AirTies: boom in OTT, faster speeds and dumb mobile devices brewing up a “perfect Wi-Fi storm” for service-providers

By Barry Flynn, Contributing Editor

A boom in OTT viewing and ever-faster broadband speeds, combined with the dumb behaviour of modern mobile devices connecting to domestic networks, is creating a ‘perfect Wi-Fi storm’ in the home, according to Bülent Çelebi, Chairman and Co-Founder of AirTies Wireless Networks.

For some broadband service providers, this means that up to 70% of call-centre traffic is Wi-Fi related, claimed Çelebi, “and they get more calls the higher the broadband speed they’re offering. […] Everyone is about to run into a brick wall.”

In a presentation to delegates on the second day of the Connected TV Summit in London, Çelebi argued that the crisis was caused by consumers seeing only “a fraction of what they pay for,”and he underpinned his claim with real-world Wi-Fi test data from territories as diverse as Istanbul, Hong Kong, Switzerland and the USA.

In an Istanbul flat with 20 Mbit/s VDSL broadband access, streaming video from an iPhone in the same room as the Wi-Fi router, “you’re getting a performance that is higher than the broadband speed that you are paying for. But if you go a few walls over, it drops down to zero. If it’s not zero, it’s 1 Mbit/s. Anything below 5 Mbit/s on Wi-Fi is very, very poor performance.”

This was not a special case, Çelebi stressed. In a Hong Kong apartment with a 1 GBit/s fibre connection, using the best available routers and clients*, where the receiving device was located in the same room as the router, it was able to download at 177Mbit/s.

But “if you go down the hall, it drops to 108 Mbit/s, if you go down a little bit further it drops to 0.8 Mbit/s. There’s gigabit fibre coming into the place!”

Çelebi supplied other test-cases showing similar results, one of them involving the worst-case scenario of a three-storey Swiss home build with concrete walls and floors.

A number of common phenomena were at work across all these test-cases, he said, caused by lack of intelligence both at the device and network level.

At the network level, broadband service-providers were unable to monitor and diagnose issues within the in-home Wi-Fi networks constructed by their clients, who stream video using mobile devices they do not control, and – when they encounter bad reception – typically install new access points or repeaters they do not own.

Meanwhile, at the device level, there are a number of different behaviours that conspire together to aggravate the situation, chief amongst which is the fact that responsibility for deciding which broadband access point to connect to at which frequency is down to the laptop, tablet or mobile itself, which “has no knowledge of what the network looks like. The only thing it’s typically relying upon is the signal strength,” claimed Çelebi.

A modern dual-band repeater is capable of operating in the 5 GHz Wi-Fi band as well as the 2.4 GHz one, which in theory allows whichever band is the least congested to be used. But in real-world domestic situations, a repeater’s 2.4GHz band will generally offer better signal strength than 5 GHz, so most Wi-Fi clients in consumer devices will tune in to the 2.4 GHz band even if capable of accessing 5 GHz.

Unfortunately, many legacy devices still have low-end Wi-Fi chipsets which are not 5 GHz-capable – so when attempting to re-connect to a local repeater, can only access the 2.4 GHz band. “So [the 2.4 GHz band is] already crowded and then it gets even more crowded,” commented Çelebi. From a network optimisation perspective, the ‘right’ decision might well have been to connect to the service-provider’s gateway directly, despite the signal strength being weaker, but the lack of network intelligence available to the consumer’s device does not allow this.

“The other thing is that mobile clients are typically ‘sticky’,” points out Çelebi. â€œOnce they make a connection to a particular access point, they tend to stick with that access point even though the signal strength has dropped significantly or there is a better device that is closer to it. So it makes a connection and then as the consumer starts moving throughout the office or throughout the home it stays connected to that [access point] unless there’s something bad that actually triggers it to do a rescan – which would typically be when it loses a connection – and then it restarts.”

What a ‘sticky’ device does is to “bring down the total capacity of the Wi-Fi network,” said Çelebi. As the device moves away from the access point it was initially connected to and moves behind walls, the original access point devotes more and more capacity to serving that particular device’s needs – when in reality, from a network management point of view, it should simply have re-connected to a different access point, if available, or switched to a different channel.

Çelebi dubbed this “the bad apple phenomenon.” The effect on a second user who, say, is located close to the same access point and successfully streaming video, is that his available capacity will collapse, and he may either be switched to a lower-quality stream, experience re-buffering, or lose his picture entirely.

Airties’ solution to these issues is “to set up mesh links between Smart Access Points, and then route on a best-path basis.”

These access points are situated at certain network ‘nodes’, which can be repeaters or stable devices such as set-top boxes, and connected together through high-speed channels called ‘mesh links’ – allowing them not just to communicate with each other but to pass network traffic between themselves.

These effectively re-insert intelligence into the home network, and allow the service provider to monitor and diagnose connectivity issues as they arise.

But they also significantly improve the user experience of streamed video in the home, said Ã‡elebi. For instance, combined with a technique called ‘client steering’, such an approach means “no bad apples” – mobile devices are simply told which access point to connect with to ensure they are performing at their maximum capability. Meanwhile, traffic is routed around the network in a manner analogous to the way in which IP packets are routed around the Internet: across a ‘best path’ instead of a single one, which is dynamically configured and re-configured depending on the number of active devices, the traffic profile, and the Quality of Service required by the customer’s agreement with the service provider.

Çelebi demonstrated ‘before and after’ results for the Airties mesh-plus-routing solution using test data from the problematic all-concrete Swiss home mentioned above, where “the only coverage that this person is able to get is essentially where the residential gateway comes into the house.”

With the Airties solution installed, however, the owner was able simultaneously to stream three HD streams around the house (one to the basement, one to the 1^st floor and one to the second floor). Çelebi showed a snapshot of how the home network coped with the load: “He’s got 90 Mbit/s worth of traffic going throughout this really concrete building and he’s only consuming 50% of the airtime capacity. So this [snapshot] basically says that in this network he should be able to get 180 to 200 Mbit/s on a sustained basis.”

In a separate example from the USA, where Airties’ mesh solution was combined with a ‘client steering’ approach in a home with 27 separate mobile devices, Çelebi showed data which demonstrated that, despite the potential for a large number of ‘bad apples’, “all the devices are operating at 5 GHz if capable of 5 GHz, and […] in general you see that [the network is] operating at 30 to 40% of the total capacity. There are a few spikes there where it goes up to 70% or so – but in general the network is being managed and this guy has a very good user experience.”

* A 3×3 11ac wireless router and MacBook Air client using beam-shaping at both the access point and on the receiving side