Posted on by — Leave a comment

“HaLow” sets stage for multi-channel Wi-Fi

The Wi-Fi Alliance’s announcement of the low power version IEEE 802.11ah, dubbed “HaLow”, was dismissed by some analysts as being too late to make a significant impact in the fast growing Internet of Things (sector). That view is wrong and seriously discounts the power and momentum behind Wi-Fi, to the extent that HaLow has already received extensive coverage in the popular as well as technical press. It is already far closer to being a household name than other longstanding contenders as wireless protocols for IoT devices such as Zigbee and Zwave.

It is true that certification of HaLow compliant products will not begin until 2018, but with IoT surging forward on a number of fronts including the smart car, digital home and eHealth, SoC vendors such as Qualcomm are likely to bring out silicon before that. There are good reasons for expecting HaLow to succeed, some relating to its own specifications and others more to do with the overall evolution of Wi-Fi as a whole.

Another factor is the current fragmentation among existing contenders, with a number of other protocols vying alongside Zigbee and Zwave. This may seem to be a reason for not needing yet another protocol but actually means none of the existing ones have gained enough traction to repel a higher profile invader.

More to the point though HaLow has some key benefits over the others, one being its affinity to IP and Internet through being part of Wi-Fi. Zigbee has responded by collaborating with another wireless protocol developer Thread to incorporate IP connectivity. But HaLow has other advantages, including greater range and ability to operate in challenging RF environments. There is already a sense in which the others are having to play catch up even though they have been around for much longer.

It is true that Bluetooth now has its low energy version to overcome the very limited range of the main protocol, but even this is struggling to demonstrate adequate performance over larger commercial sites. The Wi-Fi Alliance claims that HaLow is highly robust and can cope with most real sites from large homes having thick walls containing metal, to concrete warehouse complexes.

 

The big picture is that Wi-Fi is looking increasingly like a multi-channel protocol operating at a range of frequencies to suit differing use cases. To date we have two variants, 2.4 GHz and 5 GHz, which tend to get used almost interchangeably, with the latter doubling up to provide capacity when the former is congested. In future though there will be four channels, still interchangeable but tending to be dedicated to different applications, combining to yield a single coherent standard that will cover all the basses and perhaps vie with LTE outdoors for connecting various embedded IoT and M2M devices.

HaLow comes in at around 900 MHz, which means it has less bandwidth but greater coverage than the higher frequency Wi-Fi bands and has been optimized to cope well with interference both from other radio sources and physical objects. Then we have the very high frequency 802.11ad or WiGig standard coming along at 60 GHz enabling theoretical bit rates of 5 Gbps or more, spearheaded by Qualcomm, Intel and Samsung. WiGig is a further trade-off between speed and coverage and it will most likely be confined to in-room distribution of decoded ultra HD video perhaps from a gateway or set top to a big screen TV or home cinema.

Then the 5 GHz version might serve premium video to other devices around the home, while 2.4 GHz delivers general Internet access. That would leave HaLow to take care of some wearables, sensors and other low power devices that need coverage but only modest bit rates. As it happens HaLow will outperform all the other contenders for capacity except Bluetooth, with which it will be on much of a par.

 

HaLow will be embraced by key vendors in the smart home and IoT arena, such as Paris based SoftAtHome, which already supports the other key wireless protocols in its software platform through its association with relevant hardware and SoC vendors. SoftAtHome can insulate broadband operators from underlying protocols so that they do not have to be dedicated followers of the wireless wars.

AirTies is another vendor with a keen interest as one of the leading providers of Wi-Fi technology for the home, already aiming to deliver the levels of coverage and availability promised by HaLow in the higher 2.4 GHz and 5 GHz bands. It does this by creating a robust mesh from multiple Access Points (APs), to make Wi-Fi work more like a wired point to point network while retaining all the flexibility of wireless.

 

All these trends are pointing towards Wi-Fi becoming a complete quad-channel wireless offering enabling operators to be one stop shops for the digital home of the future, as well as being able to address many IoT requirements outside it.

At the same time it is worth bearing in mind that the IoT and its relative M2M is a very large canvas, extending to remote outdoor locations, some of which are more far challenging for RF signals than almost any home. In any case while HaLow may well see off all-comers indoors, it will only be a contender out doors in areas close to fixed broadband networks. That is why there is so much interest in Heterogeneous Networks (HetNets) combining Wi-Fi with LTE and also why there are several other emerging wireless protocols for longer distance IoT communications.

One of these others is Long Range Wide Area Network (LoRaWAN), a low power wireless networking protocol announced in March 2015, designed for secure two way communication between low-cost battery-powered embedded devices. Like HaLow it runs at sub-GHz frequencies, but in bands reserved for scientific and industrial applications, optimized for penetrating large structures and subsurface infrastructures within a range of 2km. LoRaWAN is backed by a group including Cisco and IBM, as well as some leading Telcos like Bouygues Telecom, KPN, SingTel and Swisscom. The focus is particularly on harsh RF environments previously too challenging or expensive to connect, such as mines, underwater and mountainous terrain.

Another well backed contender is Narrowband-LTE (NB-LTE) announced in September 2015 with Nokia, Ericsson and Intel behind it, where the focus is more on long range and power efficient communications to remote embedded sensors on the ground. So it still looks like being a case of horses for courses given the huge diversity of RF environments where IoT and M2M will be deployed, with HaLow a likely winner indoors, but coexisting with others outside.

Posted on by — Leave a comment

Operators unhappy over Wi-Fi and unlicensed cellular coexistence plans

Controversy has raged for well over a year now over plans by some mobile network operators (MNOs) to extend their spectrum into unlicensed 5GHz bands currently occupied by Wi-Fi. The arguments have been both commercial and technical, centering on the rights of MNOs to compete with established Wi-Fi networks and at the same time the efficiency or fairness of mechanisms for coexistence between the two.

LTE-U enables 4G/LTE cellular services to be extended into the 5GHz unlicensed bands, which is obviously attractive for MNOs because it gives extra precious spectrum without having to pay for it while making it easier to support high bandwidth applications like premium live video streaming. But the initiative, initially proposed by Qualcomm and Ericsson, has gained some traction within the 3rd Generation Partnership Project (3GPP) primarily because many MNOs want to gain full control of heterogeneous networks combining licensed and unlicensed spectrum, so there is a major commercial force here.

MNOs have expressed frustration over Wi-Fi offload, which is necessary to avoid overload on their networks and give their subscribers the best quality experience, but means they have less control over end-to-end traffic. Not surprisingly though those Telcos with extensive Wi-Fi hot spot networks take a different line and are opposed to LTE-U. Therefore we find that operators like AT&T and BT with huge investment in Wi-Fi hotspots but smaller presence in cellular are opposed to LTE-U. On the other hand Telcos that have not bet so much on Wi-Fi but have major cellular operations now support LTE-U, including big hitters like Verizon, China Mobile, NTT DoCoMo, Deutsche Telekom and TeliaSonera.

Notably though some of the world’s biggest providers of mobile services are ambivalent about LTE-U, which some of them see as complicating rather than simplifying the drive towards heterogeneous services combining licensed and unlicensed spectrum. The view there is that Wi-Fi is best placed to occupy the unlicensed spectrum with a lot of momentum and investment behind it. The LTE-U camp counter that the technology can carry twice as much data as Wi-Fi in a given amount of 5 GHz spectrum through use of carrier aggregation via LTE-LAA. This was already defined in the LTE standards and enables multiple individual RF carrier frequencies, either in the same or different frequency bands, to be combined to provide a higher overall bit rate.

This may be true as far as it goes but is largely irrelevant for users wanting to access broadband services in their homes or public hot spots, according to the Wi-Fi community, a view shared by some MNOs as well. Birdstep, a leading Swedish based provider of smart mobile data products enabling heterogeneous services combining cellular and Wi-Fi, argues that the story is not just about the wireless domain itself but also the backhaul infrastructures behind it. Any spectral efficiency advantage offered by LTE-U would be more than cancelled out by inherent inefficiencies in the backhaul. By offering access to the world’s broadband infrastructures Wi-Fi offers greater overall scale and redundancy.

Another Wi-Fi specialist, Turkey based AirTies, contends that LTE-U is just a spectrum grabbing bid by MNOs and should be resisted. Air Ties has developed mesh and routing technologies designed to overcome the problems encountered by Wi-Fi in the real world and these are only going to get worse as unlicensed spectrum reaches even higher frequencies. The next generation of Wi-Fi based on the emerging IEEE 802.11ad standard will run in the much higher frequency band of 60 GHz, which will potentially yield a lot more capacity and performance but increase susceptibility to physical obstacles and interference. It will only work with further developments in the sort of intelligent beam forming, meshing and steering technologies that AirTies has invested in.

It is true that LTE-U proponents have worked hard to mitigate any impact of coexistence with LTE-U on Wi-Fi. In Europe and also Japan they were forced to do so anyway by regulations that required LTE-U to adhere to similar rules over fair access to spectrum as Wi-Fi. These rules insist on incorporation of LBT (Listen Before Talk) into LTE-U, a mechanism originally developed for fixed line Ethernet networks where there was a shared collision domain (it was called Carrier Sense Multiple Access or CSMA). Stakeholders that are not in favor of rapid LTE-U deployment point out that in the old Ethernet days before 10BaseT/switching, CSMA proved inefficient when there were to many devices trying to get onto the same collision domain. Total capacity could drop drastically and this issue could be reborn into the wireless world.

The European Union specified two options for LBT, one the scheme called DCF/EDCA already adopted for Wi-Fi standards and a newer scheme known as Load Base Equipment (LBE), differing in the procedure for backing off when detecting traffic in a given channel.

Naturally enough there has been an assumption in the LTE-U camp that any deployments will be safe if they do adhere to the EU’s LBE LBT standard. But this assumption has recently been challenged by CableLabs in a simulation modeling a million transmission attempts on sets of nodes following the EU LBE LBT rules. The EU LBE turned out to scale badly with increased numbers of devices, with growing numbers of collisions. This will only amplify concerns expressed by broadcasters such as Sky, as well as by some major vendors like Cisco with feet in both the Wi-Fi and LTE camps, that LTE-U poses a threat to quality of service for premium video especially.

There are no signs yet of the LTE-U camp giving up on their efforts to infiltrate the 5 GHz domain, arguing correctly that by definition unlicensed spectrum is free for all and cannot be owned by any one wireless technology. But there is a strong case for holding off from LTE-U deployments until further extensive tests and simulations have been carried out to assess the impact on capacity and QoS in real life situations.

Posted on by — 1 Comment

Google getting it right at last with Android TV

It may still be too early to be sure, but there are signs that with Android TV Google’s connected video strategy is at last starting to look joined up. There is not yet any killer blow in sight, which could only really come from the content angle, but evidence is mounting that Google now has a clearer and more focused strategy that is winning over TV and device makers as well as app developers. The key lies partly in a much mature Android based ecosystem better geared towards TV than was the case when the abortive first attempt called Google TV was launched with much fanfare in 2010.

Then Android was little more than a mobile OS optimized for smartphones and subsequently tablets, with TV supported as a cumbersome add-on that was hard to develop apps for. But the latest version 5.0 codenamed Android Lollipop, first unveiled during the Google I/O developer conference in June 2014, has been revamped for TV with a completely redesigned user interface. This is a significant enhancement based on Google’s own language called Material Design incorporating tools for easy layout of screens with responsive animations, transitions, padding, and depth effects such as lighting and shadow. It comes with new guidelines for developers that make it easier to create a consistent look and feel across the whole Android device constellation, including big smart TVs down through tablets and smartphones to diminutive smartwatch screens. To encourage app creation further, Google sent out developer units, dubbed “ADT-1”, to those that signed up for a test unit at Google I/O 2014.

These nuances were initially missed by many commentators, myself included, at the time of that conference, perhaps partly because the new Lollipop version was still shrouded in a little mystery. My initial reaction to Android TV was therefore quite negative, suggesting it sent confused messages given that Google was also promoting Chromecast and that it offered little more than already existing competitive offerings such as Apple TV, Roku and Amazon Fire TV.

The reason for being more sanguine about Android TV now is not so much that Google has raised its game. If anything it is the opposite in that the horizons have been narrowed to the confines of an operating platform for TV but crucially now aligned with Chromecast as well as with its developer community. What Google has succeeded in doing is strike a balance between encouraging innovation and yet exercising some control over the environment with an emphasis on a consistent UI across all devices, which is something its competitors have not quite matched yet. We have already seen the fruits of this approach through a few OEMs such as Razer, which has announced Forge TV, a set top for Android TV that throws in some of its gaming streaming.

Casting is now at the center of Android TV and pivotal to delivery of content, with the various new boxes, including Google’s own Nexus, Player, being the first dedicated hardware units to support it. This does though pinpoint the challenge of persuading consumers to pay the extra for the full Android TV experience when they can get Netflix and all the basic content they want from Chromecast. Effectively then Chromecast is the entry level version of Android TV with the full monte running on set tops as well as smart TVs, including models announced by Sony, Philips and Sharp at CES 2015.

The killer feature though would be premium live content and all that can be said at this stage is that Google has prepared the ground with its dummy app called ‘Live Channels for Android TV’.  It remains to be seen what will be on it and how far this goes beyond the content currently available either via Chromecast or YouTube. But at least Google is much better placed to strike a major blow in the intensifying connected TV wars.

Posted on by — Leave a comment

Operators plan massive carrier grade Wi-Fi expansion

Huge pent up appetite for carrier grade Wi-Fi has been confirmed by recent research commissioned by customer experience specialist Amdocs, suggesting that both cable operators and MNOs (Mobile Network Operators) will deploy this at massive scale over the next three years. The research, conducted by Real Wireless and Rethink Technology Research, identifies how operators in both these camps are in turn responding to rapidly growing dissatisfaction among their customers with the Wi-Fi performance and reliability they are getting in public hot spots in particular. The fault does not really lie with Wi-Fi itself, which has actually improved in leaps and bounds, but instead the failure to keep up with escalating expectations. People now expect levels of availability for Internet access that used to be confined to enterprise data networks and Wi-Fi has come into the firing line as the new de facto “last mile” of the broadband access infrastructure.

So as Amdocs pointed out, service providers are seeing that “best-effort” Wi-Fi is becoming less profitable and a guaranteed higher quality of experience (QoE) is needed for emerging revenue generating services such as TV everywhere and online gaming. Yet as we all know Wi-Fi QoS at public places like hotels and trains is all too often poor and inconsistent, too susceptible to data traffic congestion as well as varying spectral conditions.

Carrier Wi-Fi implies guaranteed QoS for specific services such as TV, which in turn depends on traffic management techniques in order to meet varying requirements for bandwidth and latency by giving some IP packets priority whole holding up other packets associated with less urgent applications like email. Above all carrier Wi-Fi requires strong tools for network planning and management to ensure that QoS can be maintained even at peak times. In the Amdocs survey, two thirds of respondents identified lack of such strong tools as one of the top three risk factors that might deter or delay investment in carrier-grade Wi-Fi.

Fortunately such tools are now available from a clutch of vendors that now specialize in carrier Wi-Fi after cutting their teeth in offload to broadband via Wi-Fi from cellular networks. One of them, Aptilo, now emphasizes the importance of integrating Wi-Fi at the service management level with existing backend OSS/BSS operational systems as a foundation for policy enforcement and new revenue generation. Another, Birdstep, has been focusing increasingly on the bigger picture of heterogeneous networks (HetNets) that combine Wi-Fi with cellular with the catch line of “Experience Continuity” to describe the goal of delivering optimum QoS to users wherever they are and whatever device they have.

Of the two operator categories covered by Amdocs in its research, HetNets are of greatest interest to MNOs, but carrier grade Wi-Fi itself is a major goal for many cable operators seeking to give their subscribers access to premium TV content on the road and underpin their quad play offers. The interesting aspect of the research is the suggestion that operators will be clutching carrier Wi-Fi technology almost as soon as it comes out. As a result the prediction is for penetration of carrier-grade Wi-Fi hotspots to increase from 14% at the end of 2014 to 72% by 2018.

This will not be a case of technology leaking gradually out to the market as it comes along, but being pulled hard by consumer demand. Just as high speed broadband Internet access has come to be taken for granted, carrier Wi-Fi will quickly follow.

Posted on by — 1 Comment

Google still playing catch up with Android TV

In this first blog of our Android TV series, Philip Hunter looks at some of the reasons why Google doesn't seem to have been third-time lucky with its new Foray into the sitting room.

Android TV, Google’s heavily leaked third effort to crack the living room, arrived as expected at its annual Developer conference at the end of June 2014, but once again failed to ignite the field. Google is still playing catchup in its attempts to conquer TV, with the only consolation being that its rivals such as Apple and Microsoft are also floundering in their attempts to stamp their authority on the big screen. Apple TV after all is seven years old now and can still at best be counted as only a moderate success, having failed to establish any sort of market dominance even close to that achieved by iOS for both tablets and smartphones. One reason is that Apple’s iron grip on its ecosystem has proved a handicap in keeping the device isolated and discouraging developers.

Google was determined with Android TV to avoid that fate and continue with its open approach to maximize third party app development with a variety of incentives. But this immediately begs the question for developers of whether Android TV is worth the trouble, unlike the versions for smartphones and tablets where it was always clear there would be a big market. In the case of TV it is still not obvious at all that Android will succeed, given that the market for streaming platforms is already very crowded, with Roku, Amazon Fire TV, Xbox One as well as Apple TV among leading established contenders. There is also Google’s own Chromecast HDMI dongle, representing its second attempt at TV after the abject failure of Google TV. Chromecast has been a reasonable success because of its low price tag of $35 and flexibility, encouraging users to try it alongside their existing pay TV package if they have one, rather than as their primary source of content. But it means Google is now sending mixed and confused messages to both consumers and developers. With Chromecast, Google had given the impression that developers could stop building apps and instead create webpages optimized for the TV screen that would receive commands from an Android smartphone. If Chromecast stays around as it looks like it will, developers now face having to build and support two interfaces to cover the Google TV universe, one for Chromecast and one for Android TV.

The underlying problem though may be with Google’s new strategy of shoehorning Android for TV rather than creating a new operating system. Google is obviously trying to make Android into a ubiquitous operating system with variants for all device platforms, as was evident at its recent so called I/O developers’ conference. There Google unveiled Android Wear for wearables like wristwatch computers, Android Auto for car dashboards and Android One for a new brand of affordable smartphones prices at under $150, as well as Android TV. There is every reason to expect that Android Wear and Android One will be great successes as they are still very much in the heartland of mobile handsets, but with Auto and to an even greater extent TV Google is stretching the envelope of the operating system a long way. History tells us that attempts to create an operating system of everything are doomed to failure, as Microsoft seems to be finding with Windows Phone. Apple had the sense to create a radically new operating system for mobile devices in iOS rather than attempting to adapt the MacOS from its desk top perch. It reaped the rewards with the iPhone and then the iPad.

This leads to the other problem, which is that with Android TV Google is still cast in the role of follower rather than leader, which is not how it grew up to become the world’s most valuable brand. In search Google originally rose to dominance over rivals like Yahoo, AltaVista and Microsoft because it had superior technology and was quickly able to assume a leadership position that came to be reinforced by its human and financial resources. Of course it has been able to reallocate those resources to TV, but without so far being able to conjure up any killer technology.

Android TV actually seems rather similar to Apple TV for the GUI and is modelled on Amazon’s Fire TV in its support for voice input for searching video content. Indeed Google has instructed its developers to avoid need for any text input at all if possible and to rely largely on voice. Google has also stripped out well known features of Android on smartphones, such as support for VoIP, cameras, touchscreens and NFC (Near Field Communication), which are all deemed superfluous for a streaming set top box like device. This is well and good, since it avoids an Android TV box being an over bloated version of a smartphone or tablet. But this may also expose the limitations of the platform, especially as Google is indicating that the operating system will really come of age with its next generation called Android L, which was also previewed at that I/O conference and scheduled for launch towards the end of the year. Android L is a radical rewrite, with improved animation and audio, as well as 3D and contextual awareness, which will all feed into the TV version. Developers may therefore decide to wait until this next generation has arrived before committing to Android TV. Android L will also include many features specific to mobile handsets, including a new battery saving mode, which are mostly irrelevant for the TV version. In this sense Android L will compound rather than solve the problem of becoming too bloated. It may be just a matter of time before the TV version of Android becomes divorced altogether from the mainstream of the operating system, but meanwhile it will most likely fail again to put Google on the podium for OTT TV.

Posted on by — Leave a comment

Wi-Fi offload can help mobile operators deliver network neutrality

Network neutrality has come back to the boil in 2014 following US carrier Verizon’s famous Federal court victory in January over the regulator FCC (Federal Communications Commission), allowing it to differentiate between services delivered to its broadband customers. This was followed in April by the European Union approving strict network neutrality with the message it would take a much tougher stance than the FCC in upholding the rules. Naturally this was widely interpreted as setting Europe apart from the US, but the reality is that both are taking a more nuanced approach than in the past. Even the EU proposals allow for provision of specialized services, providing they do not intrude into network capacity set aside for the general Internet. The tones may be different but the broader implication both in the US and Europe is that network neutrality can never be fully attained through legislation, any more than true equality of wealth can be achieved via measures such as progressive taxation – both are aspirations or focal points.

For mobile operators the aspiration of network neutrality has assumed a logistical and economic dimension with the great proliferation of data hitting their infrastructures. Many have opposed strict net neutrality for the simple reason that their core and backhaul networks have limited capacity and would be unable to cope without traffic engineering and the ability to differentiate between different service or application types.

But now Wi-Fi offload has entered to change the game, giving operators an option for relieving their overstretched backhaul networks and for that matter their radio access capacity as well, by taking advantage of broadband infrastructures. It was at the Mobile World Congress in 2013 that offload first seemed to have risen right up the agenda for mobile operators. Generally, particularly before deployment of 4G/LTE, broadband networks had greater capacity and crucially lower costs than the fixed backhaul networks serving radio base stations. For this reason those major Telcos with their own network of hot spots have been leading the march towards Wi-Fi offload. In the US AT&T has built large Wi-Fi hot zones in mostly urban areas with high levels of cellular traffic, specifically for offload to help relieve congestion on its core mobile network.

The implications of such offloading for network neutrality have not attracted much attention, but are likely to be profound nonetheless. The fundamental point is that by freeing up capacity on the mobile network, offloading can help mobile operators meet their net neutrality obligations as laid down by regulators in the region concerned, while still having scope to offer specialized services. An operator could say offer an OTT video service such as Netflix with guaranteed QoS over the cellular network, resorting to Wi-Fi offload for third party OTT services such as YouTube. Alternatively Wi-Fi could be used for specialized services, especially by operators like AT&T that have their own overlapping hot spots and cellular networks on a large scale.

We are already seeing this happen, with Sprint in the US now offering calling and messaging over Wi-Fi when within range of suitable hot spots. Sprint incidentally was one of the first major carriers in the world to make serious use of Wi-Fi offloading for data.

We are going to see plenty more such offerings over the coming years. It will be interesting to see the extent to which operators will align Wi-Fi and cellular within heterogeneous service offerings effectively to escape the shackles of net neutrality while obeying the basic rules as stipulated by regulators.