Resolving Cable Wireless Conundrum Requires Clarity on Network Evolution

Jay Fausch, head of cable sector marketing, Nokia

Jay Fausch, head of cable sector marketing, Nokia

Interplay Between 5G and Virtualization Is a Key Consideration

By Fred Dawson

August 28, 2017 – Amid another flare-up in speculation over what big cable companies are going to do about mobile the overarching question is whether they are ready to act on a vision of the future of networking that is coming into focus at warp speed.

The debate over pursuing the MVNO path versus taking the M&A route into a big MNO play is a business question colored for much of this year by speculation about a dizzying array of possible deals, from a Comcast-Charter partnership looking to acquire Sprint or T-Mobile to Sprint seeking a merger with Charter to Comcast and Verizon combining into a $315-billion behemoth. In the latest wrinkles both Comcast and Charter have dumped a lot of cold water on the rumor mill by saying they’re committed to forging ahead with their MVNO deals with Verizon.

On an earnings call in late July Comcast chairman and CEO Brian Roberts, noting the MSO was already rolling out the Xfinity-branded mobile service using Verizon’s infrastructure, said, “We really feel we’re not missing anything ….I don’t see something happening in that industry that we envy a position that we don’t have today.”

Charter, responding to reports that Sprint owner SoftBank was looking to sell some or all of the company to the MSO, issued a statement saying “While we understand why a deal is attractive for SoftBank, Charter has no interest in acquiring Sprint. We have a very good MVNO relationship with Verizon and intend to launch wireless services to cable customers next year.”

Whether these statements are just negotiating ploys or a signal that everything has hit a dead end remains to be seen. Meanwhile, the real issue for cable operators, big and small, is whether they’re inclined to use their assets to full advantage to create the kind of network that mobile and telecom companies can only aspire to at this point.

Right now, as Verizon CEO Lowell McAdam noted recently, even his company with its extensive FTTH footprint is a long way from where it needs to be. In an interview with Bloomberg that focused on possible acquisitions, McAdam suggested he’d be open to talking about merging with any of three entities that came up in the discussion – Comcast, Disney or CBS. But then he said, “Given what I know about architecture, financial requirement, cultural fit, there’s never a dream deal.”

The dream deal, he added, would be one that involved an entity with all-fiber infrastructure attached to 5G microcells at the end points. “If I can find a company that had the fiber built for this architecture I’d scoop them up in a minute, but they don’t exist,” McAdam said.

Comparing 5G and Wi-Fi

While the telecommunications industry expects to be able to use 5G for mobile services, currently most agendas are focused on the advantages of using the technology for fixed wireless connectivity. 5G radios are designed to support robust performance at millimeter wave frequencies, where an abundance of spectrum allocated to 5G by the FCC and other bodies will enable fixed wireless connectivity at multi-gigabit speeds throughout the premises and in public places.

The mobile standards body 3GPP recently announced it was accelerating 5G standards development, prompting expectations that mobile uses of 5G will be possible sooner than previously expected. In May AT&T announced it could launch 5G mobile as early as year’s end 2018, but it will likely be a long time before 5G is widely available as a mobile service. Any practical implementations for mobile will require use of already saturated cellular spectrum, which is likely to lead to pressure on regulators to allocate more spectrum for migration to 5G.

Some cable executives have expressed strong interest in 5G, notwithstanding the prospects for ongoing expansion of Wi-Fi capabilities as embodied in emerging IEEE standards, which rely on some of the same advanced radio technologies used with 5G. For example, one commonly used core technology now employed with widely deployed 802.11ac Wi-Fi chipsets is MIMO (Multiple-Input, Multiple Output), which uses multiple antenna arrays in transmitters and receivers to support more robust transmissions through spatial separation of bit segments using SDMA (space division multiple access) multiplexing.

The newest 802.11ac chipsets, implementing the second wave of specifications issued for the platform, employ what’s known as Multi-User MIMO (MU-MIMO), which uses precoding to identify multiple receiving devices, thereby enabling a kind of broadcast mode that uses SDMA to transmit content delivered over a given frequency channel to more than one device. This has the effect of doubling or even quadrupling the number of simultaneously supported clients on each access point (AP).

5G goes much farther with Massive MIMO technology, which outdoes the 8×8 antenna configurations envisioned for the third wave of 802.11ac with support for 64×64 configurations in compact, commercially viable components. These antennas can focus the transmission and reception of signal energy into very small regions of space, providing new levels of spectral efficiency and throughput for more users in a dense area without causing interference.

Depending on spectrum availability for Wi-Fi, the MU-MIMO and other advances envisioned with Wave 3 802.11ac are expected to take effective throughput per AP to around 2.8 Gbps compared to 800 Mbps and 1.4 Gbps, respectively, for Wave 1 and 2 802.11ac. (Maximum PHY rates utilizing all available spectrum for the three waves are pegged at 1.3 Gbps, 2.34 Gbps and 6.933 Gbps.)

Wi-Fi doesn’t stop there. Specifications for the next generation of Wi-Fi in the existing 2.4 and 5 GHz spectrum zones, 802.11ax, are under development with the goal of pushing the PHY rate to the 10 Gbps level. And vendors are already certifying products supporting the 802.11ad WiGig protocol for transmissions at the 60 GHz tier reaching PHY throughput of 4.6 Gbps. Farther out, an enhancement to WiGig known as 802.11ay will raise the PHY ceiling to 100 Gbps.

In light of these developments it’s no wonder that many cable operators question whether they’ll ever need to use 5G to remain competitive. But it seems safe to say that Wi-Fi will never catch up with 5G as a fixed wireless solution as radio and chip technologies supporting both continue to evolve.

AT&T, describing its 5G rollout plans earlier this year, said it’s already running fixed 5G connectivity in lab tests at up to 14 Gbps. Given the much greater bandwidth available, the potential for mobility and other 5G advances that are not part of the Wi-Fi migration path, there’s a persuasive case to be made for cable operators’ transition to 5G at some point.

Just what the cable industry’s version of McAdam’s dream network might look like (with some coax running between the fiber and the microcells) can be seen in an emerging portfolio of next-generation products coming into the cable industry from the mobile and telecom sectors. Tight integration between HFC and 5G is definitely part of the picture, but so, too, is the power of cloud-orchestrated network virtualization to coordinate functions across a wide array of virtualization-optimized network elements.

As a case in point, several developments underway at Nokia, which acquired Alcatel-Lucent last year, can be pieced together to envision what’s in the offing for cable operators who are ready to move beyond the restrictions of legacy approaches to network migration. These touch on things like multi-terabit routing capabilities, remote PHY access infrastructure, virtualized network elements, distributed cloud architecture and big-data intelligence as well as 5G.

A New Benchmark in Routing

Nokia became a conduit for flowing advanced telecom technology into the cable industry by virtue of Alcatel-Lucent’s success marketing its edge routing solutions to MSOs as broadband took off in cable. With a longstanding corporate unit devoted to cable and the RF requirements of the HFC network, the company is well positioned to propose new ideas to operators and react to the resulting demand by adapting innovations coming out of Nokia Bell Labs and the company’s product development teams to cable’s requirements, notes Jay Fausch, who leads global marketing for Nokia’s cable MSO segment.

“Edge routing got us on the map with cable, and now there’s strong demand for deployment of our routers on cable backbone and core networks as well,” Fausch says. “The opportunities cable operators have to compete in a fixed-line, all-IP gigabit world are playing to our strengths.”

The same is true of the industry’s growing reliance on advanced wireless technology, he adds. “We see great opportunities around mobility in cable,” he says, “As operators address the mobility question and how to keep customers on their networks when they’re not in reach of Wi-Fi, there’s a lot of technology in the Nokia portfolio that can play in that transition.”

When it comes to routers, Nokia’s recently introduced 2.4 terabit-per-second FP4 network processor is driving capabilities in the company’s edge and core routers that will allow cable as well as other telecom companies to keep up with traffic and functional demands of all-IP service operations, Fausch says. “The FP4 is pushing the envelope on capacity without compromising on the capabilities you need to have in edge and core routers,” he notes. “It’s a significant improvement for us.”

The company says the FP4 increases edge routing capacity two to three fold across various models in the 7750 SR portfolio and by six fold in its 7950 XRS core router. For example, according to company specs, the single-shelf 7750 SR-14s supports a 144 Tbps configuration and can scale up to 288 Tbps. The 7950 XRS scales to 576 Tbps in a single system through chassis extension, without requiring separate switching shelves.

“Our next generation of terabit class routing leapfrogs other suppliers out there,” Fausch says, a point confirmed by Frank Ostojic, senior vice president and general manager of the ASIC Products Division at Broadcom. “Nokia is charting a course that others will have to follow,” Ostojic says, citing the firm’s use of several cutting-edge silicon technologies. These include 16nm finFET Plus process technology (where a fin-shaped electrode in a field effect transistor allows multiple gates to operate on a single transistor), Broadcom’s embedded SerDes (Serializer/Deserializer, a means of maximizing input/output capacity on chipsets) and advanced packaging.

Of course, in the seesaw battle among routing suppliers this might only be a temporary lead. But timing is essential as operators push ahead with preparations for future requirements. “Router interfaces and optical transport pipes will have to keep up” to enable full exploitation of these capabilities, Fausch acknowledges. Nonetheless, he adds, “This positions us very well for the cycle of scalable infrastructure upgrades taking place today.”

The Cable Virtualization Mandate

Whereas the move to higher capacity routing is more or less a no-brainer for cable operators, another key but less certain step in preparations for competing in an environment where everybody and everything is connected all the time has to do with adopting approaches to network virtualization. One area of debate where the rubber is meeting the road right now concerns best approaches to virtualizing CCAP (Converged Cable Access Platform) systems.

As previously reported, vendors are lining up on different sides of the debate with introduction of CCAPs that are meant to work with the CableLabs-defined Distributed Access Architecture (DAA) model, which relies on digital optics terminated at the HFC node by Remote PHY electronics that manage modulation, multiplexing, forward error correction and other physical layer processes in the conversion to RF for distribution over coax. The question posed by the new vendor options is, once the CCAP is relieved of performing these PHY layer processes, which components of the DAA–enabled CCAP should be virtualized.

Some vendors are virtualizing both the control and data plane components of the CCAP to run on COTS (commodity off-the-shelf) servers. Nokia, Huawei and possibly others are supporting a Remote MAC-PHY configuration with solutions that virtualize the control plane functions running in core locations to orchestrate provisioning, quality control and tie-ins with other back-office elements but move the CCAP data plane or MAC (Media Access Control) into the node.

It’s still unclear which way operators will go, Fausch says. “There’s a lot of lab and field trial activity as operators deal with explosive traffic growth and the need to reduce service group sizes and deal with increased demands on space and power in headends,” he notes. While the Nokia approach offers significant savings in power consumption and space utilization, “not everybody is ready to cross that bridge.”

“We feel traditional players in the CMTS business have a bit of embedded business to protect, so they’re not that anxious to see operators moving to remote PHY applications,” he adds. “But if you don’t go that way, you end up with a lot of big iron in the network you don’t need.”

The idea of a fully distributed virtualization architecture where software-based functions performed on COTS hardware in remote locations can be orchestrated through highly automated cloud-based workflows is gradually taking hold in the broader telecom industry. But it’s an evolutionary process where different carriers are focusing on different areas of virtualization rather than converting to full virtualization all at once. So far, most cable companies have yet to take these first steps.

“Elegant evolution is a pretty big pill to swallow,” Fausch says. “We’re emphasizing that as the industry heads into this all-IP fiber-rich gigabit-enabled world, the more you can recognize that and cloud-enable it to get the flexibility and agility in the network that you need for making adjustments to customer behavior and demand for services, the better.”

“If that’s where you want to go, we can get you there sooner,” he adds. “But until it smacks you in the face, it’s tough to bite the bullet.”

Virtualization and 5G

Looking to where things are going with respect to virtualization necessarily impacts how the industry views use of next-generation wireless. 5G has great significance here with capabilities enabling allocation of spectrum for data flows tuned to specific service categories.

A just-announced project getting underway in Europe points to what this could mean for service providers of all stripes. The 5G Mobile Network Architecture research project (5G-MoNArch) brings the architectural concepts articulated by phase 1 of Europe’s 5G Infrastructure Private Public Partnership (5G-PPP) into play with industry-driven use cases and two real-world testbed implementations.

Coordinated by Nokia under the auspices of the European Union’s €80-billion multi-technology 2020 Framework Programme, the project involves a consortium of 14 industrial and academic partners who have aligned to kick off Phase 2 of 5G-PPP, marking a significant step toward ensuring a common approach to service launches in the years ahead. Ultimately, the flexible and programmable architecture will support the vast variety of services, use cases and applications that will be part of 5G-enabled networks, participants say.

“We follow a shared architecture of what the next-generation communications infrastructure needs to look like to enable and meet the network demands of the next decade,” says Peter Merz, head of end-to-end mobile networks solutions at Nokia Bell Labs. Underscoring the magnitude of what has to be accomplished, he adds, “5G communication needs both private and public entities to invest in the infrastructure and ensure Europe remains competitive.”

The goal of the 5G MoNArch’s project is to use network slicing, which capitalizes on the capabilities of software-defined networking (SDN), network functions virtualization (NFV), orchestration and analytics to support a variety of use cases in vertical industries such as automotive, healthcare and media. Network slicing, a technique where the network is logically rather than physically sectorized, is deemed crucial to flexibly mounting and simultaneously supporting various services with widely varying requirements.

In other words, it’s all about maximizing the benefits of multi-terabit throughput in coordination with the service agility enabled through network virtualization. The 5G MoNArch consortium expects to develop detailed specifications and extensions of 5G architecture utilizing key enabling innovations such as inter-slice control and cross-domain management, experiment-driven modeling and optimization and native cloud-enabled protocol stack.

The two use cases earmarked for the project will deploy the architecture in live testbeds, one supporting heavy communications usage in a high tourism urban environment and the other enabling reliable and secure communications in a seaport environment. Such use cases, of course, are far from the traditional role played by cable companies. But in a marketplace where pay TV is no longer the defining service, operators have already gone a long way toward positioning themselves as competitors in the broader telecom environment with superior broadband connectivity and aggressive expansion into ever larger segments of the commercial services market.

Going forward, competitive strength will depend on operators’ ability to deliver value-added services over their broadband pipes, including superior converged entertainment, smart home and other benefits for the residential market as well as support for things like virtual VPNs, SaaS (software-as-a-service), video surveillance, IoT (Internet-of-Things) applications and a host of other extras that are central to serving the fast-evolving business market. Of course, ubiquitous wireless connectivity suited to meeting demand in a video-saturated environment where 4K UHD and virtual reality will be part of the bandwidth-guzzling mix will be mandatory.