestigial barriers to the deployment of any application on the WLAN. The wireless LAN thus continues its evolution from nice-to-have, to primary and default access, to preferred or only access for any mobile worker anywhere.
An era of relative stability on the standards and technology front will also spur, we believe, massive sales of .11ac, as IT departments appreciate the benefits and need consider only minimal risk. The attention of IT managers will thus shift up the protocol stack, to a focus on ever-more functional applications that yield new productivity for end users. And, if the LAN isn't enabling productivity, then why have it in the first place?
The Farpoint Group is an advisory firm specializing in wireless networking and mobile computing. Founded in 1991, Farpoint Group works with technology developers, manufacturers, carriers and operators, enterprises and the financial community. Craig is an internationally recognized industry and technology analyst, consultant, conference speaker and author, and is the writer of the Nearpoints blog. He is also the chairman of the Mobility track at the Interop conferences. Craig holds a bachelor's degree in computer science from Brown University, and is a member of the IEEE and the Society of Sigma Xi.