Open RAN architecture set to disrupt 5G

landscape
Operator Buying Pa!erns to 2025

RAN RESEARCH: THE RESEARCH ARM OF WIRELESS WATCH

Companies mentioned in this report; Airrays, Airspan, Allot, Alti-

ostar, Amarisoft, Anikowave, ARM, Askey, ASOCS, Athonet, AT&T, RETHINK
TECHNOLOGY
Beicells, Ball Aerospace, Bell Labs, Bharti Airtel, British Telecom, China RESEARCH
Mobile, China Mobile Research Institute, China Unicom, CHIPS Alli-
ance, Ciena, Cisco, Cloud-RAN Alliance, Comba, Deutsche Telekom, Bristol & Exeter House,
Docomo, Ericsson, Facebook, Fujitsu, HFR, Huawei, H3C, Innoeye, Lower Approach Road,
Temple Meads,
Intel, KDDI, Keysight, KT, Lenovo, Linux Foundation, Mavenir, Bristol
Metaswitch, NEC, Netcracker, Nokia, NTT, OKI, ORAN Alliance, Or- BS1 6QS
ange, Parallel Wireless, Phluido, Pivotal Commware, Quanta, Qual- Phone: +44 (0)117 925 7019
comm, Radisys, Rakuten, Red Hat, Reliance Jio, Ruckus Wireless, Sam- Fax: +44 (0)117 329 1480
Email: info@[Link]
sung, Sercomm, Singtel, SK Telecom, Softbank, SOLiD, Sprint, Tech
Mahindra, Telefónica, Telstra, TIM, Verizon, Viavi, Vodafone, WiFi
Alliance, Xilinx, ZTE.

Authored by Caroline Gabriel

“Lea !n" $%& w'( )i$* !n+,-,n+,n.

$h/u0h. 'n1 n2 3e45r4ita6e1 'n78y9i:”
 C O NT E NT S

Table of Contents 2
Table of Figures 3
Introduction 4
Many forces are driving the mobile industry towards an open RAN 5
The key drivers for an open ecosystem 5
The initiative lies with the operators, old and new 7
Semi-open interfaces of the past have been inadequate 8
Operators look for ways to bypass X2 in 5G 9
CPRI shortcomings have driven push for new fronthaul standard 11
Ethernet alternatives pave way for full openness 13
The economic impact of openness will be seen in all layers of the network 15
Several open initiatives are starting to make the case for the open RAN 18
ORAN Alliance 20
ORAN makes progress with PoC tests 23
Telecom Infra Project 24
Steps to reality for TIP 27
Potential convergence of open groups? 29
Small Cell Forum 30
The Open Air Interface 32
Case studies 34
Telecom Italia works with three new vendors on vRAN 34
Nokia and China Mobile 35
Rakuten claims first cloud-native, multivendor mobile network 36
There are still many barriers in the way of a fully open RAN 38
How will open RAN platforms affect the pa#ern of 5G deployment? 40
Baseline forecasts for base station deployment 42
Scenario 1 - large-scale adoption from 2022 44
Scenario 2 - at-scale deployment limited to small cell layer 46
Scenario 3 - limited deployment before 2025 50
Scenario 4– limited open ran but agreed to open fronthaul interface 51
Conclusions 53
Methodology 54
RAN Research: Forecasting disruption in wireless 56
About Rethink Technology Research 58

Copyright © 2019 Rethink Research, All rights reserved.

 TA B L E O F FI G U R E S

Figure 1. Figure 1. Top drivers to adopt open architectures in the RAN – percentage of operators placing
each driver in their top three 7
Figure 2. The role of X2 in 4G/5G interworking 9
Figure 3. The evolution of SK Telecom’s 4G and 5G RAN Source: Netmanias 12
Figure 4. Fronthaul options related to different 3GPP/IEEE functional splits in a disaggregated RAN 13
Figure 5. Projected deployment cost of a 5G macro cell, open vs traditional architecture 15
Figure 6. Primary areas of impact of open architectures on the macro network - percentage of operators
placing each factor in their top two 16
Figure 7. Primary areas of impact of open architectures on the small cell network - percentage of operators
placing each factor in their top two 17
Figure 8. Selected challengers which are active in the open RAN, and their key relationships with top
MNOs. 20
Figure 9. Conceptual architecture of xRAN (now ORAN). Source: xRAN white paper 2016 21
Figure 10. The TIP OpenRAN group 26
Figure 11. Percentage of operators active or interested in each major open mobile network group 30
Figure 12. Top barriers to adopt open architectures in the RAN – percentage of operators placing each bar-
rier in their top three 39
Figure 13. Total deployments and upgrades of macro and micro base stations by region 2018 to 2025, all
architectures 43
Figure 14. Total installed base of small cells by region 2018 to 2025, all architectures 44
Figure 15. In Scenario 1, deployments of open macro and micro base stations by region 2020 to 2025 45
Figure 16: In Scenario 2, deployments of open macro and micro base stations 2020-2025 46
Figure 17: In Scenario 2, installed base of open small cells 2020-2025 47
Figure 18: Deployment of open small cells by operator type 2020-2025 48
Figure 19: Deployment of open small cells, indoor vs outdoor 2020-2025 49
Figure 20: In Scenario 3, total installed base of small cells, all architectures 2018-2025 50
Figure 21: In Scenario 4, fronthaul and midhaul links deployed, by architecture 2019-2025 52

This document contains explanatory notes and commentary to accompany the Ex-
cel spreadsheet ‘Rethink Research Open RAN Module 2018-2025’. The surveys and
forecasts on which the outputs are based were conducted in January to March 2019.

Copyright © 2019 Rethink Research, All rights reserved.

Introduction

There are many ways in which operators expect 5G to be very different from
its predecessors. Only some of these relate to the performance of the radio
technology itself – the higher data rates, lower latency, enhanced availability
and so on. Others, which could prove just as important to the economics of
deploying the new networks, relate to the wider architecture, the supply
chain and the deployment models.

Operators hope that 5G will support an open, multivendor network with an

enlarged supply chain and greater price competition and open innovation.
This may always have been a goal, but it has not been easily achievable be-
fore.

Indeed, in the 4G era, the supply chain has contracted considerably because
of the consolidation of suppliers. In 2019, there are five major providers of
macro base stations, and for operators in countries which are restricting pro-
curement from Chinese OEMs, notably the USA and Australia, there are on-
ly three choices. This greatly reduces price competition and encourages lock-
in.

Although most MNOs have a dual-source policy for their macro RANs, this
means li#le more than spli#ing the network into geographical zones, with
one supplier for each and minimal need to interwork.

But the industry is changing and MNOs are under unprecedented commer-
cial pressure as they move towards 5G. The disconnect between rising data
usage and flatlining ARPUs is making it very difficult for most operators to
justify a significant investment in another generation of networks. It is essen-
tial to their business case that costs are reduced significantly, at the same
time that the new networks enable new revenue streams.

Most operators report that they are finding it hard to make a firm commer-
cial case for new revenues generated by 5G – particularly those that harness
its low latency, high availability qualities to support industrial and Internet
of Things (IoT) services. Over time, as the technology is enhanced and the
most promising use cases identified, this perception of high risk in the reve-
nue model will change, but for the first few years of a 5G deployment, it is

Copyright © 2019 Rethink Research, All rights reserved.

essential that operators can be confident of at least one improvement in
the model – the total cost of ownership (TCO).

One way to drive down cost will be to move towards a more open,
multivendor ecosystem with high levels of vendor competition and
lower barriers to new suppliers. In Rethink’s analysis, there are many
factors which are encouraging operators to push for a more open eco-
system. But other drivers relate to a trend that is more threatening to
the established MNOs - the potential for 5G, especially when it is de-
ployed in shared spectrum and with a cloud-native, sliceable core, to
enable new service providers and deployment models.

Many forces are driving the mobile industry towards an

open RAN

A convergence of commercial imperatives will drive openness and

interoperability. Some of these will mainly relate to pressures on the
hard-pressed MNOs, to deliver 5G profitably; while others will relate to
the new entrants which want to add mobile connectivity to their own
business models in the enterprise or cloud.

T H E K EY D R I V E RS F O R A N O P E N E CO SY S TE M:

· Virtualization – this leads to commoditization of hardware, and

entry to the market of new suppliers with experience in the open
data center market. It is often the trigger for a deep review of an
operator’s architecture and processes, and even the ownership and
management of the network (there will be more RAN sharing,
and more reliance on managed services in 5G than in 4G).

· The rise of merchant chips, to replace proprietary chips in complex

systems like base stations. These drive a more open overall platform
based on general purpose architectures e.g. ARM, Intel.

Copyright © 2019 Rethink Research, All rights reserved.

· Densification, and the emergence of smaller base stations - from
mini-macro to microcells to small cells - which can be more easily
commoditized and have a more diverse vendor landscape.

· The urgent need for new TCO model, without which most operators
will not have the business confidence to deploy at scale. Many oper-
ators plan to deploy a limited number of 5G base stations in areas of
high data traffic and high value users, but to rely on 4G for wide ar-
ea services until the cost of a bigger 5G roll-out is well understood.

· Operators are resistant to further vendor lock-in, and large players

like AT&T are taking initiative to drive open platforms, in order to
encourage new suppliers and put pressure on existing ones to re-
duce their costs and adopt open interfaces.

· Major ecosystem influencers e.g. Facebook are driving alternative

supply chains across the telco market. Facebook’s Telecom Infra
Project (TIP), like its Open Compute Project in the cloud world, is
designed to demonstrate how the cost of telco networks can be dra-
matically reduced through the use of open platforms and commodi-
tized hardware.

· New deployers are entering markets with their own supply chains
e.g. webscalers, private enterprise network operators and neutral
hosts see more open 5G architectures, and shared spectrum, as de-
velopments that will help them enter the cellular space. They bring
their own suppliers and processes with them.

· The rise of private networks for enterprise, industrial and critical

IoT services also encourages a new, broader ecosystem around 5G,
including enterprise integrators and developers, and these often
have experience of open or open source technologies and processes,
from the enterprise world.

· The increasing robustness and acceptability of open and open

source processes and platforms in telecoms. This begun on the de-
vice side, notably with Android, and now many major infrastruc-
ture projects are taking place in open source. Companies like Nokia,
Samsung, Mavenir, Metaswitch, Radisys and others have donated
code to open source initiatives related to the RAN and core.

Copyright © 2019 Rethink Research, All rights reserved.

T H E I N IT I A TI V E LI E S W I TH TH E O P ER A T O R S, O L D A N D N E W

These developments will set the scene for a new open ecosystem, but
the initiative will rest firmly with the large operators – mobile carriers,
and other large service providers such as cloud providers - to make this
a reality.

In Rethink’s survey of 76 mobile and non-traditional operators, con-

ducted in the first quarter of 2019, seven key drivers emerged which
will motivate the service providers to push a new supply chain based
around an open architecture. Figure 1 illustrates these drivers. The
leading ones relate to cost – simple TCO reduction, and improved
competition in the chain.

But others relate to the need to accelerate development of new plat-

forms and expand the base of innovation beyond a few large vendors.
Over 40% want to put the technology agenda back in the hands of op-
erators rather than the OEMs, while 39% see multivendor interopera-
bility as an end
in itself, and 32%
believe that open
interfaces are
necessary, spe-
cifically to make
it possible to mi-
grate to virtual-
ized RAN archi-
tectures (see
Chapter 3).

Figure 1. Top
drivers to adopt
open architec-
tures in the RAN
– percentage of
operators placing
each driver in
their top three
(respondents contributed all their drivers in a free answer. They were
then asked to prioritize the most-cited factors).

Copyright © 2019 Rethink Research, All rights reserved.

Semi-open interfaces of the past have been inadequate

This will not be a smooth path. The top five OEMs (Huawei, Ericsson,
Nokia, ZTE and Samsung) are entrenched in existing networks to an
unprecedented degree thanks to years of consolidation.

As the vendors have consolidated, operators have increasingly found

themselves in an uncompetitive supply chain, which they have strug-
gled to broaden, because they have been locked in to end-to-end,
supplier-specific networks.

The radio standards may be uniform, but the networks have been im-
plemented in highly individual ways. Vendors have made proprietary
‘enhancements’, as well as their own unique interpretations of suppos-
edly open interfaces like CPRI and X2 – the result being incompatibili-
ties which make it hard for operators to mix and match equipment in
the same network, or to move easily to a new vendor. So the goal of
multivendor interoperability is not new, but has never been fully
achieved.

Mobile networks are entirely based on the work of standards organiza-

tions (SDOs), notably the 3GPP and ETSI. This has, to some extent,
ensured interoperability of many suppliers’ devices with the cellular
networks. However, this has still required considerable testing of each
device to perform optimally on each carrier’s RAN, by contrast with the
WiFi ecosystem, in which a single industry organization, the WiFi Alli-
ance, takes charge of testing and certification of all devices for all WiFi
networks.

There was an a#empt, when WiMAX was considered a candidate to be

a 4G technology, when that community sought to bring WiFi-style certi-
fication to the cellular world, but when WiMAX failed, the industry re-
verted to its norm – common base standards, but a significant amount
of testing and optimization required to make these work optimally.

Copyright © 2019 Rethink Research, All rights reserved.

The situation has been even more complex when we look beyond the
device-to-network interface and examine the interfaces between differ-
ent elements of the network itself – between the access point and the
core network, or between different components of the RAN.

Here, supposedly standard interfaces have been deployed in semi-

proprietary ways, with each vendor designing slightly different, and
incompatible, variations. Two particularly significant examples in 4G
and 5G are X2 and CPRI.

O P E R A TO R S LO O K F O R W AY S TO P AY P A SS X 2 I N 5 G

The X2 interface, which sup-

ports handover between 4G
base stations – and will also
support handover between 4G
and 5G – is not fully interopera-
ble. “Each vendor tends to im-
plement it slightly differently to
get superior performance on
their own systems,” acknowl-
edged Nokia’s CTO and head
of Bell Labs, Marcus Weldon.

Figure 2. The role of X2 in

4G/5G interworking

Copyright © 2019 Rethink Research, All rights reserved.

The lack of uniformity in X2 is serious in the early years of 5G. As oper-
ators start to deploy 5G New Radio, they will want plenty of interwork-
ing with their LTE networks.

In the first phase, Non-Standalone, they will retain the LTE core, and in
most cases, will only introduce 5G base stations for particular locations
or applications with high demand. So difficulty in enabling those new
base stations to communicate with the existing LTE majority would
clearly be a show-stopper. Even when operators move to the Standalone
mode, and the 5G core, there will need to be close interworking with
LTE for many years to come.

In fact, one of the 11 options for the functional split in a disaggregated

RAN, defined by the 3GPP in Release 15, can support multivendor in-
terworking by avoiding the use of X2. Option 3a, which defines how
dual connectivity works in a 4G/5G, NSA deployment, calls for a direct
S1-U interface from the 4G core to the 5G NR site, and the traffic flow
for each 4G or 5G NR is split at the core. This reduces the processing
demand on the 4G site and works well in multivendor scenarios.

However, many operators are more interested in deploying one of the

other two dual connectivity options for NSA, particularly 3x, which
reduces the need for expensive backhaul and LTE capacity upgrades
because the 5G NR is connected simultaneously to both the EPC and the
LTE cell (via the traffic plane S1-U and the X2 interfaces, respectively).

Vendors and operator groups are looking for other solutions which may
help MNOs to get the best of both worlds. For instance, Nokia has pro-
posed an overlay approach to 4G/5G coexistence based on dynamic
spectrum sharing (DSS), a new standardized feature which is just ap-
pearing in some radios, and allows the existing LTE to communicate
with the new base stations without X2. With DSS, spectrum can be
reallocated to different radios every millisecond.

The DSS technology allows an MNO to move spectrum between differ-

ent radio technologies as required, and was designed mainly to give
greater flexibility and efficiency in spectrum usage, in high bandwidth
situations. An MNO could buy 5G equipment, with DSS support, from
a new vendor and initially run the system in 4G mode, using some ex-

10

Copyright © 2019 Rethink Research, All rights reserved.

isting 4G spectrum. This overlay would support interoperability
between the incumbent 4G and the new 5G systems, with roaming be-
tween the two 4G radios happening at packet core level as usual.

C P RI SH O R T CO M IN G S H A V E D R I V E N P U SH F O R NE W FR O N TH A U L S T A N DA R S

As for CPRI, the disaggregation of the RAN has made it increasingly

common to separate the radio head from the baseband. Initially, this just
entailed moving the baseband unit (BBU) to the base of the tower, with
the remote radio unit (RRU) on top with the antenna. The next step was
for several RRUs to share one centralized BBU, creating a ‘base station
hotel’, usually on a localized basis, but allowing for flexible allocation of
network resources between different cell sites. These initial steps are
evolving into several new trends, which will reach scale in 5G:

· Centralized RAN over a larger area, with more RRUs sharing a

remote baseband.

· Integration of radios and active antennas into a single unit, often

related to deployment of high-order or Massive MIMO arrays (a key
technology for 5G, as it makes high frequency millimeter wave
spectrum more usable).

· Cloud-RAN or virtualized RAN (vRANs), in which the shared base-

band is deployed as virtual network functions (VNFs) in the telco’s
cloud data center, or in future, even a public cloud.

· Hierarchical vRANs, with network functions divided between

remote centralized servers and more local controllers, often under-
pinning a cluster of small cells or microcells.

· SDN (software-defined networking) in the control plane to allow

data paths and allocation of network resources to be programmed
on an agile basis according to traffic pa#erns and the demands of
a particular service.

11

Copyright © 2019 Rethink Research, All rights reserved.

An example of this evolution, from SK Telecom, is shown in Figure 3.

Figure 3. The evolution of SK Telecom’s 4G and 5G RAN Source:

Netmanias

These architectures have many benefits for the economics of cellular net-
works, notably the flexibility to allocate and prioritize precious network
resources on the fly, supporting agile response to changing commercial
requirements.

But to achieve the maximum flexibility, it is essential that operators can

mix and match BBUs, RRUs and small cells from multiple vendors, and
that makes the fronthaul interface the most crucial one to the economics
of 5G and to the potential for an open, WiFi-like ecosystem.

12

Copyright © 2019 Rethink Research, All rights reserved.

E T H E R N E T A L T E RN A T I V E P A V E W AY F O R F U L L O P E N NE S S

As fronthaul has become packet-based, Ethernet alternatives are being

trialed and deployed. These come with higher standards of uniformity
because they are based on such a well-established and broadly adopted
standard, and this approach is moving from relatively small networks
into virtualized macro RANs.

There are two main Ethernet options available – the packet-based imple-
mentation of CPRI, called eCPRI, which supports Ethernet or IP; and
IEEE’s 1914.3 Radio over Ethernet (RoE). The la#er is more fully open,
but the la#er, of course, offers be#er coexistence for companies which
have already deployed some CPRI fronthaul to link remote radio heads.

In some network architectures, where most of the network functions are

centralized, eCPRI will be optimal (as in 3GPP’s and IEEE’s higher
options), but in more distributed architectures, the IEEE technology is
coming into play.

13

Copyright © 2019 Rethink Research, All rights reserved.

Figure 4. Fronthaul options related to different 3GPP/IEEE functional
splits in a disaggregated RAN
But while Ethernet addresses some of the performance and economic
shortcomings of other solutions, it has drawbacks of its own, many of
which are finally being mitigated by a recent IEEE standard called
802.1CM.

The big challenges of Ethernet fronthaul relate to timing and synchroni-

zation. Ethernet was originally not defined for low-latency applications
like fronthaul. It was unable to support synchronization and ji#er re-
quirements for CPRI-to-Ethernet bridging and so was confined to rela-
tively small vRAN designs. And it needed to address the strict latency
requirement – a maximum of 100 microseconds one-way delay between
baseband and radio head, which includes both delays through the fiber
or air (5 microseconds/km for fiber) and any delays caused by intermedi-
ate switches.

Now, the 802.1CM standard addresses time-sensitive networking for

fronthaul over a bridged IEEE 802.3 Ethernet link. It was developed in
collaboration with the CPRI Cooperation, as well as Study Group 15 of
ITU’s Telecommunication Standardization Sector (ITU-T). The la#er con-
tributed its synchronization solutions.

The new standard is critical, because it gives operators the option of an

end-to-end Ethernet architecture for cellular networks, for the first time,
lowering costs by enabling the use of common components.

Despite all the progress with Ethernet and eCPRI, more open implemen-
tations are being sought by operator-driven groups like Telecom Infra
Project (TIP) and ORAN Alliance. One of their first areas of focus, be-
cause it is so critical to the 5G business case, is the fronthaul interface,
with the aim of preventing the new packet-based technologies from
becoming too vendor-specific (see Chapter 5 for details).

14

Copyright © 2019 Rethink Research, All rights reserved.

The economic impact of openness will be seen in all
layers of the network

As operators pursue open interfaces that will avoid the lock-ins of CPRI,
X2 and others, they are also looking towards openness in the broader ar-
chitecture, right down to open source reference designs for base stations,
and even open source processors emerging for base station use cases
from initiatives like RISC-V and the Linux Foundation’s CHIPS Alliance.

The impact of such openness on the cost of deployment and ownership

of a 5G network would be significant. In a detailed analysis, carried out
by Rethink Technology Research based on inputs from over 70 operators,
and our own inhouse model, an open architecture is seen reducing the
deployment cost of a macro base station by over 60% compared to a tra-
ditional platform. If an open architecture is adopted, the cost of a macro
cell will fall by, on average, 51% compared to 2017 costs (see Figure 5).
Even traditional systems will fall because of price competition and com-
moditizing hardware, but only by an estimated 20%.

Figure 5. Projected deployment cost of a 5G macro cell, open vs tradition-

al architecture

15

Copyright © 2019 Rethink Research, All rights reserved.

More broadly, operators of all kinds expect open RAN architectures to
have a significant impact on the wider economics of deploying a 5G net-
work. These impacts will vary between the macro layer, in which opera-
tors will be migrating, overlaying or enhancing traditional networks; and
the small cell layer, which will be largely a greenfield roll-out to support
dense 5G hot zones.

Figure 6 indicates the most important areas in which operators believe an

open platform would transform the economics of the macro network.
These are led by the ability to address the challenges that have arisen in
deployment of a virtualized RAN – while most MNOs believe this will be
the most cost-efficient, agile architecture to support 5G services, the semi-
proprietary nature of the fronthaul interface has made the economics
very una#ractive to date.

This emerged as the most important area of impact, if an open architec-

ture gains scale, followed by broad TCO improvements. However, flexi-
bility in the network is almost as important as sheer cost. Being able to
split network functions flexibly between the central baseband and the cell
site, to support different use cases optimally, is seen to be easier with an
open platform;
as is the agile
allocation of
network re-
sources accord-
ing to the de-
mands of new
applications
and traffic
pa#erns.

Figure 6. Pri-
mary areas of
impact of open
architectures
on the macro
network - per-
centage of op-
erators placing each factor in their top two (respondents contributed all
their drivers in a free answer. They were then asked to prioritize the most
-cited factors).
16

Copyright © 2019 Rethink Research, All rights reserved.

In the small cell layer, the expected impacts of openness are somewhat
different (see Figure 7). The most important is to encourage a broad sup-
ply chain with improved price competition, to help bring small cell net-
works down to the low cost levels that are required if these are to be de-
ployed in very dense configurations.

That relates directly to the second factor, improved scalability via com-
moditized designs, followed by an easier route to the disaggregated, vir-
tualized small cell network; and then improved ability to deploy multi-
vendor systems and to drive towards commoditized hardware.

Figure 7. Primary areas of impact of open architectures on the small cell

network - percentage of operators placing each factor in their top two
(respondents contributed all their drivers in a free answer. They were then
asked to prioritize the most-cited factors).

17

Copyright © 2019 Rethink Research, All rights reserved.

Several open initiatives are starting to make the case for
open RAN

The disappointment with limited openness in 3G and 4G has driven two

types of development, and these are representative of the two paths down
which the whole mobile industry could go. One is to work through official
standards organizations (SDOs) to drive standards that have sufficient speci-
ficity to limit the potential for semi-proprietary variations. In fronthaul,
these are mainly based on Ethernet (see Chapter 3).

The other route is to go through the open source communities, or open

groups such as Telecom Infra Project (which has a dual system offering
either open source or licensing on Frand – fair reasonable and non-
discriminatory – terms). As operators have become impatient with progress
in the SDOs, some of them have become increasingly willing to drive tech-
nology through open source initiatives, which in the past have been entirely
alien to the telecoms industry.

The catalyst for change is not really the 5G radio, but the broader platform
which will surround it, and enable most of its high value use cases. Virtual-
ization of the network – running many of the network’s digital functions on
commoditized servers or switches – opens the ecosystem to new vendors,
and to open source platforms. The addition of software-defined networking
(SDN) and associated orchestration technologies, to manage and optimize
all the virtual network functions automatically and on the fly, further breaks
down the traditional vendor landscape. It is not given that a major network
equipment OEM will also be a strong supplier of SDN controllers.

So operators will have more choices and a different set of supplier relation-
ships, and many are working hard to foster start-ups via their incubators, or
programs like Facebook’s Telecom Infra Project (TIP), in order to shake up
the supply chain.

Even more radically, some operators are looking to define the new
technologies themselves, and seize back control not just from incumbent
suppliers but from the vendor world in general. This goes further than just
increasing R&D to support inhouse technologies. Now operators are looking
to turn those inhouse systems into platforms which other carriers can adopt,
18

Copyright © 2019 Rethink Research, All rights reserved.

and which will, given sufficient momentum, force vendors to support
frameworks which have been defined by the operators themselves, not
by suppliers or even traditional standards bodies.

Management and orchestration (MANO) of the virtualized network has

been a particular focus, because it is such a critical point of control of the
future platform. AT&T, China Mobile, Telefónica and SK Telecom have
all placed internal MANO developments into open source initiatives (the
first two into ONAP, Telefonica into ETSI Open Source MANO, SK into
its own T-MANO project). All are hoping to increase their influence over
how 5G era fixed and mobile platforms will look, by gathering sufficient
industry support to become de facto standards.

Now, operators are turning their a#ention to the biggest single area of
capital and operating expense in their networks, the RAN. Once a virtu-
alized, software-defined platform has been established, there is a strong
chance to disaggregate the RAN architecture and, with it, the supplier
landscape.

AT&T has been a leader here too, contributing inhouse specifications to

the xRAN open source group, which has now merged with the China-
oriented Cloud-RAN Alliance to form the ORAN Alliance. And in Face-
book TIP (which is not entirely open source), leading lights include Brit-
ish Telecom, Deutsche Telekom, Orange and SK Telecom.

Projects within these, the leading open groups focused on the RAN, in-
clude those to define open fronthaul and small cell interfaces; open
source reference designs for RAN hardware; and interoperable virtual
network functions (VNFs) for the multivendor RAN.

Open source groups are an increasingly common way for start-ups to get
their innovations noticed by large operators and other partners. Figure 8
summarizes some of the challenger vendors in the RAN (some start-ups,
some established like Airspan), and their key relationships with any top
30 MNOs.

19

Copyright © 2019 Rethink Research, All rights reserved.

Figure 8. Selected challengers which are active in the open RAN, and
their key relationships with top 30 MNOs.

OPEN ALLIANCE

The ORAN Alliance was formed in 2018 with the merger of the xRAN
Forum, which was an open source initiative set up to house technology
developed by AT&T, with the China-oriented Cloud-RAN Alliance.

The new Alliance was founded by AT&T, China Mobile, Deutsche Tele-
kom, NTT Docomo and Orange, and they were later joined by Bharti Air-
tel, China Telecom, KT, Singtel, SK Telecom, Telefonica and Telstra. The
vendors have clearly taken note of this prestigious list of operators and
are signing up. Nokia and Samsung contributed code to the first xRAN
release, while Ericsson recently joined ORAN too.

In April 2018, the first xRAN specifications were published, addressing

the fronthaul link in a disaggregated RAN architecture (see Figure 9).
In xRAN, Verizon chaired a fronthaul working group and contributed to
20

Copyright © 2019 Rethink Research, All rights reserved.

the xRAN Fronthaul Specification Version 1.0, as did two of its 5G sup-
pliers, Nokia and Samsung. Those initial xRAN specs will feed into the
combined ORAN platform, and now the ORAN Alliance is working on
the first technology under its new brand, with AT&T and Nokia taking
the lead.

The two companies are jointly developing an open source RAN Intelli-
gent Controller (RIC) which will align with the broader ORAN platform
as that evolves. The RIC will incorporate a set of functions and interfaces
that support advanced RAN optimization through policy-driven, closed-
loop automation. The software will allow services to be deployed more
quickly and flexibly because it will make the RAN more programmable.
Most importantly, the RIC will be one enabler of the operators’ dearest
dream, an open multivendor ecosystem of interoperable RAN elements.

Figure 9. Conceptual architecture of xRAN (now ORAN). Source: xRAN

white paper 2016

21

Copyright © 2019 Rethink Research, All rights reserved.

It will achieve this by taking a very modular approach, harnessing key
tools from the open virtualized network community. So the RIC will take
the form of an extensible real time microservices framework, in which
interoperable components and virtual functions, such as the radio infor-
mation database and open control plane interfaces, can be combined as
required. These will support use cases such as mobility management,
management, load balancing, radio resource control and RAN slicing.

The platform will also support interfaces to third party applications to

add more mobility functions such as cross-layer optimization and ma-
chine learning.

“We’re delighted to work with Nokia on co-creating an open source im-

plementation of the RIC platform to accelerate innovations and interoper-
ability in the RAN,” said Mazin Gilbert, VP of advanced technology and
systems at AT&T Labs. “We continue to look for
“It is going to break the vendor and
opportunities to drive open platforms and open
technology lock-in … and open up
interfaces in the community.”
a whole new level of interoperabil-
Andre Fuetsch, CTO and president of AT&T
ity. … Traditional areas of the net-
Labs, said: “Equipment that supports open speci- work that have been controlled by
fications from xRAN (and ORAN in the future), a few are now being opened up to
combined with increasing RAN virtualization many.”
and data-driven intelligence, will allow carriers to
reduce complexity, innovate more quickly and Andre Fuetsch, CTO and president
significantly reduce deployment and operational AT&T Labs
costs.” He was very clear about the end goal for
all this sharing of code, saying: “It is going to break the vendor and tech-
nology lock-in … and open up a whole new level of interoperability. …
Traditional areas of the network that have been controlled by a few are
now being opened up to many. I think it’s fair to say that all the ORAN
operators are highly motivated to drive more innovation and interopera-
bility, and thus competition into their radio access networks.”

22

Copyright © 2019 Rethink Research, All rights reserved.

O R A N M AK E S P RO G R E S S WI TH P O C T E S T S

In Mobile World Congress 2019, ORAN Alliance members showcased six

proof of concept (PoC) platforms and promised imminent release of the
body’s first would-be standard since its formation – the Open Fronthaul
Specifications, which comprise control, user, synchronization and man-
agement plane protocols. The six PoCs were arranged around four
themes – intelligent RAN control, open interfaces, virtualization and
white box.

23

Copyright © 2019 Rethink Research, All rights reserved.

A long line of operators pledged support, some saying they would intro-
duce the specs to commercial networks from late 2019 onwards, and
were already starting on multivendor interoperability testing. Those op-
erators include KDDI, AT&T, Bharti Airtel, China Mobile, China Tele-
com, Deutsche Telekom, KT, NTT Docomo, Orange, SK Telecom, Soft-
bank, Telefónica, TIM and Verizon.

Among the vendors promising support are Cisco, Fujitsu, HFR Net-
works, Intel, Keysight Technologies, Mavenir, NEC, Nokia, Pivotal
Commware, Radisys, Samsung, SOLiD and Viavi (but not Ericsson,
though it has recently joined the Alliance; or Huawei).

The ORAN Alliance has also demonstrated the scale of its ambition by
launching a new collaboration with The Linux Foundation, which hosts
it, to establish a broad open source community dedicated to RAN soft-
ware. To support this, it released its first white paper late in 2018,
‘ORAN: Towards an Open and Smart RAN’. The ORAN Software Com-
munity will provide an application layer for RANs, aligned with the
ORAN architecture, and will encourage the development of an open
source infrastructure platform for 5G.

T E L E C O M I NF R A PR O J E C T

The Telecom Infra Project was established by Facebook to drive general

purpose processor architectures and open interfaces into every area of
the telecoms network, to reduce costs and encourage an open ecosystem
and multivendor, disaggregated platform. It has been heavily backed by
several large operators, mainly from Europe, some of which have set up
TIP Ecosystem Acceleration Centers (TEAC), to incubate the kind of start
-ups which will create the new ecosystem. These include BT, Orange,
Deutsche Telekom and SK Telecom. In all cases, the selected start-ups get
the chance to pitch their products to the TIP community and potentially
to gain venture investment from their sponsoring telco.

The MNOs are engaged in incubation efforts and loud words – Orange
executives have said that some of its supported TIP start-ups, Athonet
and Amarisoft, could replace Ericsson one day. But are these token

24

Copyright © 2019 Rethink Research, All rights reserved.

efforts aimed at browbeating existing (and, despite everything, mainly
trusted) suppliers, or a genuine move to open up the market?

Last autumn, BT was the first to move to a second wave of start-ups, and is
running a competition focused heavily on mobile technology and on intent
-based networking (IDN). The winners will be based in the TEAC at the BT
Innovation Labs in Martlesham, Suffolk and in London’s Tech City.
Another way in which companies can access support and funding is
through TIP’s OpenCellular Grants (OCG) program, which was launched
to fund and accelerate non-traditional and open platforms, particularly
those focused on bringing internet access to underserved rural areas (a key
objective for Facebook, to drive usage of its services, and of its [Link]
initiative).
So far, 11 awards have been made under the OCG program, with success-
ful projects including community networks in Nicaragua, Columbia, Brazil
and South Africa; a humanitarian connectivity project in Nigeria; remote
connectivity solutions in Mozambique, Mexico and the Democratic Repub-
lic of the Congo; and a training center in Thailand.

It launched its OpenRAN Project Group at its TIP Summit 2017 to bring a
more flexible approach to RAN equipment design and deployment.

The group describes itself as “pursuing the development of fully program-

mable RAN solutions based on General Purpose Processing Platforms
(GPPP) and disaggregated software. In parallel, we have been working to
define GPPP-based RAN specifications and test new network tools to en-
sure potential OpenRAN solutions meet our own internal key performance
indicators.”

To achieve this, the project aims to foster an open ecosystem of solutions

and solution components that take advantage of the latest capabilities of
GPPPs, both at a software level and also using programmable offload
mechanisms such as field-programmable gate arrays (FPGA).

Figure 10. The TIP OpenRAN group

25

Copyright © 2019 Rethink Research, All rights reserved.

Other TIP projects of relevance to the RAN are the development of an
open disaggregated cell site gateway router, an essential component of
future 5G deployments, via a group called CANDI.

TIP also recently announced an extension of the OpenRAN project group,

called OpenRAN 5G NR, which will be led by Sprint and Vodafone,
while vendor participants will include small cell and vRAN suppliers
(Airspan, Altiostar, Parallel Wireless), chip providers (ARM, Intel and
Qualcomm) plus Aricent and Tech Mahindra.

The initial focus is one sub-6 GHz 5G NR small cells for outdoor and in-
door use cases, with millimeter wave small cells to follow. The aim is to
define an open reference design to combine open source software and
general purpose hardware, to enable a flexible and modular 5G NR de-
sign.

The project group will also develop modeling techniques and a test
framework for specific applications and deployment scenarios as priori-
tized by operator members, and these will apply to either Standalone or
Non-standalone 5G NR networks.

26

Copyright © 2019 Rethink Research, All rights reserved.

S T EP S TO RE A L I TY F O R T IP

In Mobile World Congress 2019, TIP showed off its interoperability ena-

blers in an end-to-end network demonstration which involved several of
its working groups – OpenRAN, vRAN Fronthaul, Edge Computing and
Applications; and Disaggregated Cell Site Gateway; along with some
specific platforms that have emerged from these groups or from Face-
book’s own R&D, including Cassini, CrowdCell, OpenCellular and new
technologies for radio access, backhaul, core and network management.
The demonstration involved 20 technology providers, and was the first
time TIP’s projects were integrated on a single network.

TIP processes and designs are starting to find their way into operator ac-
tivities. At the TIP Summit in November 2018, Vodafone and Telefónica
named the vendors which complied with the terms of an RFI, based on
TIP specifications, which had been issued some months earlier. All those
on the list were independent or start-up suppliers in the small cell, vRAN
and vEPC fields (see inset).

Now the two operators have started working with some of these suppli-
ers on OpenRAN field trials. These started in Turkey, where TIP said 25
trial sites were set up to carry commercial traffic from Open RAN macro-
cells in the 900 MHz band. The trials are now being expanded to parts of
Africa and Latin America. Vodafone is focused on the former region, via
its South Africa-based joint venture, Vodacom, and working initially
with Mavenir and Parallel Wireless. Vodafone UK also plans to launch a
live field trial using indoor 5G small cells based on TIP specs. Meanwhile,
Telefónica is deploying macro and small Open RAN sites in Peru, Colom-
bia and Argentina for pilots.

Both operators, along with Italy’s TIM, are also working to bring the Dis-
aggregated Cell Site Gateway, to the commercial trial stage, and are cur-
rently engaged in lab tests.

27

Copyright © 2019 Rethink Research, All rights reserved.

When TIP announced the RFIs, it promised that “the vendor technologies
presented in response to these RFIs will be evaluated using clear bench-
marks that will help guide operators towards the selection of a shortlist
of vendors … Only RAN nodes which demonstrate open interoperability
from the perspective of the baseband processing platform, radio hard-
ware, software and business model will be evaluated within this pro-
cess.”

In one way, this looks like a radical open source approach being pulled
back towards telco norms. After all, the RFI is a very traditional process
and has been driven by two very established operators.

Already, TIP has had to introduce a paid licensing process, in addition to

its open source activity, to enable manufacturers of new-style equipment
to participate. It’s one thing to offer software freely and then build a ser-
vices model around that in order to monetize; that becomes a lot harder
when base stations are involved and some TIP start-ups had said they
were unwilling to contribute the inventions which were their key differ-
entiator.

But the insistence on interoperability, and the selection of challenger sup-

pliers, both suggest there is still disruptive potential in TIP. And at least,
the RFI demonstrates that, after just over a year, the OpenRAN group has
achieved concrete requirements which are sufficiently unified and spe-
cific to be put out to the market.

28

Copyright © 2019 Rethink Research, All rights reserved.

P O T E N TI A L CO N VE R G E N C E O F O P EN G R O UP S

The new TIP 5G NR group was keen to play down the notion that TIP is
in competition with other open source initiatives, or conventional stand-
ards bodies, in driving open RAN interfaces. It said it would coordinate
with the ORAN Alliance, 3GPP and others. Indeed, the success of the
open RAN vision is likely to rely on a measure of convergence of the
open source efforts in the near future – unless one of them clearly be-
comes the industry’s de facto standard by sheer weight of adoption.

Deutsche Telekom’s Axel Clauberg, who is also chair of TIP and a lead-
ing light in several other groups, has called for greater collaboration be-
tween different initiatives. He recently hinted at future convergence
when he said that Telefónica and Vodafone, as well as being leaders of
the TIP Open RAN project, also plan to base their work on the ORAN
specifications. That, he intimated, would lead to coordination of efforts,
though not a merger of the groups.

“They have not done any specs from the ground up. They have not built
architecture but are looking at how the architecture specified by the
ORAN Alliance could be built and deployed in a commercial way,” he
said. “In this area, TIP would do solutions integration and prove it is pos-
sible to build such an open RAN in real life.”

Clauberg said: “I wouldn’t say TIP is competition to 3GPP but it is im-

portant to understand the relevance of global standards, because without
them we wouldn’t have interoperability. Sometimes there are global
standards and you need acceleration and a standard doesn’t necessarily
say how you should build something. Opening up the RAN is not cov-
ered by the 3GPP but by the ORAN Alliance work and what we’re driv-
ing here in TIP.”

But he believes open initiatives are vital to accelerate progress, especially

at the start of a technology’s lifecycle, and so to encourage start-ups.

And with many operators belonging to both groups – or, worse for the
progress of open platforms, staying on the fence until the market gets
clearer (see Figure 11) – some signs of convergence, or at least coopera-

29

Copyright © 2019 Rethink Research, All rights reserved.

tion, would help to reduce the number of operators (about one-third)
who have not decided on their support for any open project yet.

Figure 11. Percentage of operators active or interested in each major open

mobile network group

S M A L L C EL L F O RU M

A great deal of work on standard interfaces has been pioneered in the en-
terprise or small cell worlds. Small Cell Forum has worked on open inter-
faces between different elements of the network since its formation over
a decade ago.

In the early days of small cells, the Forum came up with Iuh, an interface
between small cells and controllers or cores, and succeeded in having it
adopted as a 3GPP standard. It went on to enhance the TR-069 and X2
interfaces, working with other standards bodies; and then to create FAPI
and nFAPI. The next step will be to take these into the denser, more open
world of 5G.

30

Copyright © 2019 Rethink Research, All rights reserved.

FAPI (functional application platform interface) provides a set of com-
mon APIs to support interoperability between the 3G, 4G or 5G PHY, and
software elements such as the security coprocessor or scheduler. nFAPI
(network FAPI) extends the concept to virtualized small cell networks
and provides an interface between the remote radio unit (RRU), and a
centralized baseband unit (BBU), on which some or all of the baseband
functions are virtualized.

These two sets of APIs map to the two main approaches to dense 5G de-
ployment. In outdoor HetNets, 3GPP is already defining some of the in-
terfaces to connect RRUs with local or far-away BBUs, but FAPI will still
be important as an internal interface between the chipset and software
layers (a gap in the 3GPP or open source standards efforts). In indoor en-
terprise networks, nFAPI will be essential to support the most common
architecture – a group of cells controlled by a central, virtualized unit.

In other words, SCF is addressing a requirement that is absolutely central

to the economics of 5G, since these, for many operators, rest on the ability
to disaggregate the RAN. To date, deployment of this virtualized RAN
(vRAN) has been impeded partly by the lack of a fully unified fronthaul
interface between the RRU and BBU. SCF has already made considerable
progress on this, in a small cell context, and is also addressing aspects of
the interface which are not a central focus for other initiatives, filling
gaps in work by 3GPP or the open projects.
Densification absolutely requires an open ecosystem to make deployment
of large numbers of cells cost-effective. These interfaces are the enabler of
the innovation and price competition that comes when operators can se-
lect equipment from many suppliers.

As outlined above, the move towards open interfaces has been more ad-
vanced in the small cell layer than the macro network. Several suppliers
already offer architectures in which a number of small cells are clustered
around a centralized, virtualized controller. Once standard interfaces be-
tween the radio and the controller are supported, along with open base-
band virtual network functions (VNFs), the economic argument for den-
sification will be far stronger, and a key disadvantage vis-à-vis WiFi (its
open ecosystem) will be removed.

31

Copyright © 2019 Rethink Research, All rights reserved.

nFAPI took a big step towards larger scale when an open initiative, the
open-nfapi project, was announced to support it in February. Led by Cis-
co, the project has worked to implement the nFAPI disaggregated base
station architecture in a set of libraries and simulators. As Mark Grayson,
distinguished consulting engineer at Cisco, wrote in a blog post, Cisco
was drawn to nFAPI because it supports active RAN sharing, and there-
fore neutral host or multi-operator small cell networks – which will be
essential for many enterprise and vertical industry deployments in the
5G era, which cannot be tied to a single MNO.

“We concluded that the Small Cell Forum’s multivendor nFAPI split ar-
chitecture, together with its neutral host management model, offered a
new approach to active sharing of an LTE RAN based on a multivendor
CU-DU implementation,” wrote Grayson.

T H E O P E N A I R I N TE R F A C E

He highlighted the challenges facing many would-be open standards,

particularly in ge#ing sufficient real world adoption when there are
many open, and even open source, projects to choose from. He contin-
ued: “Even though Cisco has licensed the nFAPI libraries under the per-
missive Apache 2.0 license – enabling them to be integrated by the widest
possible set of stakeholders, including closed source proprietary RAN
products – we do not underestimate the real challenges in ge#ing such
capabilities implemented.”

To help address the issue, Cisco joined the Open Air Interface (OAI) open
source initiative and said it would use its membership to demonstrate
how to integrate the open-nFAPI libraries into an existing LTE RAN pro-
tocol stack; the integration between the lower nFAPI libraries and the
PHY layer, implemented on a software defined radio platform; and the
integration between the upper nFAPI libraries and the MAC and RRC
layers. This will help to lower the barriers to active RAN sharing in small
cell environments.

The mission of the OpenAirInterface Software Alliance (OSA), estab-

lished in 2014, is to provide software and tools for 5G R&D and product
development.

32

Copyright © 2019 Rethink Research, All rights reserved.

It states its aims as bringing open source software, running on general
purposes processors such as x86 or ARM, deep into the RAN, an area
whose demanding requirements have led to continuing dominance by
“proprietary elements that stifle innovation and increase the cost for the
operators to deploy new services/applications”, as the OSA puts it.

With software-defined networking (SDN) and open source technologies

starting to change the landscape in other parts of the network, and in de-
vices, the group believes an open source implementation of a real time
stack (base station, terminal and core) on general purpose processors is
needed.

The OSA provides a framework for intellectual property and financial

contributions to the platform. It provides a standards-compliant imple-
mentation of a subset of 3GPP Release 10 LTE for UE, eNB, MME, HSS,
SGw and PGw on Linux-based computing (x86 or ARM architectures).
The software is freely distributed by the Alliance under the terms of
the OSA license model.

33

Copyright © 2019 Rethink Research, All rights reserved.

Case Studies

If the open approach to the RAN succeeds, it will usher in not only a wid-
er selection of suppliers, but will lower the barriers for alternative opera-
tors too, especially those focused on enterprise, rural or IoT services with
specific requirements that have often proved hard for MNOs to deliver
and monetize on their mainstream networks.

There are many trials, and a few deployments, using open platforms, and
we have selected a few that represent:
· An established operator working with a new vendor.
· A partnership between and an established operator and a large OEM,
which is adapting to a more open approach.
· A new entrant operator, working with a mixture of start-ups and
large vendors in an open network.

T E L E C O M I T A L I A W O RK S WI TH TH RE E N E W V E ND O RS O N V R A N

A new project by Telecom Italia will give hope to the smaller providers.
The company is trialing a virtualized RAN (vRAN), but is not working
with Ericsson and Huawei – which have taken part in previous vRAN
tests and trials with the MNO – but with three smaller vendors, plus inte-
grator Tech Mahindra.

It has also worked in the past with a vRAN specialist, Altiostar, but the
most recent trials involved Baicells on the base station side, plus software
start-up Phluido. Technology for virtualized core and RAN was also pro-
vided by Radisys, a leading light in the open source CORD (Central
Office Re-architected as a Datacenter) group, and recently acquired by
disruptive Indian operator Reliance Jio. Finally, Tech Mahindra will be
the systems integrator, a role it is increasingly taking in projects based on
open technologies – it is also a very active player in CORD, and in Face-
book TIP.

34

Copyright © 2019 Rethink Research, All rights reserved.

“We are looking to a deeper virtualization that will increase the cost
a#ention we are looking for,” said Lucy Lombardi, SVP of technology
ecosystem innovation at TIM. “The scope is to test fronthaul in existing
networks, including microwave and fiber, as well as interoperability be-
tween different vendors. We are in the process of integrating Radisys and
Phluido on one side and producing a radio resource unit with Baicells.”

Phluido is particularly interesting in this group. The four-year-old start-

ups has developed a fronthaul technology that compresses data so that it
can be transported over non-fiber connections such as Ethernet or wire-
less.

Phluido is a leader of the TIP OpenRAN workgroup (see Chapter 5). It

calls its approach to vRAN ‘radio as a service’ and claims its approach
could make it as viable to virtualize the RAN in the cloud as the packet
core. Founded by former Qualcomm engineers Dario Fertonani and Alan
Barbieri (now CEO), it has developed a technique that allows the net-
work to pre-process and compress the frequency data at the cell site, so
that it can be realistically transmi#ed to the data center via a lower per-
formance connection such as an Ethernet cable or microwave. Once there,
the baseband processing would take place, running as an application on
centralized servers.

N O K I A A N D CH I N A MO B I L E

Nokia has been more willing than most of its key rivals to adapt its pro-
cesses to a more open, software-driven market. It has been working with
AT&T to design an open RAN Intelligent Controller (RIC) within the
ORAN Alliance (see Chapter 5G). And it has a similar collaboration with
China Mobile.

It will work with the China Mobile Research Institute (CMRI) to work on
key aspects of an open, interoperable and multivendor RAN, including:

· A decentralized, multivendor architecture in which the central unit

controls several remote distributed units, and most baseband functions
are centralized and virtualized on cloud infrastructure. This architecture

35

Copyright © 2019 Rethink Research, All rights reserved.

will support real time and non-real time processing capabilities, the for-
mer requiring more physical and virtual network functions (VNFs) to be
located at the cell site or edge node.
· Jointly specified open RAN interfaces between multivendor compo-
nents, such as the fronthaul interface connecting the central unit to the
remote radio heads.
· Smart radio intelligence control using advanced network analytics.
· Commercial off-the-shelf (COTS) hardware equipment.

“By optimizing multivendor network interworking we will be able to ac-

cess network intelligence and use end-to-end data analytics and machine
learning to support the delivery of innovative new consumer and indus-
try services in the 5G world,” said Zhiqiang Yang, a VP of CMRI.

R A K U T E N CL A IM S F I R S T CO U D - N A T IV E , M U LT I V E N DO R MO B I L E
N E T WO RK

Japan’s ecommerce company, Rakuten, acquired spectrum last year and

is building a greenfield 4G network to convert its existing MVNO busi-
ness into a full mobile platform, chiefly to drive usage of its digital ser-
vices, such as shopping and banking, and to enable new applications.

As a new entrant, Rakuten is able to take a new approach to its network,

unencumbered by legacy installations and relationships. It plans to sup-
port several key concepts of future 5G networks -vRAN, edge compu-
ting, network slicing, end-to-end automation supported by AI, an IPv6-
based transport network, a unified OSS layer, and a cloud-native packet
core with full control/user plane separation.

Its vision is a cloud-native, end-to-end network driven entirely by digital

services, not bits and bytes; and with a fully multivendor roll-out inte-
grating traditional and disruptive suppliers, held together by open inter-
faces (and a hefty outlay on turnkey services from Nokia and a lab collab-
oration with Tech Mahindra).

36

Copyright © 2019 Rethink Research, All rights reserved.

Other suppliers include Altiostar, Cisco, Intel, Red Hat, OKI, Fujitsu, Ci-
ena, Netcracker, Qualcomm, Mavenir, Quanta, Sercomm, Allot, Innoeye
and Viavi.

Everything is deployed natively on cloud infrastructure. Rakuten is work-

ing with Red Hat to build its NFV cloud, implementing three of the com-
panies’ offerings – the Red Hat Enterprise Linux operating system; Open-
Stack Platform for Infrastructure-as-a-Service (IaaS); and Ceph Storage for
massive software-defined storage functionality.

At the RAN level, the vRAN comes from Altiostar, in which Rakuten has
taken an investment. The supplier’s baseband is fully virtualized on COTS
hardware, in this case from Quanta Computing Technology (QCT) of Tai-
wan, though other vendors’ boxes could be introduced too. The servers
will run on Intel processor/FPGA combinations.

The Quanta boxes are deployed in aggregation centers, which in future

might also house edge nodes, and these centers serve clusters of radio
cells. Rakuten will build 4,000 virtual distributed units and two central da-
ta centers to support the vRAN. This infrastructure will handle baseband
processing and connect to the remote radio heads that Rakuten is sourcing
(Nokia AirScale).

Altiostar makes a high layer functional split between virtualized baseband

functions and those at the cell site, supporting 3GPP’s Option 2. This
means most functions are centralized, while another group pushing more
open networks, the ORAN Alliance, is focused on the lower layer Option
7. Most importantly, though, this RAN points to the long-held MNO
dream of a network in which basebands and radio heads come from differ-
ent suppliers.

There will also be large numbers of small cells – what Rakuten calls ‘lean’
outdoor radio/antenna units with efficient siting requirements – from an
unnamed vendor but running on Qualcomm chips. Qualcomm has also
provided test devices.

The cloud-native core is also a multivendor affair. The main core is coming
from Cisco, but Nokia is also providing elements from its AirGile family,
and is working with Rakuten on several future core functions designed to
harness AI-assisted automation to reduce opex. Another important aspect

37

Copyright © 2019 Rethink Research, All rights reserved.

of opex reduction will be that Rakuten has fewer than 10 SKUs (stock
keeping units) to maintain, because it has opted for standard, uniform in-
frastructure – this compares to hundreds for most operators.

There will be a common orchestration and OSS layer on the telco cloud,
managing all the virtual functions in the RAN, core and transport. The
cloud will include distributed or edge nodes which, combined with con-
trol/user plane separation (CUPS) in the packet core, will enable mobile
edge applications.

Finally, Rakuten will use Netcracker’s end-to-end digital BSS (business

support system).

There are still many barriers in the way to the way of a

fully open RAN

Despite all the efforts ongoing, it is still very early days for a fully open
RAN, especially as operators in many markets are feeling pressure to de-
ploy 5G macrocells more quickly than they had originally planned. In
those circumstances, they are more likely to support tried and tested archi-
tectures and vendors in order to reduce the risk associated with rapid roll-
out of new technologies.

In the operator survey, respondents were asked to name all the barriers
and inhibitors they saw to deploying open RAN. They were then asked to
rank the most-cited examples in terms of seriousness for their business
case.

The most important barriers, according to this study, relate to quality and
cost (see Figure 12). Almost half of the operators (48%) were concerned
that open platforms would not support the same level of quality and ro-
bustness as existing solutions, because they are new, commercially untest-
ed, and have been devised by new groupings with different agendas.

38

Copyright © 2019 Rethink Research, All rights reserved.

Almost as many – 45% - believe that, while open source software may be
free, the cost of implementing open platforms will be high because of the
large amount of systems integration, customization and inhouse testing
that will be needed to turn an open network into something that is fully
carrier-grade. Some operators, such as the UK’s BT, have already warned
that working with open platforms – even well-established ones with major
ecosystems, like OpenStack – incurs significant costs of skills and integra-
tion, at least for the first few years.

Figure 12. Top barriers to adopt open architectures in the RAN – percent-
age of operators placing each barrier in their top three (respondents con-
tributed all their drivers in a free answer. They were then asked to priori-
tize the most-cited factors).

39

Copyright © 2019 Rethink Research, All rights reserved.

There are also concerns about vendors – the dilemma between supporting
new suppliers in order to broaden the ecosystem and increase competition;
and the risk of trusting critical networks to untried partners. This factor
explains why many real world open RAN initiatives are focusing on the
small cell network, which is often greenfield and is less business-critical in
the early stages of 5G. It is also the main reason why systems integrators
see the open network as a major opportunity to take on the risk and effort
of ensuring that new vendors’ platforms can be assured to be fully carrier-
grade.

Other concerns including the potential for longer time to market when sev-
eral different technologies and vendors need to be supported; and a relat-
ed issue of the time and effort required to integrate new open RAN with
existing networks, especially given the limited openness of current genera-
tion interfaces.

Fragmentation is always a significant downside of open source, and over

one-third of operators believe that is a top two barrier – certainly, some
closer cooperation between ORAN Alliance and TIP, and support from
large OEMs, will help to allay the fears, especially as those OEMs still
dominate the investment in R&D in emerging technologies, and are still
placing many of their efforts mainly into conventional standards bodies.

How will open RAN platforms affect the pattern of 5G

deployment?

It is too early in the story of open RAN platforms to make firm forecasts of
how many base stations will be deployed based on open systems com-
pared to conventional technologies. But the progress of the developments
in open architectures; and in-depth interviews with operators, established
OEMs and new vendors, are all helping to make detailed and firm as-
sumptions about the likely scenarios. The market can certainly play out in
several ways, depending on a set of important variables, but the spread of
outcomes is narrowing.

40

Copyright © 2019 Rethink Research, All rights reserved.

The most important variables which will affect the level of uptake of
open RAN platforms are, according to our operator and vendor surveys:

· Speed of deployment of at-scale 5G networks – in markets where

there is some breathing space, there will be more time for operators to
get to grips with open systems.

· Speed and extent of densification, since small cells will often drive
open platforms, and very dense networks will make openness more
important.

· Decisions by top five OEMs – level of support for open initiatives vs

pushing more interoperable approach via groups like CPRI.

· Development of a services ecosystem to reduce cost and risk e.g. inte-

grators which offer pre-configured, pre-tested open components.

· The number of new entrants to the cellular operator space, often ena-
bled by shared spectrum, as greenfield players have fewer barriers to
adopt open platforms.

· Commercial success of early deployments based on open platforms.

· Availability of commercial solutions, especially to enable an ecosys-

tem, and a clear idea of pricing, to stabilize.

· Convergence of overlapping initiatives to reduce fragmentation, and

concentrate funds and operator support on one system.

· Support from a combination of new innovators, to drive more compe-

tition, as well as the big OEMs, to lend credibility and R&D weight.

· In the small cell area, increasing coexistence with WiFi, rather than a
situation in which WiFi 6, the newest standard, which has certain 5G-
like qualities, makes large-scale small cell deployment less necessary.
WiFi 6 will, of course, have access to the huge open ecosystem estab-
lished by its predecessors, complete with a centralized testing and
certification process which is already underway, under the auspices
of the WiFi Alliance.
41

Copyright © 2019 Rethink Research, All rights reserved.

We are predicting the impact of open platforms on the industry’s RAN
shipments between 2020 and 2025, based on four very broad scenarios,
namely:

Open base station designs and interfaces are adopted broadly in the
macro and small cell layers from 2022.

Open platforms have a high impact on the small cell layer, creating a
WiFi-like ecosystem, but have li#le impact on the more mission-
critical macro network before 2025.

Open platforms have limited market adoption for mainstream com-

mercial networks, outside of some greenfield roll-outs.

There is limited uptake of open source base stations and processors

before 2025, but there is broad adoption of the emerging open
fronthaul interfaces which will be supported by conventional net-
work equipment and stimulate the roll-out of vRAN and 5G.

B A S E LI N E F O R E C AS T S F O R B AS E S T AT IO N D EP LO Y M E NT

Before discussing these scenarios, we should outline our baseline fore-

casts for total deployments of macro and micro base stations, and of
small cells, to 2025. These forecasts cover both conventional and open
platforms and represent our median case for open migration.

Figure 13 predicts CAGR of just 2% in deployments and upgrades of 4G

and 5G (or dual-mode) macro and micro base stations from 2018 to 2025,
though growth pa#erns vary by region and by the phase of deployment.

North America and South East Asia will see deployments peak in 2020
and then decline somewhat after that, the former actually declining by
2% in CAGR terms over the whole period despite a strong roll-out in
2018 to 2021. The strongest CAGR is seen in the Middle East and Africa
(MEA) region, while South Asia will experience steady growth until
2024.

The main upturns in roll-outs, in 2020-2021 and 2023-2024, are driven

mainly by 5G deployments, sometimes accompanied by 4G refresh and

42

Copyright © 2019 Rethink Research, All rights reserved.

densification – the first wave in front-running regions like North Ameri-
ca, the second in emerging economies.

Figure 13. Total deployments and upgrades of macro and micro base sta-
tions by region 2018 to 2025, all architectures

As for small cells, Figure 14 summarizes our forecast that the installed
base (cumulative) will rise by 18% CAGR over the period from 2018 to
2025, also driven by a combination of 4G densification and the deploy-
ment of 5G in enterprises, indoor environments and ho+ones. The high-
est growth will be in Caribbean and Latin America (CALA) and in MEA,
though in sheer numbers terms, South East Asia will dominate through-
out the period, despite CAGR of only 11%. By 2025, the global installed
base will have reached about 8.4m.

43

Copyright © 2019 Rethink Research, All rights reserved.

Figure 14. Total installed base of small cells by region 2018 to 2025, all ar-
chitectures

S C E N A R IO 1– L A R G E S C A L E A DO P TIO N F RO M 2 02 2

In the first of our scenarios, large-scale adoption of open base stations in

the macro and small cell layers from 2022, we examined what percentage
of the deployments above would be based on open architectures
(commoditized hardware, open or open source interfaces and software).

Based on the predictions of operators, conventional and new vendors,

and the open bodies themselves, we project a median case of below 5% of
macro and micro base stations deployed in 2020 being open. However,
that will rise quickly in subsequent years, reaching 25% of the total in the
first year of mass roll-out, 2022, and ge#ing to almost two-thirds of the
total by the end of 2025 – representing CAGR of 65% for the whole peri-
od from 2020 to 2025.

In the first couple of years, adoption will be led by North America

(thanks to the push of AT&T and, to a lesser extent, Verizon behind the

44

Copyright © 2019 Rethink Research, All rights reserved.

open technologies); and by South Asia, where the weight of India will be
key - its operators, especially Reliance Jio, have already commi#ed
strongly to open source architectures in their quest to reduce cost of own-
ership and assert leadership over their own ecosystem.

RJio has even acquired Radisys, a software provider and integrator

which has been a leading light in the CORD (Central Office Re-
architected as a Datacenter) open source network initiative. By 2025, 70%
of South Asia macro and micro deployments will be open, and 67% in
South East Asia, including China (Figure 15).

Figure 15. In Scenario 1, deployments of open macro and micro base sta-
tions by region 2020 to 2025

45

Copyright © 2019 Rethink Research, All rights reserved.

S C E N A R IO 2– A T - S C A L E D EP LO Y ME NT L IM I T E D TO SMA L L C E L L
LAYER

In Scenario 2, in which adoption of open platforms takes place mainly in

the small cell layer, the deployments that do happen in macro and micro
categories will be far more heavily driven by Asia. Overall CAGR will be
50%, from a lower starting point, and by 2025 the global number of open
deployments will be less than half that in Scenario 1.

In the two Asia regions, the number of open base stations in Scenario 2
will be a li#le over half of those in Scenario 1 and Asia will account for
70% of the (smaller) total of open sites deployed in 2025.

Figure 16. In Scenario 2, deployments of open macro and micro base sta-
tions 2020-2025

By contrast, the installed base of open small cells will be similar in the
median case for both scenarios. The roll-out of open small cells will not
start in earnest until 2021, when 8.7% of the base will be based on open or
open source hardware and interfaces (up from 1.8% in 2020). South Asia,
especially India, will be the biggest driver by a long way in the early
46

Copyright © 2019 Rethink Research, All rights reserved.

years – 15% of its installed base will be open in 2021, compared to 9% in
North America and South East Asia (including China) and under 5%
elsewhere. By 2025, the proportions will be more consistent in different
regions though the two Asia regions will have the highest open base, at
almost 70% (see Figure 17).

The CAGR between 2020 and 2025 will be over 100% in all regions, and
128% globally, but the highest annual growth rates will be in MEA
(184%) and Europe (169%).

Figure 17. In Scenario 2, installed base of open small cells 2020-2025

One of the features of the Open RAN landscape is the rising presence of
non-MNO deployers of small cells. These may be enabled by shared
spectrum, or airwaves allocated specifically to a particular industry, or in
some cases, MNOs will provide their own spectrum for a neutral host,
private network operator or an enterprise ‘heavy MVNO’ (an MVNO
which deploys some of its own equipment). Most commonly, these new
deployers will build ‘sub-nets’ – localized networks optimized for a par-

47

Copyright © 2019 Rethink Research, All rights reserved.

ticular enterprise or location, often integrated with their own packet core
and edge computing. They will be important for extending high quality
cellular connectivity to industrial sectors which have often been difficult
for MNOs to target profitably.

The barriers to the new providers will be lower if there is access to afford-
able equipment from a variety of sources, so although we see enterprise
and private networks as being an increasingly significant part of the cel-
lular map in any scenario, they will grow more quickly if the same is true
of open platforms. The percentage of open small cells deployed by non-
MNOs (see Figure 17) will grow from 25% in 2020 – a community which
will be important in kickstarting the roll-out of the open systems – to al-
most 40% in 2025.

Figure 18. Deployment of open small cells by operator type 2020-2025

Probably the primary reason why alternative operators will grow in the
small cell and enterprise environment is that conventional MNOs contin-
ue to find it hard to make a strong business case for funding networks
that penetrate deep indoors, but these are necessary to support high qual-
ity enterprise connectivity and to enable many of the emerging industrial
and IoT applications which are expected to be made possible, or en-
hanced, by 5G.
48

Copyright © 2019 Rethink Research, All rights reserved.

The deployment of open small cells will be dominated by indoor systems
(Figure 19), though there will be increasing usage of these new architec-
tures, by MNOs and others, to lower the cost of densification in city cen-
ters, and for rural connectivity, based on platforms such as the Telecom
Infra Project’s OpenCellular. In 2020, 76% of open small cells, and over
80% of 5G models, will be deployed indoors, and the percentage will be
67% by 2025.

Figure 19. Deployment of open small cells, indoor vs outdoor 2020-2025

49

Copyright © 2019 Rethink Research, All rights reserved.

S C E N A R IO 3– L IM IT E D D EP LO Y ME N T B EF O R E 20 25

In Scenario 3, we assume it takes longer than current advocates are hop-

ing, for open architectures to make a significant impact on the RAN. In
this scenario, the variables listed at the start of Chapter 8 will all tend to
the negative. This does not mean open architectures will never be widely
adopted – as the 5G era progresses and networks need to support ex-
tremely large numbers of cells and devices, we believe openness to be in-
evitable. But it is possible, in the worst case, that there will be li#le adop-
tion before the middle of the decade.

One of the effects of this delay would be a lower deployment of small

cells overall. Our baseline forecast for total small cell installed base
(Figure 20) assumes a level of open systems adoption at the level of Sce-
nario 2. If a delay until 2025 or later is seen, a portion of the small cell de-
ployments will not happen at all, because they will not be cost-effective
without access to open platforms; or certain deployers will be deterred
from investing; or operators will use WiFi instead.

Figure 20. In Scenario 3, total installed base of small cells, all architectures
2018-2025

50

Copyright © 2019 Rethink Research, All rights reserved.

 We calculate that the shortfall in the cumulative total of all small cells in-
stalled by 2025 could be over four million units (including all small cell
form factors including radio heads in distributed systems). These will ei-
ther result in some locations or use cases staying neglected, or cellular
being displaced by WiFi. The global CAGR in the small cell installed base
between 2020 and 2025 will be about 6%, compared to 128% in the posi-
tive scenario (Figure 20).

S C E N A R IO 4– L IM IT E D O P E N R A N B UT A G R E E D TO O P EN F R O N T H A UL I NTE R F A C E

In the final scenario, there would be limited uptake of full open source or
open base station platforms – commoditized hardware, general purpose
processors, open source software – before 2025.

However, we believe there is one element of the RAN where openness

will be essential for the cost-effective adoption of 5G architectures, espe-
cially vRAN. That is the interface between radio heads and baseband
units in a virtualized, disaggregated RAN. This is the initial focus of the
two main open RAN initiatives and the operators’ urgent need for a com-
mon, fully interoperable interface in this area should drive adoption of
one of the options, or a convergence of the two, running on top of Ether-
net fronthaul.

There will be systems which use IEEE protocols over Ethernet, and others
that stay with the CPRI roadmap, but we expect a new open Ethernet
protocol, from TIP and/or ORAN, to take a major place in the market too,
starting with greenfield deployments and localized networks such as city
systems, and then moving into the main macro RAN as 5G is expanded.

So in Scenario 4, despite the slow pace of a full open architecture, there

would be broad adoption of the open RAN interfaces (Figure 21). In the
period from 2019 to 2025, there will be a 30% decline in the CAGR for
proprietary options, while IEEE Ethernet solutions will grow by 40%, and
the open source alternative by 126%. By 2025, these open protocols will
account for about two-thirds of fronthaul deployments.

51

Copyright © 2019 Rethink Research, All rights reserved.

Figure 21. In Scenario 4, fronthaul and midhaul links deployed, by archi-
tecture 2019-2025

52

Copyright © 2019 Rethink Research, All rights reserved.

Conclusions

A fully open RAN platform is in its early stages and there are many ob-
stacles in its path, not least the entrenched position and heavy R&D in-
vestment of the major OEMs, and the risks many operators see in an im-
mature and potentially fragmented solution. However, the economics of
driving 5G to deliver new revenues - which will entail pushing it to
many new locations and devices – will steadily make a more competitive,
cost-effective and multivendor approach impossible to avoid.

The big question is how quickly it will become mainstream, and we have
modelled scenarios in which adoption starts in 2020, and others in which
there is limited deployment until 2025 or later. In either case, much of the
momentum will come from new classes of operators and vendors, ad-
dressing underserved areas including the indoor enterprise and industri-
al IoT.

These greenfield initiatives will have the opportunity to build a more

open, WiFi-like ecosystem based on the open specifications which are be-
ing driven aggressively by major operators, via organizations like Face-
book TIP and the Linux Foundation-hosted ORAN Alliance.

53

Copyright © 2019 Rethink Research, All rights reserved.

Appendix: Methodology

This document contains explanatory notes and commentary to accompa-

ny the Excel spreadsheet ‘Rethink Research Open RAN Module 2018-
2025’. The surveys and forecasts on which the outputs are based were
conducted in January to March 2019.

The wireless forecasts included in this report are based on research on

the top 40 international mobile operator groups, which account for 80%
of the global mobile subscribers (IMG-40).

These forecasts were based on a combination of data from:

· Detailed surveys, interviews and operator-by-operator modeling of
the IMG-40 groups.
· Studies and modelling of the deployments and strategies of the top
100 4G/5G operators, as tracked by Rethink Technology Research’s
quarterly surveys, interviews and desk research.
· A survey of 76 tier one operators about their detailed plans for RAN
deployments to 2025, including new architectures.
· Input from ecosystem vendors on shipments, technology strategies
and competitive landscape, also updated quarterly.
· Surveys of alternative cellular operators, including over 45 private
cellular, neutral host, enterprise and carrier WiFi providers.
· A calculation of the resources required in each type of location to
achieve the MNOs’ stated objectives.
· A detailed, proprietary model to project best and worst cases, and
different scenarios, according to all these inputs and a detailed list of
assumptions and variables, in turn drawn from a research base com-
piled over 15 years, and supported by over a decade of historic data.

From the starting point of a calculation of the number of cell sites already
deployed worldwide, forecasts were made of the numbers of base sta-
tions and edge nodes that would be rolled out a) to brand new sites and
b) to replace or upgrade existing sites. These deployment forecasts were
then categorized by network topology, spectrum band, spectrum mode,
region and other factors. The equipment deployed in each case was also
surveyed and modeled.

54

Copyright © 2019 Rethink Research, All rights reserved.

Based on the surveys of operators and vendors, it was then calculated
how the cell and edge sites would be equipped – by base station type,
edge node type and location, technology, frequency band etc., leading to
a detailed unit and market size measurement.

The Rethink RAN Research process summarized

Most of the forecasts refer specifically to nodes deployed within MNO

networks, either by themselves or by partners. MNOs may also make use
of third party nodes deployed outside their network and connected by
wireline or WiFi; and many edge cloud services will run on infrastructure
that is not used by MNOs at all.

55

Copyright © 2019 Rethink Research, All rights reserved.

RAN Research: Forecasting disruption in wireless

Rethink Technology Research is a specialized research and consulting firm

with 12 years’ experience in surveying wireless, broadband, over-the-top
and quad play operators. This has resulted in a broad research base of over
140 service providers (MNOs, telcos, cable and satellite operators, over-the-
top providers) worldwide. These organizations are surveyed on a regular
basis about their network infrastructure and business plans, and have a re-
lationship of trust with Rethink.

Rethink also has deep relationships with the telecoms ecosystem (tier one
device OEMs, vendors, technology developers, integrators, regulators etc),
and is perceived as a thought leader in many areas of the telecoms and me-
dia sectors. Key areas of expertise and research experience include HetNet
migration, small cells and carrier WiFi; transformation strategies for the
RAN and the BSS/OSS; convergence of IT and network skills and platforms;
device and chipset roadmaps; spectrum strategy.

Here are some sample titles of reports we have produced recently:

· Will 5G really be so different after all? 5G Deployment Trends 2018-2025
· MNOs look to MEC to deliver 5G class apps: Edge Computing Forecast
2018 to 2025
· Virtualized RAN rollouts stu#er – awaiting vendor interoperability,
standards

S U B S C R IP TIO N CO S T S

Single User license - $2,000 (One Report)

Single User subscription (This report plus 5 more) $6,000
Corporate license—$4,000 (unlimited distribution inside your organization).
Corporate Subscription (This report plus 5 more) $10,000.

56

Copyright © 2019 Rethink Research, All rights reserved.

 R A N R E S E A R CH MA I N CO N T R IB U TO R S

Caroline Gabriel - Research Director

caroline@[Link]

+44 (0)207 450 1230

R E T H I NK L E A D E RS H IP

Peter White - Co-founder and CEO

peter@[Link]

+44 (0)117 925 7019

Caroline Gabriel - Research Director and Co-founder

caroline@[Link]

+44 (0)207 450 1230

[Link]

57

Copyright © 2019 Rethink Research, All rights reserved.

About Rethink Technology Research

Rethink is a thought leader in quadruple play and emerging wireless and

IoT technologies. It offers consulting, advisory services, research papers,
plus three weekly research services; Wireless Watch, a major influence
among wireless operators and equipment makers; Faultline, which tracks
disruption in the video ecosystem, and OTT video. Riot on enterprise dis-
ruption from the combination of AI/IoT and cloud.

Need more information?

Natalia Szczepanek (Client Relations and Marketing Manager)

natalia@[Link]

+44 (0)117 925 7019

58

Copyright © 2019 Rethink Research, All rights reserved.

 Bristol & Exeter House
Lower Approach Road
Temple Meads
Bristol
BS1 6QS
United Kingdom

Tel. +44 (0) 1173 291480

Tel. +44 (0) 1179 257019

[Link]

Published May 2019

59

Copyright © 2019 Rethink Research, All rights reserved.