Whether ’tis Nobler in the Wi-Fi to suffer
The Slings and Arrows of poor performance,
Or to take Arms against a Sea of troubles
Wi-Fi has been around since 1997 – over 15 years since the original standard was released to the market. In it’s original form Wi-Fi was pretty slow compared to today – the first version only supported data rates of 1 and 2 Mbps. In 1999, Wi-Fi got a big speed boost with the introduction of 802.11b which supported higher 5.5 and 11 Mbps data rates, still slow compared to today but at the time it was considered a pretty good speed increase. In fact it was considered so good that the first Apple Macs with Wi-Fi were launched in July 1999.
Since then, Wi-Fi has got faster and faster. We had the launch of 802.11g which increased the maximum speed to 54 Mbps, then we had 802.11n further increased the speed to a maximum of 600 Mbps and today we have the new 802.11ac which has a data rate if around 1 Gbps in products shipped today. With second generation 802.11ac we will see the maximum data rate increase to just under 7 Gbps.
But there is a problem with Wi-Fi. We have these maximum speeds, but networks never seem to get anywhere near them in real life. When we quote a data rate, we are talking about the maximum speed possible under ideal conditions in the air, the airput. We do not talk about legacy or payload, we just talk about speed. Payload is the data that we can transmit. In an ideal network, the maximum data payload is about 70% of the airput. So when we say we can achieve 1Gbps with 11ac, the maximum payload in ideal conditions would be 700 Mbps. Still not bad and most people would not be concerned about this.
But the ideal network is very hard to achieve. One of the problems that we have in Wi-Fi is that the network has to be able to work with legacy equipment – including the origional 802.11b standards. Since 802.11g, everything in the 2.4GHz band has been OFDM radios. The original standards used DSSS and CCK radios. These old radios are slow and the OFDM radios have to be able to work with them. When 802.11g Wi-Fi was first introduced we would get about 24 Mbps data throughput with an airput of 54 Mbps; that was OK. When we had a mixed network of 802.11g and older 802.11b devices, the data throughput fell to about 12 Mbps – over 50% reduction in data throughput – not OK. The Wi-Fi industry did things to sort this out, but we still have to support the older 802.11b devices.
2.4 GHz Wi-Fi channels for 802.11b,g WLAN (via Wikipedia)
The other problem we have is that all access points – even access points using the latest technology – which operate in the 2.4 GHz band have to support the older 802.11b devices. Access points have to send out their management frames (the stuff that is used to make sure that the network behaves itself) at a data rate that everybodycan understand – which is 1 or 2 Mbps. All Wi-Fi networks operating in 2.4 GHz have lots of data running at very low data rates. Effectively the legacy requirements are bringing the data throughput down to very low levels. That is why in very busy locations your Wi-Fi is very slow, there are lots of very slow management frames.
Currently the IEEE 802.11 Working Group – the group that creates the basic Wi-Fi standards – is doing an update to the standards. One of the discussion items is whether we should still support the legacy 802.11b data rates. Microsoft are just about to stop supporting Windows XP, should the Wi-Fi industry stop supporting the old slow data rates? The obvious answer is yes. The people saying yes are those involved in high speed, high performance Wi-Fi (laptops, Access Points, phones, tablets, etc). They see the removal of the legacy 802.11b data rates as good as they will increase the real life data throughput we see with a Wi-Fi network. Also, removing support for these from the Wi-Fi chips should reduce the power consumption of the Wi-Fi devices – helping to increase battery life.
Even though removing these old data rates will speed up the performance of the Wi-Fi network, there is a significant part of the Wi-Fi industry that is saying that maybe we should still support them. The people saying this are those involved in the IoT industry. Many IoT devices only need to support low data rates – a sensor for example only needs to send a small amount of data at a very low data rate. One of the advantages of the 11b data rates is that you get the longest range. Using a typical Wi-Fi device, you can get around 300m with the old 1 Mbps data rate. 802.11g 54Mbps will typically be around 25m maximum range.
So the Wi-Fi industry is in a bit of a quandary. Should we support 802.11b – to 802.11b or not to 802.11b, that is the question. There are good reasons for removing 802.11b and there are good reasons for keeping 802.11b for the growing IoT industry. What will the Wi-Fi industry do? The 802.11 Working Group is hoping to come to a decision by July 2014. There is a case for making 802.11b optional – where we can have the best of both worlds! But it could be very difficult to come up with a solution where 802.11b can work when it is deployed without having problems from the newer OFDM based networks. Hopefully we can come to a decision soon.