If you are used to conference panel sessions being boring festivals of agreement, you should have seen this one. Speakers from Sky, CableLabs and (to a lesser extent) Cisco, tore into Qualcomm’s Neville Meijers at the WiFi Global Congress for having the temerity to stand up at a WiFi conference and propose LTE-U/ LAA and WiFi aggregation as means to increase spectrum utilisation and efficiency.
Meijers presented (or re-presented depending on how familiar you are with the subject matter) Qualcomm’s work on methods of enabling coexistence of LTE-U/LAA and WiFi in 5GHz spectrum bands.
His gist: Qualcomm has done a great deal of research into LTE-U and LAA, and is pretty convinced that either duty cycle (CSAT) or Listen Before Talk co-existence mechanisms will work OK.
Just as a quick primer: LTE-U allows LTE cells (usually small or metro cells) to combine unlicensed spectrum with licensed spectrum. LAA is the 3GPP standardised version of LTE-U that is expected to be confirmed in early 2016. To do this the cellular base station listens to which channels within the unlicensed spectrum are sparsely used or unused, and then transmits in those channels.
The trouble is that the methods being proposed to do that in a “fair” way, and without interfering, do not convince everyone. Many within the WiFi community resent the “greed” of LTE within their spectrum. Cellular LTE-U proponents point out it is not “WiFi” spectrum, it is unlicensed spectrum and anyone can use it as long as they do so nicely.
So… back to the panel session, where Qualcomm’s Meijers met with resistance from Vikas Sarawat of CableLabs, Mark Grayson of Cisco and BSkyB’s Sami Susiaho.
Sarawat was first to fire shots: “Co-existence with fairness is a requirement on day one. Duty cycle based approaches are not good enough, will not listen and will transmit when they think it is their turn. And it’s not simply about throughput: if LTE does not listen to transmissions it will create other side effects like delay and will affect apps like video streaming and VoWiF. So co-existence is very, very, very critical.”
Meijers: VoWiFI and streaming have been analysed in LTE-U and because of the duty based cycle frequency we are able to support VoWiFi and video streaming – there is no dropped calls or performance. When it comes to venues and enterprise – the environment is very well planned and adding LTE in means it becomes part of that planning exercise. So you mitigate against any interference by planning out which bands WiFi will be using and which 20MHz channels LTE-U could use – and have a very nice symbiotic situation as a result of true planning.
Sarawat: But let’s say in the lab [tests] there’s a duty cycle of 5% in the deployment. What if an operator changes that to 80% – what’s the guarantee [against that?] – there’s none in LTE-U.
Meijers: Vendors are used to working within 3GPP in a certified environment. Vendors will make sure co-existence is fairly handled and it’s not like knobs that can be turned. it’s not in their or in operators’ interest to be treading on their own toes.
Sarawat: If it can be in 3GPP and proven it’s working – that makes sense. Doing LTE-U in a proprietary manner is not in the favour of people who have multi-billion businesses running in unlicensed bands.
Cisco’s Mark Grayson put forward a common criticism – that coexistence tests struggle to simulate real world conditions, or to anticipate what future networks will look like.
Grayson: “People look at CSAT and see that no standard algorithms exist. And you could identify testing cases but not represent where we are in 10 years’ time. We need a breadth of simulations within 3GPP showing how we can get coexistence. Just to say we have this body of work and we have CSAT and nothing is defined… that’s part of the cause [of the opposition].
Meijers: It’s not in Qualcomm or the industry’s interest to see a technology deployed that will damage our own multi-billion dollar business. We have huge credibility at stake by launching LTE-U, and in making sure that carrier aggregation’s foundation is to ensure fair sharing and coexistence. That does not mean that LBT is only way to coexist. You can coexist with other mechanisms… we believe we have proven that. We have an open door, so CableLabs, Ruckus, Aruba, Cisco are all welcome to come and see – and have been since this first phase of LTE-U.
Sarawat: If we see a marriage of unlicensed and licensed on these terms – marriage is a benign word. If these two get married – after the marriage these two do not have equal rights. Licensed drives unlicensed spectrum however it wants – and that’s not a marriage.
There then followed an exchange triggered by Cisco’s Grayson, who pointed out that as a company with two million licensed spectrum small cells in the market and with a million WiFi APs shipping every quarter, Cisco has a foot in both camps. Grayson said that in terms of building systems that support a venue, the direction of travel is to make access points that behave more like unlicensed, rather than the other way round. “Perhaps LAA is making it sway the other way,” he mused.
Meijers: I’m not sure why want to make a tech that is very efficient [LTE] less efficient [by being more like unlicensed)
Grayson: But you will with LTE-U – you will make LTE less efficient
Meijers: You will get all the improvements from synched MAC and be bringing the best parts of LTE into unlicensed spectrum to make it more efficient. And so long as you are using that spectrum fairly and it coexists with the existing infrastructure there should be no reason why you can’t do that. 5GHz isn’t a Wifi band it’s an unlicensed spectrum band. The fact LTE is going there and will be more efficient should be making them [the WiFi side] happier.
Susiaho: If there is coexistence then how much more efficient can it be?
Meijers: 2 x more efficient
Susiaho: I would like to see that
Meijers: You are welcome to dig under the covers in San Diego – happily. We will show our claims are not unproven and we welcome anybody.
Susiaho: As soon as the standard is written and you have something developed according to the standard then I’ll come.
And that, dear reader, was just about that, save for some sparring over the need or otherwise for standalone LTE-U. But that’s one for another day.
