CSP Evolution to 5G and Beyond

hpe-intel-lockup

‘5G’ seemed to be everywhere at this year’s Mobile World Congress in Barcelona. It was printed on the new products, prominent on the signage and was a major topic in the conference. Everyone was talking ‘5G’ but had you listened carefully you might have heard them asking: what is it exactly and what made it different from Gs 1, 2, 3 and 4?

What follows is our answer, gathered and distilled from the many interviews and videos TelecomTV conducted at the show. To assist the explaining we’ve developed a fun new metaphor. In fact 5G is a chocolate and marshmallow cookie. Read on.

‘5G’ seemed to be everywhere at this year’s Mobile World Congress in Barcelona. It was printed on the new products, prominent on the signage and was a major topic in the conference. Everyone was talking ‘5G’ but had you listened carefully you might have heard them asking: what is it exactly and what made it different from Gs 1, 2, 3 and 4?

What follows is our answer, gathered and distilled from the many interviews and videos TelecomTV conducted at the show. To assist the explaining we’ve developed a fun new metaphor. In fact 5G is a chocolate and marshmallow cookie. Read on.

We think we’ve developed a new way to visually describe 5G. Taking a leaf from the Google playbook (its Android versions are mostly calorific – Cupcake, Donut , Eclair, etc. ) we think 5G – or H1.0 for Heterogeneous 1.0 as we’re recommending – is best explained and understood by a chocolate and marshmallow cookie. The biscuit base is the physical network infrastructure, the chocolate is the exciting but thin radio network edge with all its antennas, and the marshmallow is the fluffy virtual network cloud which dynamically controls all the connections.

We think we’ve developed a new way to visually describe 5G. Taking a leaf from the Google playbook (its Android versions are mostly calorific – Cupcake, Donut , Eclair, etc. ) we think 5G – or H1.0 for Heterogeneous 1.0 as we’re recommending – is best explained and understood by a chocolate and marshmallow cookie. The biscuit base is the physical network infrastructure, the chocolate is the exciting but thin radio network edge with all its antennas, and the marshmallow is the fluffy virtual network cloud which dynamically controls all the connections.

Welcome to the heterogeneous network: So here’s the final shape of our cookie. Virtualization is all very well, but you still need a big chunk of crunchy infrastructure to abstract your networks from – that infrastructure runs through the biscuit base and note how it gets physically very close, on the perimeter of the base, to the dome of chocolate radio access network. That’s where the network edge meets the radio network for MultiAccess Edge Computing. But the key element of this diagram is, of course, the marshmallow filling… the crucial SDN/NFV element that will turn multiple and overlapping network types into the next generation, post 5G, heterogeneous network.

Welcome to the heterogeneous network: So here’s the final shape of our cookie. Virtualization is all very well, but you still need a big chunk of crunchy infrastructure to abstract your networks from – that infrastructure runs through the biscuit base and note how it gets physically very close, on the perimeter of the base, to the dome of chocolate radio access network. That’s where the network edge meets the radio network for MultiAccess Edge Computing. But the key element of this diagram is, of course, the marshmallow filling… the crucial SDN/NFV element that will turn multiple and overlapping network types into the next generation, post 5G, heterogeneous network.

Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01 Image 01

Why the 5th generation might be the final G

and why it’s all best understood as a chocolate and marshmallow biscuit

It’s a truth universally acknowledged that any market in possession of a viable access technology must be in want of a generational narrative with which to sell it (apologies to Jane Austen).

Narratives are usually a good thing. I couldn’t structure this text and companies couldn’t market their products and services without them. But sometimes a well-worn narrative gets worn out.

Let’s think about the mobile industry’s generational narrative. That’s the mobile story as a series of distinct technical progressions – from 2G (GSM) to 3G to 4G (which is actually 3G as it’s the Long Term Evolution of 3G) and now on to 5G.

Up to now the generational approach has been reasonably successful. By and large it has kept the mobile industry focused on formulating and supporting specific standards in a timely fashion, so that the world’s networks and their terminals could interwork properly. Just as important, users could be persuaded to discard the current ‘G’ in favour of the next one – so we could harness ‘aspiration’ for each new ‘G’ progression and really shift those units.

Track now: click in box

But now, with the advent of 5G, we’ve noticed a change. Our TelecomTV interviewees and panel members increasingly question whether the generational tag is still the best framework for understanding the market they are trying to grapple with. They say the descriptor, ‘5G’ no longer captures the market’s complexity, with this generation’s range of standards at different frequencies being prepared for a whole gamut of use cases – from low bandwidth IoT, through to high speed mobile broadband, fixed link broadband and Wi-Fi, ultra-low latency broadband and high speed IoT; and all of these being introduced at different times in different markets and all of them being ‘5G’.

In fact some observers believe that the ‘G’ lost its relevance at around #4

5G also implies something standardised and static – implying a network offering x, y and z for ten years after which 6G will arrive to start the process over again. As we’ll show, this won’t be the case.

5G changes the paradigm

Far from being nailed down, the SDN/NFV network expected to be built behind the radios to support 5G will go on to support the next iterations of radio technology and even fixed line interfaces because it will likely be driven by software and able to adapt gracefully for, as far as we can envisage now, forever.

Last but not least it appears difficult for everyone not totally immersed in the technical side of the mobile industry, let alone the general public, to grasp the fifth G as a concept. Even top telecoms executives have come to expect a ‘G’ to have a tight definition, not a pic ‘n’ mix of relatively vague use cases, and often appear frustrated that it does.

So we’re setting out the case for the end of G and urging a logical progression to ‘H’ for ‘Heterogeneous’, which might be a better way to understand and characterize the ‘network’ in all our imaginations.

20 seconds: NFV first, 5G later
50 seconds: From 5G-Ready to 5G Deployment: Telco strategies for mobile evolution

How 5G will evolve from and overlap with 4G

and where does mobile edge computing fit in?

So why is 5G so different from the Gs that came before it?

There has always been considerable generational overlap in mobile. The analogue cell network (1G) stayed switched on as the first digital 2G (GSM) network was built out. 2G stayed around for 3G which extended the same courtesy to 4G/LTE. In fact my own operator (who shall remain nameless… Vodafone) has only recently stopped attaching my smartphone to the painfully slow 2G GPRS service when 3G was absent (which was most of the time). LTE now shows up on my screen most of the time but my service, like many, defaults to 2/3G voice when a call comes in or goes out – there’s no Voice over LTE (VoLTE) yet.

But our interviewees are expecting more than just a little overlap between 4G/LTE and 5G. Telcos are being urged to virtualize their networks in advance of 5G specifically so they may deploy 5G-like services using their 4G technologies. Plus, LTE will not simply stick around for a while, it will continue to develop in its own right, more or less in parallel. This alone is enough to make the transition to 5G very different from what we’ve experienced before.

One way to think about this glide factor, is to see plain old LTE as 3.5G or pre-4G, LTE-A as 4G, and LTE-A Pro as 4.5G or pre-5G. LTE-A (for advanced) is taking top speeds up to 1 Gigabit/s and reducing latency, but many eyes are on the ultra-enhanced version – LTE-A Pro – which manages more than 3 Gbit/s by aggregating multiple channels and reducing latency down to 2 milliseconds from LTE-A’s 10 milliseconds. Obviously there is a lot of life left in LTE for many years to come. Target applications include high definition video and virtual and augmented reality.

Most intriguing is the idea of ‘synergies’ where 4G and 5G radio services work with each other to overcome difficulties. LTE at 600MHz might be good at penetrating buildings while ‘new radio’ 5G at very high frequencies is bound to find walls a problem. A service might be anchored by its LTE radio but then pick high speed 5G channels or indoor small cell service under instruction – there are many permutations.

Track now: click in box

The new MultiAccess Edge

Edge or ‘Fog’ computing is where cloud facilities are distributed right to the edge of the network to improve latency for demanding applications – autonomous car is often mentioned but there are others.

Recently, ETSI’s ‘Mobile Edge Computing’ ISG was renamed to Multiaccess Edge Computing to "embrace the challenges in the second phase of work and better reflect non-cellular operators’ requirements," the organization said. It seems to have taken time and much gentle argument to begin to define the network edge as heterogeneous thus enabling other, non-cellular protocols (ie fixed broadband access/WiFi) to get a look in. Vociferous supporters of multiaccess inclusiveness included telcos with both mobile and fixed operations.

40 seconds - "It’s really important to keep evolving your 4G network and then introduce 5G in that context."
20 seconds - "When operators move to a virtualized architecture they don’t want to be investing in one technology now and then have to upgrade later."
30 seconds - "We looked for the killer app, but turned out that every vertical had their own."

The ETSI MEC says its future work will take into account heterogeneous networks using LTE, 5G, fixed and Wi-Fi technologies. So while MEC is most often paraded as a ‘5G’ capability because of its ability to help get round trip delay down below 1 millisecond for demanding applications, once it’s properly installed at the edge of an NFV/SDN driven network it can work on reducing latency across 4G, WiFi and fixed applications as well.

View full video: Angela Whiteford, VP Product Management and Marketing, Affirmed Networks

The idea that 5G must involve sudden change and standards ‘deadlines’ also impacts preconceived notions of what it might mean to ‘go early’ with network investment in preparation for accommodating 5G on the network. Installing pre-standard hardware in the days of 3 and 4G might have been a big risk, installing pre-standard software is different.

20 seconds - "There’s very little risk [involved in moving ahead early] because most of the critical aspects of 5G are software based"

Adjacencies, enterprise solutions and spectrum sharing

The next generation of radio access technologies is being designed to economically reach out to every wireless and mobile application that we know, plus quite a few more that we haven’t yet thought of.

In doing that the technologies won’t necessarily be deployed only by telcos, all manner of organizations would like to step up and play a part. These new or adjacent players include: metropolitan authorities looking to build smart city infrastructure, big verticals such as media companies and large industrials, cable companies and enterprises of all kinds. In fact it’s thought that many of the more ambitious network types – autonomous vehicle support networks for instance – will need adjacent players to put finance in just to make network build viable.

Track

Track now: click in box

Then there’s network sharing between telcos – again, the level of investment required is such that regulatory authorities are being urged to relax competition laws to ensure new generation network goals are met.

More than anything, next generation services and applications will live in the general-purpose virtualized and software driven network of networks – that fluffy cloud that we’ve characterized as a marshmallow filling, will manage and orchestrate the data flows in a highly agile and flexible way, which is why we’ve characterized it as we have: as adaptable, able to accommodate thousands of virtual connections and just as able to make changes to them all in milliseconds.

Track

Track now: click in box

However, it’s likely that the sort of large-scale telco cloud deployments capable of supporting a comprehensive 5G virtual network strategy won’t materialise in great numbers until after 2020. So the first focused early 5G deployments will probably be well-understood services such as mobile and fixed broadband and won’t need a new telco cloud network.

Track

Track now: click in box

5G should also enable new telco business models. There will be scope for operators or MVNO/retail service providers to specialise to pursue different user segments – especially in the enterprise market. There may also be scope for geographically specialised players: the idea here being that local operators could build both indoor and outdoor small cell infrastructure and offer local context-related services via locally-issued spectrum licences, perhaps on a shared spectrum basis.

These operators could own the infrastructure and support the users of all macro networks if and when they were within range. (the banks’ shared ATMs are an interesting example of this approach), thus offering efficient local service delivery at a given location with guaranteed quality. Everyone wins: macro MNOs don’t have to overbuild each other’s small cell networks in dense urban locations – in big shopping malls or transport terminals, for instance. Municipal authorities don’t have to suffer the disruption of multiple network builds in the same locations.

Spectrum sharing is expected to be an important development

With all that high-frequency radio signal failing to make it through walls, attention is being paid to indoor coverage and how best to enable it. Shared-access regulatory frameworks provide more spectrum for all the players to play with and can also foster a Wi-Fi/cellular integration approach that leads to new models for cost sharing and the ownership of the wireless infrastructure.

Track

Track now: click in box

This comes at a time when venues are taking a greater interest in providing communications amenity for their customers as operators, struggle with how to deploy outdoor or indoor small cells in mid-size venues in particular.

One of the trends is that venue owners – keen to improve their own user experiences – seem more and more willing to pay for both the access and the infrastructure, taking the burden off operators. This way the venue owner gets control and can provide economic access to multiple operators.

45 seconds - "An understanding of the vertical is really critical in order to implement it properly"
25 seconds - "Enterprise players, now with access to telco grade equipment, can be bigger and bigger players"

Are we nearly there yet?

Standards, technologies and progress so far

The first wave of 5G smartphones, supporting the ultra speedy 3GPP 5G NR (New Radio) standard, are expected to arrive as early as 2019 thanks to a recently agreed ‘intermediate’ specification called ‘non-standalone’ which relies on an LTE network to act as an anchor and provide the mobility management and LTE packet core for the dual-mode handsets.

Track

Track now: click in box

View full video: Adrian Scrase, CTO, ETSI

It means large-scale trials and deployments can start in 2019. Training wheels, you might say. Meanwhile full standalone mode will be part of the next 3GPP Release (Release 15) so telcos (or a new wave adjacent operators) will be able to launch compliant networks without needing access to an LTE network to anchor it in.

Track

Track now: click in box

Obviously this gives existing operators an early-start advantage over ‘greenfield’ challengers since the former can get to market first, but it will be interesting to see how long non-standalone (or tethered, we might call it) persists, given that finicky high frequency mmWave spectrum as well as sub-6 GHz spectrum, is to be supported in NR. It could be that non-standalone quickly becomes accepted as the best solution for operators.

Track

Track now: click in box

View full video: Paul Crane, Head of Mobile and Wireless, BT

There may be widespread uncertainty about how quickly telcos will adopt the various elements of 5G and how they will adapt their business models to operate in a more open and heterogeneous environment. What’s not in doubt is the huge effort going into technical development – both software and hardware – for all the different technical strands of 5G. TelecomTV’s Tracker – try it –  which selects Insights from the industry’s press release stream and provides tools to analyse them, reveals how the industry’s attention has switched to 5G over the past year. We’ve logged 654 ‘Insights’ related to 5G, 61 per cent of the total TelecomTV Tracker entries.
 Track

Track

Tracker entries for 5G spotlight both the upsurge in collaborative activity – joint trials, PoCs and participation in multiple open source initiatives – as well as product launches, market research and standards development. It also illustrates the level of ferocious competition between telcos, between vendors and between countries. The urge to secure ‘bragging rights’ has seen our ‘Towards 5G ‘ category simply swamp the other categories – NFV/SDN and IoT included.

Track

Track now: click in box

The killer is that there’s no killer: diversity is the killer app with 5G

In TelecomTV’s recent 5G survey we asked respondents to rank the five 5G services they were going to tackle first. IoT and video streaming came in as the two most often mentioned areas: no surprises there. What did come as a surprise was the sheer range of first rank choices, indicating that 5G in one form or another was clearly going to be targeting every application going.

So we arrive at the heterogeneous network: There’s no lack of awareness that something called 5G is on the way. There is, however, less clarity about what it is and when it will start to be available. Already some telcos are claiming to have 5G phones now, whereas in truth "5G" – as defined by the 3GPP standards body – won’t assume its first ‘frozen’ form until March 2018. In reality, as our many MWC conversations show, 5G will be embodied in a range of overlapping network types introduced at different times in different markets for different applications and what we currently call 4G will be with us for a while longer also.

The first wave of 5G smartphones, supporting the ultra speedy 3GPP 5G NR (New Radio) standard, are expected to arrive as early as 2019 thanks to a recently agreed ‘intermediate’ specification called ‘non-standalone’ which relies on an LTE network to act as an anchor and provide the mobility management and LTE packet core for the dual-mode handsets.

Track

Track now: click in box

View full video: Adrian Scrase, CTO, ETSI

It means large-scale trials and deployments can start in 2019. Training wheels, you might say. Meanwhile full standalone mode will be part of the next 3GPP Release (Release 15) so telcos (or a new wave adjacent operators) will be able to launch compliant networks without needing access to an LTE network to anchor it in.

Track

Track now: click in box

Obviously this gives existing operators an early-start advantage over ‘greenfield’ challengers since the former can get to market first, but it will be interesting to see how long non-standalone (or tethered, we might call it) persists, given that finicky high frequency mmWave spectrum as well as sub-6 GHz spectrum, is to be supported in NR. It could be that non-standalone quickly becomes accepted as the best solution for operators.

Track

Track now: click in box

View full video: Paul Crane, Head of Mobile and Wireless, BT

There may be widespread uncertainty about how quickly telcos will adopt the various elements of 5G and how they will adapt their business models to operate in a more open and heterogeneous environment. What’s not in doubt is the huge effort going into technical development – both software and hardware – for all the different technical strands of 5G. TelecomTV’s Tracker – try it –  which selects Insights from the industry’s press release stream and provides tools to analyse them, reveals how the industry’s attention has switched to 5G over the past year. We’ve logged 654 ‘Insights’ related to 5G, 61 per cent of the total TelecomTV Tracker entries.
 Track

Track

Tracker entries for 5G spotlight both the upsurge in collaborative activity – joint trials, PoCs and participation in multiple open source initiatives – as well as product launches, market research and standards development. It also illustrates the level of ferocious competition between telcos, between vendors and between countries. The urge to secure ‘bragging rights’ has seen our ‘Towards 5G ‘ category simply swamp the other categories – NFV/SDN and IoT included.

Track

Track now: click in box

The killer is that there’s no killer: diversity is the killer app with 5G

In TelecomTV’s recent 5G survey we asked respondents to rank the five 5G services they were going to tackle first. IoT and video streaming came in as the two most often mentioned areas: no surprises there. What did come as a surprise was the sheer range of first rank choices, indicating that 5G in one form or another was clearly going to be targeting every application going.

So we arrive at the heterogeneous network: There’s no lack of awareness that something called 5G is on the way. There is, however, less clarity about what it is and when it will start to be available. Already some telcos are claiming to have 5G phones now, whereas in truth "5G" – as defined by the 3GPP standards body – won’t assume its first ‘frozen’ form until March 2018. In reality, as our many MWC conversations show, 5G will be embodied in a range of overlapping network types introduced at different times in different markets for different applications and what we currently call 4G will be with us for a while longer also.