It has been said that there are two certainties in life – death and taxes. If you said the same about technology it would read ‘confusing acronyms and over hyped performance claims’. Although 802.11ac, the latest in a long line of Wi-Fi standards, has been in development for several years it was finally approved in January and more client devices are now appearing which support it, including the Samsung Galaxy S4, newer iPads and the rumoured iPhone 6.
Claiming speeds of 433Mbps up to 6.77Gbps, multi user MIMO and beamforming it would sound like we should all be rushing to implement this technology as soon as we can to solve our Wi-Fi woes. For the home user a shiny new 11ac Wi-Fi router and compatible tablet may indeed offer some benefits but if you look at the limiting performance factor in most households it is the broadband connection itself and not the Wi-Fi which throttles everything to a crawl.
For those of us deploying large scale, high density Wi-Fi, particularly at events and stadiums, the potential impact of 11ac is far more important and if not considered carefully could easily reduce performance rather than improve it. There are many enhancements and extensions within 11ac and as before with 11n it will take time for all the features to be implemented and used effectively.
One of the big changes is with MIMO or Multiple Input Multiple Output streams. MIMO is like moving from a single carriageway road to a dual carriageway or motorway – the data travels from your device to the Wi-Fi access point using multiple paths. Most business quality Wi-Fi access points have supported MIMO since 802.11n but many handheld devices have only just started to implement it. It can provide better overall speed and improve coverage especially where there are lots of obstacles. 11ac allows for up to 8 streams, whereas 11n is limited to 3, however, in reality most devices will not implement more than 3 and in fact most handheld devices will be limited to 1 or 2 because of the cost, complexity and extra power drain of adding more streams.
Those extra streams are not necessarily lost though as 11ac will eventually offer multi-user MIMO where different streams can be directed to different clients providing a much needed boost in situations such as events where the pinch point is the number of connected devices rather than absolute speed. Unfortunately version 1 of 11ac does not support multi-use MIMO so we will have to wait another year or two for that.
Beamforming is another aspect which 11ac requires, a technology which aims to optimise performance based on the direction of signals and provide a higher interference rejection. Beamforming is already supported in 11n and, when combined with adaptive antenna arrays, is very powerful in ‘noisy’ environments like event sites, however, many wireless vendors do not implement it so 11ac aims to standardise beamforming across clients and vendors, which over time will provide performance improvements.
So far it all sounds good so what is the problem? To answer that we need to look at why we have a problem today. Wi-Fi is a shared medium, a Wi-Fi ‘access point’ has to simultaneously talk to a number of client devices and split the available capacity between all the devices it is talking to. For example an 11n wireless access point (without MIMO) can at best deliver 150Mbps of capacity, if there are 100 users connected to it then each user would see a maximum speed of 1.5Mbps. This is absolute best case, real world would be far, far lower.
To add more capacity more wireless access points are used but they all need their own ‘space’ to operate in otherwise they would just interfere with each other like a room full of people shouting. To do this there are a number of standard ‘channels’ defined and each Wi-Fi access point is assigned a channel. The most common form of Wi-Fi today runs at 2.4GHz which has 14 channels but these channels overlap and not all of them are usable in all countries, in fact there are really only 3 usable channels when it comes to designs for large scale deployments. On top of this 2.4 GHz Wi-Fi has to contend with Bluetooth devices, baby monitors, microwave ovens and a whole host of other things which also use the same frequency range!
At home where there are likely to be only a few devices connecting to the Wi-Fi network these issues are not generally a great concern but on an event site where hundreds, or now more typically thousands of users have to be connected simultaneously the combination of the lack of capacity and interference creates a huge problem.
All is not lost though as there is a second Wi-Fi frequency range at 5GHz which offers 23 non-overlapping channels (although that is before you factor in indoor, outdoor, DFS and country restrictions) and much lower interference. Today most of the wireless backbone infrastructure on event sites uses 5GHz – this includes normal data transmission, CCTV cameras and other wireless devices such as video senders. There are enough channels to do this successfully provided it is all managed carefully.
Until recently most client devices did not support 5GHz but now many do meaning that client access can also be provided at 5GHz avoiding the problems of 2.4GHz. The downside of this though is that 5GHz is no longer the quiet frequency it used to be with many domestic Wi-Fi routers supporting it and permanent wireless links using it, all of which increases interference and limits available free channels. 11ac however could make things far worse.
Whereas 802.11n was a standard for both 2.4GHz and 5GHz, 802.11ac is a 5GHz only standard which means we will see an acceleration in the adoption of 5GHz in all devices. This in itself is not a bad thing but it will change the dynamics of Wi-Fi deployments with more and more focus on 5GHz client access leading to less room for 5GHz backhaul. The likely result is that backhaul will have to move to licenced frequencies or higher unlicensed frequencies such as 24, 60 or 80GHz but there are cost and implementation considerations.
The second problem is that 11ac focusses on delivering more speed but one of the ways it does this is by using a wider channel in which to send data and this is implemented by in effect ‘bonding’ channels reducing the number of independent channels available. 11n can already bond two channels but 11ac can bond four which could reduce the available channels by 75% leading to interference issues.
All of these factors are configurable and manageable and the design for a large event site will be considerably different to say an office environment but for everything to work in harmony there will need to be an even greater focus on ‘spectrum management’ ensuring that all parties using wireless equipment do so in a controlled and agreed manner. Without this structure and control the user experience will deteriorate rather than improve. 11ac can bring benefits, albeit without the headline speed claims, but there are greater risks in terms of poor design.
We will be starting to deploy some 11ac access points in a controlled manner over the coming months, working closely with vendors to optimise designs for the challenging needs of event sites.