Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for

or 5G Transport

2 3
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

Contents

1 Requirements and Vision.............................................................................. 1

1.1 Driving Force................................................................................................. 2

1.1.1 Wireless Services....................................................................................2
1.1.2 Residential Services................................................................................3
1.1.3 Enterprise Services..................................................................................4
1.1.4 DCI Services.............................................................................................4
1.2 Key Requirements for new Transport Networks........................................... 5
1.3 SPN Vision..................................................................................................... 6

2 Principles and Architecture........................................................................... 7

2.1 SPN Design Principles.................................................................................... 8

2.2 Technical Architecture................................................................................... 9

3 Key Technologies......................................................................................... 13

3.1 Slicing Ethernet........................................................................................... 14

3.1.1 Slicing Ethernet Requirements.............................................................14
3.1.2 Slicing Ethernet Technology..................................................................15
3.2 Efficient High-Bandwidth Technology.......................................................... 17
3.2.1 Efficient High-Bandwidth Requirements..............................................17
3.2.2 Efficient High-Bandwidth Technology...................................................17
3.3 Deterministic Low Latency Technology....................................................... 19
3.3.1 Slicing Packet Deterministic Low Latency Technology........................19
3.3.2 Slicing Ethernet Deterministic Low Latency Technology......................20

4 5
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

3.4 Flexible and Reliable Connection Technology............................................. 22

3.4.1 Flexible Connection Requirements.......................................................22
3.4.2 Flexible Connection Technology...........................................................22
3.5 Ultra-High Precision Synchronization Technology....................................... 23
3.5.1 High-Precision Time Synchronization Requirements...........................23
3.5.2 SPN Time Synchronization Network Model..........................................24
3.5.3 Ultra-High Precision Synchronization Technology for the SPN............25
3.6 Centralized Management and Control Technology of SDN......................... 27
3.6.1 Requirements........................................................................................27
3.6.2 Key Technologies for Intelligent Operations........................................28

1
3.6.3 Architecture of the Intelligent Operating System................................30

4 Application Scenarios.................................................................................. 33
Requirements and Vision
4.1 Mobile Service Transport........................................................................... 34
4.1.1 Fronthaul Network...............................................................................34
4.1.2 Midhaul/Backhaul Network.................................................................35
4.2 Data Center Interconnection....................................................................... 35
4.2.1 Multi-point Interconnection Between Edge DCs over the SPN...........36
4.2.2 Point-to-Point Interconnection Between DCs over the SPN...............36
4.3 Video Everywhere....................................................................................... 37
4.4 Cloud Private Line....................................................................................... 38
4.4.1 Cloud Connection.................................................................................38
4.4.2 Cloud Application.................................................................................39

5 Summary and Prospect............................................................................... 41

6 1
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

As 4G has profoundly changed people’s life, 5G will further bring fundamental Lower latency and higher reliability
changes to the society. It is expected that a fully mobile and connected world will Compared with 4G, 5G should have 5 to 10 times of decrease in the latency
be performed. In particular, networks are and will always be the cornerstone with from the terminal device to the core network. Moreover, for the services relat-
the rapid development of the information era. The 5G transport network should ed to human life and property security, the E2E reliability should be increased to
be a part of a converged multi-service network that supports wireless services, res- 99.999%.
idential services, enterprise services, and Data Center Interconnect (DCI) services. Improved spectral efficiency
According to ITU, the minimum value of the mean spectral efficiency in out-
door for IMT-A should be 2 bps/Hz to 3 bps/Hz. To resolve the shortage of spectral
1.1 Driving Force resources caused by explosive growth on traffic, the mean spectral efficiency of 5G
should be 5 to 10 times higher than that of 4G.

1.1.1 Wireless Services Figure 1-1: 5G system requiring new transport networks

3GPP has defined three typical scenarios of 5G, namely eMBB (enhanced Mobile
Broadband), URLLC (Ultra-high Reliability, Low Latency Communication) and mMTC
(massive Machine Type of Communication). These new applications will bring
great challenges to the transport network as follows:
Traffic rate will grow 1000 times
It is estimated that the total global mobile data traffic will be up to 1000 times
from 2010 to 2020. The corresponding throughput will reach 100 Gbit/s per square
kilometer, which is also required to grow 1000 times.
Number of connections will increase 100 times
With the rapid development of the Internet of Things (IoT), the total number
of devices connected by the mobile network will be between 50 and 100 billion
by 2020. Thus, the connection density will also be increased significantly. In sever-
al cases, the connection density of 5G mobile networks will reach 1 million/km2,
which is 100 times larger than that of 4G networks.
10 Gbit/s peak data rate
Nowdays, the peak data rate of 4G networks is around 1 Gbit/s. In the 5G era, 1.1.2 Residential Services
this rate will be improved to 10 Gbit/s to meet the user requirements.
10 Mbit/s user experienced data rate With the rapid development of new services, such as 4K/8K video, virtual reality
In 2020, the user experienced data rate will be larger than 10 Mbit/s in the (VR)/augmented reality (AR), IoT, Big Data, Industry 4.0, and artificial intelligence
vast majority of cases. Furthermore, the 5G network could provides 100 Mbit/s (AI), etc., existing residential broadband is faced with enormous challenges. As
data rate for several high-priority services, such as HD medical image transmission people’s desire for better experience is never-ending, it will further drive up the
in emergency vehicles. quality of residential broadband. However, there is still an obvious performance

2 3
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

difference between the user-experienced bandwidth and the subscribed band- 1.2 Key Requirements for New Transport Networks
width, especially in a Wi-Fi network. To pursuit better experience, video service
will have more strict requirements on transmission rate and latency. In addition,
carriers have spent so much time on service provisioning and failure processing, The advent of the digital society will exert profound influence on the communica-
resulting in degradation of user experience. Cloud-based strategies could be intro- tions infrastructure network, and the traffic model of the existing communications
duced to improve the management efficiency. network will change dramatically. The connection density and traffic density will
increase at an unprecedented rate and the geographical network coverage will be
greatly extended. Users of different types pose personalized requirements on net-
works, and the requirements for real-time interaction of specific industries are ex-
1.1.3 Enterprise Services
tremely high. Therefore, the next-generation transport network architecture must
meet the following requirements:
For government and enterprise customers, carriers provide broadband access and High bandwidth: The transport network needs to provide low-cost and
lease line services among different nodes. The rapid development of the IT indus- high-bandwidth capabilities. The throughput capability should be larger than 1
try will exhaust the bandwidth of the existing enterprise private lines in the near Gbit/s per user, and the network should support the continuous development of
future and promote continuous demand on high-speed private lines. For different video, holographic, and VR applications.
scenarios, there are three main types of private lines: IDC interconnection private Low latency: It should support an E2E latency better than 1 ms to meet the
line, high-value government and enterprise private line, and small and medium stringent requirements of interactive experience and industrial control.
enterprise private line. Cloud services are benefit to enhance the management Flexible connection: The total number of connections has increased by sever-
capability of private lines. Meanwhile, the clouded MEC will be further allocated al dozen times, and the new network should satisfy Full-Mesh data connection in
toward the edge of the network which will bring tremendous challenges to the the huge network with more than 10,000 nodes.
transport network. Simple bandwidth leasing cannot satisfy the requirements of Network openness: It should support standardized southbound interface and
large enterprises. Compared with the traditional private line service, virtual net- northbound interface. In this manner they have the potential of leveraging the ver-
work leasing may be a promising way in the future. satility and openness of the SDN control plane for implementation of the service
control they are generating in either a virtual or a physical deployment or both.
Network slicing: Diversified services bring about diversified network require-
1.1.4 DCI Services ments. In the future, the transport network should be equipped with dynamic
network slicing capabilities to meet diversified service requirements. The network
should provide the capability of both hard isolation and soft isolation.
The DC-based network architecture will change the traditional traffic model of the High reliability: It should provide ultra-high reliability connections for new
transport network. The transport network should support more flexible flows and services such as AR, industrial control, and telemedicine.
diversified reliability. It is noted that the SDN technology is the key part of DCI ser- Intelligent O&M: The service model changes fundamentally. Network slicing
vices. The controller should achieve the WAN level interconnection between DCs and Full Mesh traffic require a more intelligent network O&M system to reduce
and the orchestrator should realize resource scheduling across DCs. The SDN tech- the OPEX.
nology also supports forwarding and control of element separation and centralized
routing computation.

4 5
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

1.3 SPN Vision

Slicing packet network (SPN) focuses on building an efficient, simplified, and ul-
tra-broadband transport network to support the deferent services on the metro
network.
New architecture: The new technology architecture provides a low-cost and
simplified transport network. The bandwidth of the network is 100 times higher,
and the cost per a single bit is 10 to 100 times less.
New services: SPN focuses on the support for new services on the transport
network. The latency is 10 to 100 times less, and the number of service connec-

2
tions is 100 times greater.

Principles and Architecture

New operations: The brand new O&M platform provides agile service deploy-
ment and operation capabilities. The OPEX is 10 times lower.
The SPN is positioned as a next-generation converged transport network
based on the Ethernet ecosystem to implement comprehensive service transport
with high bandwidth, low latency, and high efficiency.
The SPN carries the following services: wireless services, enterprise services,
residential services, and cloud interconnection services.

6 7
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

2.1 SPN Design Principles connections, and services on the network, which implements all-round
network reliability and supports high-reliability service transport.
 High-precision synchronization: In-band clock and time synchroniza-
The SPN takes advantage of efficient Ethernet ecosystem to provide low-cost and tion-based transmission is implemented, with high reliability, high preci-
high-bandwidth transport network services. With the efficient convergence of sion, and high efficiency.
multi-layer network technologies, flexible soft isolation slicing and hard pipe slicing  Flexible service scheduling: The SPN uses flexible tunneling, addressing,
are implemented, and multi-layer service transport capabilities from L0 to L3 are and forwarding techniques to flexibly schedule P2P, P2MP, and MP2MP
all provided. SDN centralizes management and control to innovatively implement services. Slicing Ethernet addressing and forwarding are used to imple-
open, agile, and efficient network operations. SPN principles are as follows: ment Layer 1 service scheduling. MAC and MPLS addressing and forward-
 Packet friendly: Drawing on Ethernet technology (IEEE 802.3 Ethernet, ing are used to implement Layer 2 service scheduling. IP addressing and
OIF FlexE, and innovative slicing Ethernet), the SPN shares the IP/Ethernet forwarding are used to implement Layer 3 service scheduling
ecosystem (optical modules, protocols and chipsets), and support main-
stream packet clients friendly. Figure 2-1: SPN network architecturee
 Multi-layer network technology convergence: With the efficient conver-
gence of IP, Ethernet, and optical technologies, hierarchical networking at
L0 through L3 can be implemented to allow for the construction of various
types of pipes. Ethernet packet scheduling is used for flexible connections
of packet services. Innovative slicing Ethernet data unit stream scheduling
is used to support hard pipe isolation and bandwidth guarantee for ser-
vices and provide low-latency service transport network channels. Opti-
cal-layer wavelength grooming is used to support smooth capacity expan-
sion and large-granularity service grooming.
 Highly efficient soft and hard slicing: Both highly reliable hard slicing and
elastic scalable soft slicing capabilities are provided. Such capabilities iso-
late resources of a physical network to run multiple virtual networks and
provide differentiated SLA-based transport network services for multiple
types of services.
 SDN centralized management and control: SDN helps implement open,
agile, and efficient network operations and maintenance. Service provi-
sioning and O&M are automated. SPN can monitor network status and 2.2 Technical Architecture
trigger network self-optimization in real time. In addition, the architecture
with converged SDN-based management and control provides the other
capabilities, such as simplified network protocols and open networks, as Based on the ITU-T network model, the SPN uses Ethernet as the basic technology
well as cross-network domain and cross-technology service coordination. and supports the integrated transport of IP, Ethernet, and constant bit rate (CBR)
services. The SPN technology architecture consists of the slicing packet layer (SPL),
 Carrier-grade reliability: Network-level hierarchical OAM and protection
capabilities are supported. OAM is used to monitor logical layers, network slicing channel layer (SCL), slicing transport layer (STL), time and clock synchroniza-

8 9
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

tion module, and management and control function module. provide E2E Ethernet-based virtual network connections, and establish
low-latency and hard isolated slicing channels for multi-service transport
Figure 2-2: Layered SPN network model at L1. Drawing on OAM and protection functions for slicing Ethernet chan-
nels, the SCL implements E2E performance monitoring and failure recov-
ery.
 STL: uses IEEE 802.3 Ethernet physical layer technology and OIF FlexE tech-
nology to implement efficient high-bandwidth transmission. The Ethernet
physical layer is composed of new high-speed Ethernet interfaces, such as
50GE, 100GE, 200GE, and 400GE. By virtue of the Ethernet industry chain,
low-cost and high-bandwidth network construction is implemented, and
mainstream networking applications with up to 80 km between hops are
supported. For applications that require higher bandwidth scalability and
longer transmission distance, the SPN uses the Ethernet+DWDM combi-
nation to implement networking of the 10T-scale capacity and hundreds
of kilometers of long distances.

 SPL: By implementing addressing, forwarding, and transport channel en-

capsulation for IP, Ethernet, and CBR services, the SPL provides various
types of services, such as Layer 2 virtual private network (L2VPN), L3VPN,
and CBR transparent transport. The SPL performs service mapping using
multiple addressing mechanisms, such as IP, Multiprotocol Label Switch-
ing (MPLS), 802.1q, and physical ports, and provides service identification,
traffic engineering, and quality of service (QoS) guarantee. For packet
services, the SPL provides the segment routing-transport profile (SR-TP)
tunnels and provides both connection-oriented and connectionless trans-
port channels. The source-route-based segment routing technique allows
the ingress to use a set of segments (MPLS labels) to identify each tun-
nel forwarding path. Unlike conventional tunneling technologies, transit
nodes do not need to maintain path status information for segment rout-
ing tunnels, which allows for flexible path adjustment and network pro-
grammability. SR-TP tunneling technology enhances O&M capabilities of
SR tunnels and supports bidirectional tunnels and E2E service-level OAM
detection.
 SCL: provides E2E channelized group channels for network services and
slices. The SCL uses innovative slicing Ethernet (SE) to implement timeslot
processing on Ethernet physical interfaces and FlexE bonding groups,

10 11
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

3 Key Technologies

12 13
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

3.1 Slicing Ethernet 3.1.2 Slicing Ethernet Technology

3.1.1 Slicing Ethernet Requirements There are some efforts in the industry Ethernet based hard slicing/isolation. For
example, FlexE and previous MLG developed by OIF provide a hard slicing/isola-
tion mechanism based on Ethernet physical interfaces. However, FlexE as an inter-
The future multiple services (such as 5G eMBB, URLLC, and mMTC) support raised face-level technique at this stage have no end-to-end networking consideration
the end- to-end network slicing requirement for the future transport network. to meet the carrier network requirements. In the SPN architecture, an innovative
Different services of different industry have differentiated requirements could be new network layer, the Slicing Ethernet (SE) is proposed to provide Ethernet based
supported by different QOS guaranteed network technologies (e.g. Packet/TDM). slicing/isolation capability. Slicing Ethernet prevents L2/L3 packet storage and ta-
The end-to-end sliceable future transport network with allocated isolated network ble lookups, provides end-to-end hard isolated pipe Layer 1 networking capability
resource should be able to prevent negative impact of adding new service to the with node low latency. Slicing Ethernet (SE) has the following features:
existing network and thus reduce operation costs.
 Slicing Ethernet cross-connect (SE-XC): Supports extremely low node for-
Network slicing enables relevant service functions and network resources of warding latency and hard isolation for services. Providing end-to-end SPN
a physical network to be organized together to form multiple completely separat- Slicing Ethernet Channels and support E2E Ethernet L1 networking.
ed, autonomous, and independent logical network slices, with each slice meeting
 On-demand E2E OAM: The on-demand Slicing Ethernet OAM&P messag-
specific user/service requirements. es replacing Ethernet IPG idle blocks are used for supporting strong and
The high-bandwidth, massive-connection, and low-latency services pose good enough OAM and protection functions in an end-to-end Slicing Eth-
higher requirements on the transport network. Effectively isolating transport net- ernet Channels based on requirement and configuration. E2E protection
work resource to satisfy differentiated SLA requirements of various services is a switching can be implemented within 1 ms, and system bit error detection
new challenge for the future-proof transport network. Current slicing/isolation can be performed as required.
mechanism can be classified as soft slicing/isolation and hard slicing/isolation.
 Transparent mapping of CBR services: Specific transcoding and data rate
Packet based (e.g. MPLS) VPN technology is considered as an efficient solution adjusting mechanism is used to transparently map various services into
for soft slicing/isolation in packet transport network with statistical multiplexing Slicing Ethernet channels, over which Ethernet and/or non-Ethernet CBR
capability. However, packet based VPN technology as a soft slicing/isolation mech- services could be transparently transmitted.
anism cannot solve the bandwidth preemption problem of different services and
cannot guarantee the SLA of the services. Especially in the case of considering
improve bandwidth utilization, many data center networks use the jumbo packet
technique, which prolongs packet scheduling between slices and adversely affects
the SLA capability for the other slices. The IP/Ethernet friendly future multiservice
transport network need to be enhanced with hard slicing/isolation capability. Eth-
ernet has become the mainstream technology for its simplicity, high efficiency, and
low cost, the hard slicing/isolation mechanism should be Ethernet based.

14 15
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

Figure 3-1: Slicing Ethernet technology 3.2 Efficient High-Bandwidth Technology

3.2.1 Efficient High-Bandwidth Requirements

With the rapid development of 5G mobile, HD video, data center interconnection,

and IoT services, the bandwidth of transport networks will grow rapidly. The 5G
mobile New Radio (NR) uses wider spectrum and higher spectral efficiency to pro-
vide bandwidth experience at a peak rate of 10 Gbit/s for each cell, improved by
100 times compared with 4G mobile networks. With the development of 4K, 8K,
and VR video services, the bandwidth requirement of a single user will increase to
100 Mbit/s to 1 Gbit/s. Thus the transport network urgently requires new efficient
technologies to accommodate service bandwidth development demands. In ad-
Slicing Ethernet L1 networking technique is an effective enhancement to the dition, the average revenue per user (ARPU) of a single user cannot be increased
L2 Ethernet L2.5 MPLS and IP L3 networking. With the introduction of Slicing Eth- greatly. Therefore, the bandwidth growth should not increase network construc-
ernet, the SPN has the Ethernet-based multi-layer networking capability and satis- tion costs largely.
fies differentiated multiservice transport requirements. End-to-end SPN Slicing Eth- The Ethernet interface is the most widely used interface technique in the
ernet Channels are used for high-value private lines. Layer 2 or Layer 3 per packet communications field. It is also the most cost-effective interface. After nearly four
scheduling and/or soft tunnels which support packet statistics multiplexing is used decades of development, Ethernet interfaces have formed a mature industry chain
to achieve efficient bandwidth utilization. For low-latency packet services, Layer 2 and have been widely used on telecom and IT networks. Owing to the develop-
and Layer 3 packet scheduling could be limited at edge PE nodes, on P nodes with- ment of the Internet, high-speed Ethernet interfaces have been developed rapidly
in the network Slicing Ethernet cross-connecting (SE-XC) at Layer 1 could rapidly in recent years to meet ever-increasing service bandwidth requirements.
forward packets of the same SPN Slicing Ethernet Channels with low latency.

Figure 3-2: Slicing Ethernet application

3.2.2 Efficient High-Bandwidth Technology
PE P PE
Common
services
IP/MPLS Driven by requirements and technologies, the physical layer of Ethernet interfaces,
IP/MPLS (TSN+low IP/MPLS the low-cost single-channel technique, and the high-performance multi-channel
Low- latentcy) Low-latency rapid
latency forwarding path technique are developed at the same time to achieve optimal cost-effectiveness.
services MAC MAC
MAC The 50GbE, 200GbE, and 400GbE rates with the four-level pulse amplitude mod-
High-value
private line Ethernet transparent
services SE SE services
ulation (PAM4) and 25G optical components gradually become the mainstream
SE
next-generation Ethernet interfaces.
FlexE/Ether FlexE/Ether
FlexE/Ether
net net
net
Timeslot switching+ Timeslot Timeslot switching+
low-latency forwarding switching low-latency forwarding

16 17
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

Figure 3-3: Ethernet port rate evolution 3.3 Deterministic Low Latency Technology

The deterministic ultra-low latency requirement is an important change in 5G ser-

vices, compared with the LTE era. eMBB services require 10 ms E2E latency and
URLLC services require 1 ms E2E latency, which poses higher requirements than
LTE services.
When forwarding service packets, traditional packet devices place packets
in queues on outbound interfaces, resulting in long latency of tens of μs. In case
of network congestion, latency is longer and even reaches the millisecond level,
which cannot meet the requirements of low-latency services in the 5G era. The 5G
Based on 25G optical components and key technologies such as forward er-
URLLC service requires the low forwarding latency at the μs level and low delay jit-
ror correction (FEC) and PAM4, the rate is doubled, achieving the Lane 50 Gbit/
ter on a single node.
s data rate and reducing the cost per bit. On the basis of 50GE over a single lane,
the multi-lane mode is used to develop low-cost high-speed Ethernet interfaces,
such as 200GE and 400GE. PAM4-based high-speed Ethernet technique standards,
including 50GE, 200GE, and 400GE standards, have been developed in 802.3bs 3.3.1 Packet Low Latency
(200GE and 400GE) and 802.3cd (50GE). The standards will be released in 2018,
which are widely recognized by the whole industry.
The 50GE, 200GE, and 400GE interfaces involve the following key technolo- The packet low latency forwarding technique improves the forwarding mechanism
gies: that rapidly forwards packets which are identified as low latency packets, and do
1. FEC: The mature KP4 FEC technique is implemented for long-distance trans- not need queue processing. If an outbound interface is forwarding other packets
mission. with lower priorities, the low-latency services can preempt resources and are rap-
2. PAM4: PAM4 is used to double the data rate when the baud rate remains idly forwarded in real time.
unchanged, and effectively reduces the cost of optical interfaces.
Proposed in 2014, FlexE was expected to apply to DCI whose main require- Figure 3-4: Fast packet forwarding
ment is to decouple the MAC layer data rate from the PHY layer data rate. FlexE DRAM

supports flexible mapping between MAC and PHY layers and the bonding of mul- Normal Normal
Normal
tiple PHYs, and implements various high-bandwidth interfaces. In the transport MAC
forwarding
TM
MAC
803.br 803.br
Normal
network field, FlexE is primarily used to provide variable data rates, which matches Express
Rapid
forwarding
Time-aware
scheduler Express

available bandwidth values on a network and support multi-wavelength applica-

tions, and saves optical fibers.
On an SPN network, FlexE bonds multiple optical interfaces and provides
high-rate Ethernet interfaces based on a low-cost low-rate optical module. For
example, four 200GE Ethernet interfaces are bonded to provide the 800 Gbit/s
capacity on a single port. In addition, with the combination of FlexE bonding and
The fast packet forwarding technique reduces the forwarding latency of a sin-
DWDM, the SPN system can support single-fiber with Terabit-level capacity.

18 19
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

gle node to the 10 μs scale. Figure 3-5: Figure 3-5 SE-XC principles

3.3.2 Slicing Ethernet Low Latency Technology

Traditional packet devices forward customer service packets hop by hop. On each
node, those Packets experiences PHY layer processing, packet assembly, address-
ing and label lookup for forwarding, and queue scheduling before it was transmit-
ted. The processes induce unpreventable random high node latency up to dozens
of microseconds. Even with the packet fast forwarding mechanism, node latency
is reduced at certain level but end-to-end latency variation (jitter/PDV) still cannot
meet the requirements of latency-sensitive 5G services. Therefore, a further laten-
cy optimization solution targeting at deterministic low latency needs to be consid-
ered.
The slicing Ethernet-based deterministic low latency solution has introduced
SE-XC technologies to replace the traditional store-and-forward mechanism. User
packets do not need to be parsed on intermediate nodes of a network, and the
service flow forwarding process is almost completed in real time. The node latency The preceding figure shows the SPN channel switching process. FlexE Groups
of a pizza box device can be optimized to the microsecond level. A and B indicate two logical interfaces on a device. Each group can have m physical
For instance, a PE maps user service packets to SPN channels. The intermedi- links. Assume that three FlexE client services are configured.
ate node performs SE-XC and then directly forwards the service packets to the out-
 Client1 in blue is mapped to the timeslots 1, 4, and 7 and switched to the
bound interface without the time consuming queuing process. Therefore it could timeslots 5, 8, and 10 on the right.
achieve extremely low forwarding latency.
 Client2 in red is mapped to the timeslots 12, 16, and 18 and switched to
the timeslots 9, 11, and 12 on the left.
 Client3 in yellow is a pass-through service, which is mapped to the
timeslots 13, 15, 17, and 19 on the left and switched to the timeslots 1, 3,
4, and 7 on the right.
For the pass-through service in yellow, Group A receives code blocks from m
PHY chips and restores the m x 20 code block sequences. Group A extracts code
blocks from timeslots 13, 15, 17, and 19 based on the preconfigured timeslot table
and restores the code block stream of Client3. Then, the code blocks are inserted
into the timeslots 1, 3, 4, and 7 in the m x 20 code block sequences on the right,
implementing the SPN channel switching. Finally, the code blocks are forwarded to
the next node through the PHY chips.

20 21
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

The SE-XC-based forwarding technology implements service forwarding at programmability. However, SR-TE only maintains path information on the source node,
the timeslot layer. Similar to the L1 forwarding technology, SE-XC ensures that the making it impossible for bidirectional association. In this case, SR-TE lacks the tradi-
minimum delay is less than 1 µs, and the jitter is very small, which is suitable for tional transport-geared capabilities of end-to-end OAM detection and bidirectional
carrying URLLC services. SE-XC meets the stringent requirements on latency and tunnels.
jitter for services, such as financial high-frequency transactions and autonomous An SRTP-TE tunnel uniquely identifies an end-to-end connection by carrying a
driving. path segment and implements bidirectional tunnels by path segment association. As
a subset of the SR-TE tunnel, the SRTP-TE tunnel features SR-TE’s flexible features and
the transport capability of traditional tunnels. SRTP-TE is an ideal evolution solution for
3.4 Flexible and Reliable Connection Technology MPLS-TP tunnels on the existing transport network.

Figure 3-6: SRTP-TE Tunnel

3.4.1 Flexible Connection Requirements

The 5G mobile backhaul is deployed with the flat IP architecture. The east-west
traffic between base stations soars due to the requirements of ultra-dense net-
working. In addition, deploying MEC downstream requires the setup of east-west
channels between different MECs. The number of 5G network connections increas-
es dozen-fold compared with the number of 4G network connections. Traditional
static MPLS-TP tunnels cannot meet the requirements of flexible service schedul-
ing and ubiquitous connections. Therefore, a new tunnel connection technology is
required.  SRTP-BE Tunnel
An SRTP-BE tunnel is automatically generated after an IGP flooding. Full-Mesh
tunnel connections can be generated in the IGP domain. SRTP-BE tunnels are ap-
plicable to connectionless Mesh services and simplify tunnel planning and deploy-
3.4.2 Flexible Connection Technology ment.

As a source routing technology, segment routing (SR) simplifies the existing MPLS
technology, reuses the existing MPLS forwarding mechanism, is compatible with 3.5 Ultra-High Precision Synchronization Technology
the current MPLS network, and supports smooth evolution from the existing MPLS
network to the SDN network. SRTP is a new tunnel technology that enhances O&M
capabilities in the transport domain based on the SR tunnel technology. It is classi- 3.5.1 High-Precision Time Synchronization Requirements
fied into SRTP-TE tunnel or SRTP-BE tunnel.
 SRTP-TE Tunnel 5G divides traditional BBU into two logical entities: CU and DU. CU devices process
A traditional SR-TE tunnel does not require the maintenance of the path status on non-real-time wireless upper-layer protocol stacks, and DU devices process physi-
a transit node, therefore improving flexibility of tunnel path adjustment and network cal layer functions and real-time requirements. Time synchronization requirements

22 23
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

for CU and DU are different. CU devices do not require precise time synchroniza- Table 3-1: The time error allocation for the 4G network
tion, while the DU-AAU level has the following synchronization requirements: Time Transport Base Station E2E
(1) Synchronization for the 5G TDD basic service Reference Network Synchro-
For TDD base stations, uplink and downlink signals utilize the same frequency. Source nization
To prevent signal interference between base stations, strict phase synchroniza- budget
±250 ns ±1000 ns ±250 ns
±1500
tion between base stations must be satisfied to ensure that the uplink-downlink
switching time on all the base stations is consistent. The introduction of ultra-short ns
frames to 5G New Radio (NR) speeds up the uplink-downlink switching frequency, For the 5G SPN, considering the synchronization requirements and future
which requires the time synchronization accuracy to the microsecond level. technology evolution and the balance between difficulty and cost, the time error
(2) Synchronization for the coordination technologies between base stations allocation is suggested as following table.
The coordination technology is an enhanced attribute for 5G, which has strict
requirements on the time error, e.g. base station coordination (CoMP) and base Table 3-2:The time error allocation for the SPN
Time Transport Base Station E2E
station carrier aggregation (CA) requires the time precision of about ±130 ns,
Reference Network Synchronization
whereas the evolution of the MIMO technology proposes the synchronization ac-
Source budget
curacy requirement of ±65 ns.
±50 ns ±200 ns ±50 ns ±300 ns
(3) Synchronization for new 5G services
For TDOA-based base station positioning services, the positioning precision is The 200 ns budget for the transport network can be allocated as follows:
linearly related to the time error between base stations. The 1-ns time error corre- (1)100 ns for normal time synchronization. Each node is assigned ±5 ns. A to-
sponds to the positioning precision of about 0.3 m to 0.4 m, and the 3-m position- tal of 20 hops are supported for a time chain.
ing precision corresponds to the time error of about 10 ns. (2)100 ns for time holdover during a fault.
Considering 5G service evolution requirements, 5G has higher synchroniza- Ultra-high precision time synchronization is required for 5G fronthaul,
tion precision requirements than ±1500 ns of the 4G TD-LTE technology. midhaul, and backhaul networks.

3.5.2 SPN Time Synchronization Network Model 3.5.3 Ultra-High Precision Synchronization Technology for
the SPN
The time error allocation on the time distribution chain for the 4G network is as
follows. Ultra-high precision time synchronization requires new time source and time trans-
mission technologies. In order to support the new network architecture, new syn-
chronization interface, maintenance and management technologies are required.
(1)Ultra-High Precision Time Reference Source
The ultra-high-precision time reference source needs to achieve the synchro-
nization accuracy better than ±50 ns. The following technologies can be used to
enhance the accuracy:
 New satellite receiving technology

24 25
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

Possible solutions for reducing the GNSS receiving time error include using es.
the GNSS common-view method or dual-band GNSS receiving technology.  Novel in-band synchronization interface
 High-stability frequency source technology Since the SPN utilizes new traffic interfaces, accordingly it must support new
The introduction of the clock group (such as a rubidium clock group) instead synchronization interface technologies:
of a single clock improves the stability and the holdover performance without a) FlexE interface synchronization: supports ultra-high precision time synchro-
GNSS. nization and interworking.
(2)Time Transmission Technology b) eCPRI interface synchronization: Ethernet interface synchronization is used
In the time transport process, the time error sources come from the follow- to implement eCPRI interface synchronization and interworking.
ing: (4)Intelligent Clock Technology
 Time stamping granularity The intelligent clock technology at the control layer provides the support for
 Physical-layer frequency error the deployment and O&M of ultra-high precision synchronization networks. The
 PHY-layer asymmetry core functions are as follows:
 Propagation delay inside a node  Automatic planning of the synchronization network
 Link asymmetry  Graphical display of the dynamic synchronization status
To achieve ultra-high precision time synchronization, the following techniques  Detection and analysis of the synchronization running status
are available:  Intelligent synchronization fault diagnosis
 The high-precision timestamp technology enables the time stamping gran-  Real-time monitoring and analysis of the synchronization performance
ularity to be better than ±1 ns.
 The clock module with high stability and synchronization precision pro-
vides more precise and stable frequency.
3.6 Centralized Management and Control Technology of
 Precise latency measurement and compensation for the PHY layer.
SDN
 Precise latency design and measurement for the system.
 Precise latency measurement and control for asymmetric links. This prob-
lem can be solved by using single-fiber bidirectional transmission. 3.6.1 Requirements
 Enhanced synchronization test protocols and algorithms with higher pre-
cision.
(3)Synchronization Interface Technology SDN is one of the key 5G technologies. It can be summarized as centralized net-
work control, separation of forwarding and control, open network and program-
The SPN should support dedicated synchronization interfaces and novel in-
mability. SDN uses a three-layer architecture (service application layer, network
band interfaces with ultra-high precision.
control layer, and forwarding layer) to improve 5G network resource utilization,
 Dedicated synchronization interface with ultra-high precision
accelerate service deployment, and enable flexible service scheduling as well as
The dedicated synchronization interface is used for the time source input, the
open network programmability.
time output and measurement. The following interfaces could be considered:
The SDN controller, as an intelligent network operating system, has the fol-
(a) 1GE optical synchronization interface: used to access ultra-high precision
lowing development trends:
time sources and transmit time information to downstream devices.
(1) E2E management capability
(b) 1PPS synchronization output interface: used to output the 1PPS signals of
The SDN technology allows the SPN to transform from vertical management
the local node, facilitating the measurement using the synchronization test devic-

26 27
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

of individual NEs, networks, and services to closed-loop management covering Figure 3-7: Key technologies for intelligent operations
deployment of network and service to evaluation feedback. The operating system
forms a closed-loop process from service creation, network monitoring, fault anal-
ysis, and performance evaluation to network fault switchover and recovery. This
greatly improves service deployment efficiency and network robustness.
(2) Automated OAM and analysis
To cope with dynamically changed services, the intelligent operating system
must transform towards automated management to agilely manage network func-
tions, applications, and services. This capability automatically generates OAM poli-
cies based on the Big Data analytics of the 5G network.
(3) Open operating capability
Another important principle of intelligent operations is to break through the
technical barriers of different vendors and build an open platform for unified man-
agement and control. The free and open-source software architecture and the
carrier network capabilities open to third parties can accelerate the maturity and
commercial use of new technologies and boost innovation.

3.6.2 Key Technologies for Intelligent Operations

(1)Low-coupling service architecture
To adapt to the changes of the SPN architecture, the new intelligent operations The core idea of the low-coupling service architecture is to atomize various
system must make changes in key technologies such as service architecture, ser- independent functional modules into independent atomic services, so that pro-
vice model and policy, operations process, and reliability and security assurance. viding different service scenarios through aggregation and combination. Atomic
The following figure shows the mapping between the development trend of intelli- functions are classified into different types based on basic service functions. The
gent operations and key technologies. orchestrator or controller arranges and combines sub-functions as required for dif-
ferent service scenarios.
(2) Automated management process
Standardization and automation of design, creation, deployment, manage-
ment, and iteration of new services based on the advantages of the SDN net-
work architecture can greatly improve network operations efficiency. Automated
management involves automated service configuration, automated environment
awareness, automated software and hardware installation and upgrade, automat-
ed fault detection, automated service switchover and recovery, intelligent service
data analysis, and so on.
(3)E2E policy management

28 29
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

A policy management module can be developed to formulate, parse, execute, Figure 3-8: Architecture of intelligent operating system
and optimize policy rules for automated orchestration. The orchestrator can deliv-
BBS/third-party app
er policies to the unified policy library based on design requirements. Once a net-
work event occurs, the controller or orchestrator can automatically perform pre-
defined actions to reduce manual intervention and increase the response speed.
Orchestrator
(4)Design of unified open model
A set of standard data models are the prerequisite for E2E service manage-
ment, data analysis, and an open operations platform. Based on standard specifi-
Capability opening interface
cations, the formulated models and process of unified topology reporting, alarm &
performance reporting, and service configuration can significantly reduce E2E ser- Capability Policy management Service assurance Resource storage

vice differences. In addition, the data analysis system can extract key information orchestration engine
Policy
Policy
Orchestration Capability rule

and analyze data based on these fixed models. engine list creation
parsing
Service Network

(5) Reliability and security assurance Atomic

capability
Policy
exec ution
Policy
optimization Device
resource

The orchestrator and controller, as the core of the operations system, must
guarantee reliability of the intelligent operating system for providing timely E2E Data migration, internal APIs
management. By dividing a network into multiple subnets and allocating each sub- Service
resource

net to multiple controller clusters for management and control, the cluster deploy-
Data collection and service configuration module
ment and distributed storage solution improves network robustness and prevents Route
resource
single point failure or single point traffic congestion. NETCONF SNMP BGP-LS

3.6.3 Architecture of the Intelligent Operating System

The following figure shows the architecture of an SPN-based intelligent operations

system. The core module is the SDN controller, which provides intelligent opera-
tions through centralized control and management of network devices.

30 31
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

4 Application Scenarios

32 33
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

Being the next-generation integrated transport network technology, the SPN is cross-connect technology of the SCL satisfies the 5G fronthaul network’s
applicable to diversified application scenarios. Based on brand-new network archi- demand for low latency.
tecture, multiple key technologies and new operation mode, SPN is a good choice  Complete OAM and protection mechanisms enable real-time fronthaul
for the user to carry multiple services on an integrated transport network in the 5G network monitoring, swift fault localization and fast protection switcho-
era. ver.
 Low-cost 100GE short-distance silicon optical modules are employed to
reduce network construction costs.
4.1 Mobile Service Transport  The industry-leading CPRI over FlexE technology provides compatibility
with traditional 3G/4G fronthaul interfaces.

The unified transport of fronthaul, midhaul and backhaul services enabled by the
SPN prevents heterogeneous networks from interoperation, which not only makes
4.1.2 Midhaul/Backhaul Network
the entire network planning, provisioning, O&M, and optimization much easier, The 5G base stations can be deployed in different ways. For example, the DUs
but also reduces the network CAPEX and OPEX greatly. The architecture of the and CUs can be deployed together or separately. If the DUs and CUs are deployed
SPN-based mobile transport network is as shown in Figure 4-1. together, the transport network needs to transmit services to the core network. If
the DUs and CUs are separately deployed, the midhaul network completes service
Figure 4-1: Mobile Transport transmission between the DUs and CUs, and the backhaul network transport ser-
vices from the CUs to the core network. Both deployment methods may coexist.
Despite of different deployment ways, the 5G base station raise unified de-
mands for the transport network, including: high bandwidth, low latency (lower
than that of 4G networks), stronger L3 routing capabilities, L3 to edge, and flexible
deployment and scheduling of southbound, northbound, and east-west services.
An SPN-based metro network consists of three layers (access, aggregation and
core layers), implementing unified transport of the midhaul and backhaul services.

4.2 Data Center InterconnectionI

4.1.1 Fronthaul Network
In the future, DCs will be more widely used to carry various NFV cloud software
In CRAN scenarios where the DUs or integrated DUs and CUs are deployed in and IT applications, playing a pivotal role in carrier service deployment. At the
a centralized manner, the network between the AAUs and the DUs or integrated same time, service interoperation between different DCs will be more complicat-
DUs/CUs is referred to the fronthaul network. The fronthaul network requires high ed. The SPN-based integrated transport network can be used to interconnect the
bandwidth (more than 20 Gbit/s per AAU) and low latency (less than 100 µs). DCs.
The SPN provides the following features for fronthaul network:
 The ultra-low-latency service forwarding based on the SPN channel

34 35
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

4.2.1 Multi-point Interconnection Between Edge DCs over 4.3 Video Everywhere
the SPN
When multiple DCs exist in one MAN, mesh networking is required. The SPN With the increasing popularity of the Internet, video services provide an important
can be employed to provide independent network slicing to carry the DCI services. strategic transformation opportunity for telecom carriers. In the era of big video,
The DC GWs interconnect each other via SPN L3 VPN, which satisfies the require- UHD videos such as 4K, 8K, and VR/AR raise higher requirements for the network
ments of the inter-DC traffic scheduling. bandwidth, latency, and jitter. The telecom carriers can leverage their network ad-
vantages to provide video services with higher quality and better user experience
Figure 4-2: Interconnection between edge DCs than Internet OTT players.
To transfer home broadband services such as HSI and IPTV running on optical
line terminals (OLTs), the SPN employs the SPN channel to partition the network
into multiple independent slices. The services on different slices are isolated from
each other. As per real-time traffic status, the SPN can deploy QoS to set priori-
ties for different services. So that, the video services can get corresponding pre-
cedence. To prevent IPTV service performance degradation caused by network
congestion, the independent network slicing technology can be used to carry IPTV
services. In this way, these IPTV services are strictly isolated from common Inter-
net access services, making the users enjoy highly qualified services that are quite
different from other OTT services. The service quality measurement (SQM) tech-
nology is used to monitor video traffic performance, including the latency, band-
width and jitter. Based on statistics and analysis implemented on a regular basis,
the network can be changed accordingly to ensure the service quality.
4.2.2 Point-to-Point Interconnection Between DCs over the
SPN Figure 4-4: SPN-based big video service transport

For point-to-point interconnection between DCs in a metro area, the SPN

can provide L1 transparent transport with CBR mapping and high-bandwidth hard
pipes so as to ensure low latency for DC services.

Figure 4-3: Interconnection between VDCs

With the development of big video services, the upstream bandwidth of an

OLT has expanded from NxGE to Nx10GE. The SPN that can be expanded via the
FlexE and DWDM technologies provide bandwidth guarantee for home broadband
services.

36 37
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

4.4 Cloud Private Line mapping to directly forward the bit streams at the physical layer to the SPN chan-
nels, and the intermediate nodes of the network service uses SE-XC for service
scheduling. This implementation meets the requirements for hard service isolation,
Being an excellent combination of the SDN technology and packet transport net- security, and ultra-low latency.
work, the SPN is capable of implementing fast network re-architecting, providing
suitable network slices for private lines and offering cloud connection. The SPN can Figure 4-6: CBR private line of the SPN
also address cloud application requirements by connecting DCs or deploying local
computing nodes.

4.4.1 Cloud Connection

The SPN can be divided into network slices based on resources such as ports,
wavelengths, and FlexE channels. A physical network can be divided into multiple
virtual networks, with each virtual network carrying respective private service 4.4.2 Cloud Application
without affecting each other.
In private line service scenarios involving cloud-and-network synergy, the us-
er-side CPE enables user access and enterprise virtualization services. Or access
Also, the SPN can provide differentiated L2/L3 services based on the actual
enterprise branches to provide virtual private line connections.
requirements of enterprise users. This allows flexible service scheduling between
The user services access the DC user gateway vMSR via the SPN for Internet
enterprise branches and the headquarters and provides customers with L2VPN/
and site access. The switches in branches connect the SPN through L2 VLANs, and
L3VPN capabilities.
the SPN transmits the services to the TOR switches on leaf nodes of the DC. The
DVS directs the L2 traffic to the vMSR
Figure 4-5: Cloud connection

Figure 4-7: Cloud-and-network synergy

Private line services with higher requirements for isolation, such as Ethernet
and SDH services for finance and securities sectors, can be carried by SPN chan-
nel-based hard pipes independently. The client side of the SPN device uses CBR

38 39
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

5 Summary and Prospect

40 41
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

The SPN 5G transport solution, which adheres to the vision of “New Architecture, Abbreviations
New Services, and New Operations”, supports brand-new cloud service transport
and agile network operation. The SPN uses an efficient Ethernet ecosystem to pro- 1PPS 1 Pulse Per Second
vide low-cost and high-bandwidth transport channels. By effectively integrating AAU Active Antenna Unit
multi-layer network technologies, the SPN enables the transmission of L0-L3 ser- AI Artificial Intelligence
vices and soft/hard pipe slicing. Also via the SDN-based centralized management AR Augmented reality
and control, it presents an open, agile, and efficient network operation system to ARPU Average Revenue Per User
the user. BGP-LS Border Gateway Protocol - Link State
In addition to in-depth research, China Mobile also verified the SPN network CAPEX Capital Expenditure
architecture and key technologies in labs. From August to December 2017, China CBR Constant Bit Rate
Mobile Research Institute (CMRI) organized major equipment vendors to complete CoMP Coordinated Multiple-Point Transmission/Reception
the lab tests of SPN prototypes. The test covers a variety of items, for instance, STL, CPE Customer Premise Equipment
SCL, and SPL, time synchronization and interoperation between different vendors. CPRI Common Public Radio Interface
The test results show that the SPN is able to meet the requirements for integrated CRAN Cloud Radio Access Network
transporting 5G and other MANs’ services. CU Centralized Unit
In 2018, China Mobile will conduct small-scale 5G service transport field trial DC Data Center
in multiple cities to promote SPN commercialization, so that SPN can be commer- DCI Data Center Interconnect
cially used on a large scale in 2019. DC-GW Data Center Gateway
At the ITU-T SG15 plenary meeting held at Geneva, Switzerland from Janu- DU Distributed Unit
ary to February in 2018, a project initiation is approved for the Characteristics of DVS Digital Video Server
Transport Networks to Support IMT-2020/5G (G.ctn5g) standard, which marks a DWDM Dense Wave-length Division Multiplexing
new phase of ITU-T’s 5G transport research. The SPN technology is regarded as an E2E End to End
important part of the standard. eMBB Enhanced Mobile Broadband
The year 2018 marks the beginning of 5G transport, China Mobile is ready to FEC Forward Error Correction
strengthen cooperation with global 5G organizations, enterprises, research insti- FlexE Flexible Ethernet
tutes, and universities to promote the research of SPN architecture and key tech- HD High Definition
nologies, and boost the development of SPN standards and the industry. HSI High Speed Interaction

. ICT Information Communication Technology

IDC Internet Data Center
IGP Interior Gateway Protocol
IMT-A International Mobile Telecommunications Advanced
IoT Internet of Things
IoV Internet of Vehicles
IP Internet Protocol
IPTV Internet Protocol TeleVision
IS-IS Intermediate System-to-Intermediate System

42 43
 Technical Vision of Slicing Packet Network (SPN) for 5G Transport Technical Vision of Slicing Packet Network (SPN) for 5G Transport

IT Information Technology SE-XC Slicing Ethernet Cross Connection

ITU International Telecommunication Union SLA Service-Level Agreement
L2VPN Layer 2 Virtual Private Network SPL Slicing Packet Layer
L3VPN Layer 3 Virtual Private Network SPN Slicing Packet Network
LTE-A Long Term Evolution Advanced SR Segment Routing
MAC Media Access Control SR-BE Segment Routing - Best Effort
MEC Mobile Edge Computing SR-TE Segment Routing - Traffic Engineering
MIMO Multiple-Input Multiple-Output SRTP Segment Routing Transport Profile
mMTC Massive Machine Type Communication SRTP-BE Segment Routing Transport Profile - Best Effort
MP2MP Multi-Point to Multi-Point SRTP-TE Segment Routing Transport Profile -Traffic Engineering
MPLS Multi-Protocol Label Switching STL Slicing Transport Layer
MPLS-TP MPLS Transport Profile TDD Time Division Duplex
MSR Multi-Service Router TDM Time Division Multiplexing
NE Network Element TM Traffic Management
NFV Network Function Virtualization TOR Top Of Rack
NNI Network Network Interface UHD Ultra High Definition
NR New Radio UNI User Network Interface
O&M Operation and maintenance URLLC Ultra Reliable & Low Latency Communication
OAM Operation, administration and maintenance VR Virtual Reality
OIF Optical Internetworking Forum VLAN Virtual Local Area Network
OLT Optical Line Terminal WAN Wide Area Network
OPEX Operating Expense
OTT Over The Top
P2P Point to Point
P2MP Point to Multi-Point
PAM4 Four-Level Pulse Amplitude Modulation
PCEP Path Computation Element Protocol
PCS Physical Coding Sub-layer
PHY Physical Layer
PMA Physical Media Attachment
PMD Physical Media Dependent
PTN Packet Transport Network
QoS Quality of Service
ROI Return on Investment
SCL Slicing Channel Layer
SDN Software Defined Network
SE Slicing Ethernet

44 45