The WiFi versus 3GPP topic is one that never goes away. Even when a cooperation between the two ecosystems is announced – like this week’s far-reaching joint white paper issued by the Wireless Broadband Alliance (WBA) and Next Generation Mobile Network (NGMN) Alliance – it only serves to highlight the long line of wasted opportunities to bring the two main wireless platforms closer together.
The debate has gained new life from the similar timings of the newest round of standards. Will WiFi 6 (based on the IEEE’s 802.11ax specifications) make 5G New Radio redundant? Or will 5G cover so many bases that it will squeeze WiFi into a small corner, especially when an unlicensed spectrum version appears?
Of course, neither of these extremes will materialize. WiFi is entrenched in homes, non-smartphone devices and private enterprise networks, and its vast ecosystem will ensure it remains the default technology for WLAN, indoor environments where high mobility is not required. Cellular scores outdoors, in very mobile situations, and where the control and protections of licensed spectrum are needed.
The real battlegrounds will be in the use cases and markets where both standards are exploring new ground. Ultra-dense networks, indoors and outdoors; low latency connections; massive numbers of IoT devices; even fixed wireless access (FWA); integration with edge clouds. In some of these areas, one or other has a headstart – 5G in low power IoT, thanks to GSM, for instance; or WiFi in FWA.
Both sets of standards have been extended significantly to support the new network behaviors and new spectrum bands. WiFi goes well beyond the WLAN these days, with variants like HaLow (802.11ah), which targets low power WANs, or WiGig (802.11ad) for very high capacity, short range connections in millimeter wave spectrum. 5G New Radio (NR) is also reaching up into mmWave bands, and the upcoming Release 16 will set a new bar in terms of low latency and high availability. WiFi 6 claims many of the same capabilities. It will be finalized in a similar timeframe to Release 16, but is already appearing in pre-standard silicon.
A new twist in the tale is the renewed interest in private networks. To some extent, WiFi was responsible for their decline, since it provided a solution for enterprises which wanted good indoor wireless connectivity and full control over access and performance, rather than trusting everything to an MNO. When WiFi is largely in-building, and under the management of an IT system, it can be secured so that the disadvantages of unlicensed spectrum – interference and hacking – are negated, while its cost advantage is fully realized.
With 5G, there is a renewed push to get self-contained cellular networks into enterprises, often linked to edge compute and cloud packet core deployments, and looking ahead to the provision of network-on-demand and network slices to corporations.
The enterprise, in-building space, and particularly the enterprise-managed network, will therefore be the chief area where WiFi 6 and 5G NR will need to fight it out, or reach accommodation. The home belongs to WiFi, the outdoors belongs to cellular – but industrial and enterprise markets are where most of the new revenue will be found in the 2020s. That fact increases the risk of continuing stand-off, as the powerful industries behind each standard – which remain surprisingly short on overlap – try to gain pole position, and the bulk of the revenues.
Yet, for those enterprises, and users in general, a big dose of convergence would deliver far greater benefits from next generation wireless. WiFi is incumbent and WiFi 6 will increase speed, reliability, responsiveness, automation and so on. But 5G NR will be superior in some other areas that are becoming increasingly interesting to some large vertical industries, especially for IoT applications from the early 2020s onwards. Some of this relates to the radio standards themselves, including low latency and high security, but a lot is about the wider 5G platform, with the 5G core network (5GCN) supporting full virtualisation, network slicing, automated operations and many other capabilities that will improve performance and economics.
Getting the best of both worlds will be important to many industries, and the NGMN and WBA are using their new paper to reiterate the call for greater RAN convergence. This has been heard many times before, but attempts to include the IEEE in the 5G standards process came to nothing (see inset), and though cooperations between the two sides have yielded some results in terms of interworking specifications, these are minor compared to what could have been achieved with more fundamental collaboration.
The WBA was set up to represent WiFi service providers, and has extended its remit in recent years to other unlicensed spectrum technologies. It has been bold in its claims for a central role for WiFi in the broader 5G platform; and in arguing that WiFi will deliver many of the advances that are being discussed around 5G NR. Its official line is that “WiFi will be a core component of 5G communication technologies and WiFi will have a significant impact on the definition of 5G, which will seamlessly integrate with future IMT-2020(5G) standards.”
However, as a solo voice, the arguments have always sounded somewhat defensive, especially when the WBA has been set against its mighty equivalent in the cellular world, the GSMA. By forming a united front with the NGMN – a major powerhouse in the 3GPP arena, and heavily driven by large MNOs – the WBA has greatly amplified its voice and boosted its credibility.
In concrete terms, the two bodies have published the first results of a collaboration which aims to drive greater convergence of different RAN and core technologies. The report identifies various use cases where the optimal solution woud be an integrated 5G/WiFi approach, as well as highlighting the challenges to achieve that. If there is to be any hope of real future convergence, the key hurdles to leap are, the report argues:
- tighter integration of WiFi access within 5G networks
- network manageability and policy control,
- the enablement of WiFi-only devices in a converged environment.
The organizations stated: “As society increasingly depends on fast reliable data connectivity, NGMN and WBA believe an important capability for the industry is the convergence at a network level between 5G and WiFi, so that the unique and complementary capabilities of both RANs can be leveraged to provide seamless network services. Bearing in mind that a significant amount of data traffic from smartphones use a WiFi access, this will lead to a better user experience and create new business opportunities for both WiFi and cellular providers.”
The NGMN Alliance’s CEO, Peter Meissner, said in a statement: “Convergence of 5G and WiFi can potentially bring major benefits to cellular operators, enterprise WiFi and public WiFi solution providers, giving access to 5G and enterprise services from both access networks.”
But he raised one of the key issues which has always kept MNOs at a wary distance from WiFi. They may have welcomed some measure of interworking, so they could offload excessive or low value traffic from their precious 4G RANs, but they knew that, once a user moved to WiFi, they could no longer be monetized or their behavior tracked and analyzed. Meissner said that, especially in the enterprise and public WiFi areas, “there is a demand from cellular operators for a standardized solution for improved visibility and control in the configuration and management of WiFi access networks”.
This will be the price for WiFi to be embraced fully within the 3GPP environment, rather than treated as a hard-working but mistrusted poor relation.
MNOs will need WiFi offload less in the 5G era than in 4G, at least once they get to boost capacity in mmWave bands. And they can leverage unlicensed spectrum using the 5 GHz implementations of LTE, the 3.5 GHz MulteFire technology, and in future 5G-Unlicensed (though not until Release 16 is finalized). So it is important, from the WBA’s point of view, to cancel out some of the MNOs’ motives to shift away from WiFi.
One of the advantages of using cellular technologies in ‘WiFi bands’ is to retain full control and visibility of traffic and users. If that can be offered in a mixed WiFi/cellular system, there will be less requirement for the 3GPP unlicensed solutions. Licensed 5G will handle the traffic where the advantages of cellular are needed; mobile broadband capacity can be split seamlessly between licensed cellular and unlicensed WiFi.
Full interworking will clearly be essential. It has never really materialized, despite several efforts and some limited progress. The WBA’s Hotspot 2.0 supported SIM-based sign-in to WiFi and some degree of seamlessness with cellular. The 3GPP has also developed hand-off and interworking specifications such as TWAG (Trusted Wireless Access Gateway) and non-trusted access (see inset). Individual vendors and service providers, from Ruckus Wireless to Google, have created their own solutions for intelligent hand-off between WiFi and cellular, while some MVNOs have adopted WiFi-first mechanisms in which a handset defaults to WiFi but automatically moves to cellular when there is no hotspot to hand.
But at network and core level, the two systems remain very separate, managed and controlled differently. In particular, the NGMN/WBA study identifies four challenges which must be met to enable fully converged operation, with the associated cost, performance and revenue benefits. These are:
- Enabling WiFi-only devices to connect to the 5G core
- Further study to ensure tight integration between 5G and WiFi networks
- an interface to enable a certain level of network manageability and policy control between 5G core and WiFi networks
- the ability for a client to route traffic over one or more access types in an intelligent way, making optimal use of the available connectivity.
As a next step, the WBA and NGMN have initiated further study on these challenges in order to identify potential solutions, which will be enshrined in a full strategy for future Converged RAN deployment.
Tiago Rodrigues, general manager of WBA, said: “WiFi 6 introduces new capabilities for carriers, cities and enterprises to cost-effectively provide additional coverage and capacity, mainly indoors, to address the 5G use case requirements. Now it’s time to fully capitalize on these capabilities by delivering a clear strategic path for converged RAN deployments. This is a priority. We will continue to work closely with NGMN and its members to review, develop and test potential solutions.”
The advent of 5G would seem to be a good time to revisit the issue, especially as, within a few years, 5G roll-outs will start to go hand-in-hand with network virtualization programs within many operators. A virtualized RAN, core and management and orchestration platform may be a good opportunity to incorporate other radios in a flexible way with common control – but only if the WiFi community has been able to coordinate its own efforts in the crucial area of network and services management.
Though presented as a project of equals, there is a sense that this is more about giving WiFi a seat at the MNOs’ top table, than about the reverse – opening 5G technologies up to WiFi-oriented service providers. There are prospects for that to happen, but they are being driven by different entities, such as the USA’s cablecos and their R&D arm, CableLabs. These are leading the way in leveraging LTE, and later 5G, in shared spectrum, to support self-contained home and enterprise networks independent of the MNO. They are working on ways to achieve common management and orchestration of cellular and WiFi assets, and some of CableLabs’ work in these areas is likely to find its way into 3GPP and/or open projects like Telecoms Infra Project (TIP).
How far 5G becomes a tool for non-spectrum owners in future will depend on more than interworking with WiFi. It will be driven by policy decisions about how much spectrum, including in the mmWave bands, should be unlicensed or shared; plus commercial developments like the potential rise of industrially focused neutral host platforms. But deeper integration between the two key wireless technologies would be helpful to cablecos, private network operators and webscalers, if they do indeed get their hands on more 5G. It is to be hoped that the strong initiative which NGMN and WBA have made can be expanded to address the whole wireless service provider base in future, and that – for the first time – the two sectors go beyond grand words and white papers, and drive meaningful convergence into the standards and the ecosystems.
To download the RAN Convergence Interim White Paper, click here: https://www.wballiance.com/resource/ran-convergence/
The WBA’s’ 5G White Paper, outlining its view of the role of WiFi in 5G, can be downloaded here: 5G White Paper.”
Progress to date on interworking:
Some technical aspects of 5G/WiFi interworking have already been addressed and these should be built upon. For example:
3GPP has specifications for tight integration of 3GPP with non-3GPP radio technologies, such as WiFi. These are also being developed within the 5G Core Network, though they stop well short of full convergence.
- Both 3GPP and WiFi 6 have adopted some common solutions, including frameworks based on the IEEE’s Extensible Authentication Protocol (EAP), to accommodate different wireless service types (mobile, wireless or fixed broadband) and their native authentication methods.
- 3GPP Release 15 provides some support for interworking between 5G and WiFi, including support for untrusted non-3GPP access (such as WiFi) to the 5G core via Non-3GPP Interworking Function (N3IWF); with secure transport of Control Plane/User Plane (CP/UP) messages over an IKEv2/IPSec tunnel between the terminal devices and the N3IWF.
- 3GPP Release 16 is continuing the work by enhancing capabilities for WiFi integration, including trusted WiFi support and access traffic steering, switching and splitting.
The wasted opportunity for standards body cooperation on 5G:
Ahead of the kick-off of the 3GPP’s work on Release 15, its first set of 5G standards, there were hopes that the IEEE, guardian of the 802.11 standards on which WiFi is based, might have more direct input to the deliberations than in previous generations.
But there was no 802 or WiFi official representation at the kick-off of the 3GPP’s 5G process in autumn 2015. In reality, 5G NR has been drawn up largely by the usual suspects, albeit with some participation from WiFi-centric bodies like CableLabs – but not from IEEE 802.11 itself. This has been a wasted opportunity, given that 5G was generally conceived as more than a set of 3GPP specs, but instead an umbrella framework embracing different radio technologies within a broad, software-driven platform.
The entrenched processes and politics of two standards bodies got in the way of valuable unification, it seems, leaving it to initiatives like that of the NGMN and WBA to drive convergence in a retrofitted way.
The IEEE did make various overtures to its cellular counterpart, culminating in one at the start of 2017, when it announced its 5G Initiative, calling on industry leaders, policy makers and academia to form a “neutral forum” to move 5G forward. This seemed to promise greater cooperation between the IEEE’s broad-ranging standards activities across most fields of connectivity, and the more narrowly focused 3GPP.
Although the Initiative’s working groups may well produce valuable work, it started too late to influence Release 15, whose launch date was accelerated anyway with the decision to fast-track the Non-Standalone subset. As soon as those specs were finalized, 3GPP secured its usual place as the lodestone of the new cellular generation, driving critical processes of device design, interoperability testing and consequent economies of scale. The huge WiFi ecosystem has those too, many driven by the WiFi Alliance, but they remain separate from the 3GPP processes, and addressing a different community.
That was why it was unfortunate that the IEEE 5G Initiative didn’t get started earlier, since it had strong representation from the cellular community – its co-chairs are Ashutosh Dutta of AT&T and Gerhard Fettweis, Vodafone chair professor at TU Dresden in Germany. He said: “The IEEE 5G Initiative is convening the vast breadth of IEEE resources in its members around the globe and new participants, to realize targets like one terabyte per second WiFi and 10Gbps cellular by 2025; one millisecond latency rate; and 25 bytes every 100 seconds for 10 years from a AAA battery.”
Working groups are focused around activities like the 5G Roadmap project, which will identify short (~3 years), midterm (~5 years), and long-term (~10 years) research, innovation and technology trends in the communications ecosystem for the purpose of establishing a living document with a clear set of recommendations, IEEE said.
So far, a common IEEE/3GPP approach to 5G has been elusive. In October 2016, Paul Nikolich, chair of the IEEE 802 standards committee, wrote to the chair of the 3GPP Project Coordination Group (PCG) to propose how they might work together. This was taken as a hopeful – if belated – step at the time, but no concrete news followed from that.
Nikolich wrote: “We believe it would be helpful for our organizations to identify mutually agreeable collaboration mechanisms that take into account the existing procedures of each organization.” But insiders indicated that there was little agreement on how each body might adapt its processes to the other’s, or get round the politics of prioritizing different working groups and projects.
A previous attempt to insert WiFi into the 5G process came in September 2015, when the 802.11 working group sent a liaison to 3GPP’s RAN taskgroup asking about the potential to include WLAN in an IMT-2020 submission (the ITU, via its IMT-2020 process, decides which submitted standards can be official 5G candidates).
While IMT-2020 can choose multiple candidates – as it did for 4G, when 802.16/WiMAX was on the list alongside LTE – the greater benefit for the industry would come from a converged platform, not a competition between rival candidates.
The IEEE has its own high level 5G project, which falls within its Standards Association (IEEE-SA). This, in turn, is part of the Global Standards Collaboration (GSC), which also includes the ITU plus bodies like ETSI, the TIA and ARIB (also 3GPP organizational or market partners).