NE40E

novikov.mikhail@huawei.com

gomonyuk.alexey@huawei.com
Agenda

2
 NE
Unified Mgmt., Visualized OAM (U2000, uTraffic)

Agg
Internet

DSLAM SR
/FTTx

Triple Play Softswitch

(Residential)
SingleMetro Firewall Converged IP/MPLS
BRAS
Backbone
LSW Smart CO
/E-FTTx NPE Headend
VPN
SA GGSN
(Business)
Agg

MSC Server

Mobile CSG RSG RNC/BSC

Access FMC Metro: NE40E S-POP: NE40E Core: NE5000E/NE9000 Services

NE20E-S2E/S2F NE20E-S8/S8A

NE05

NE05 NE08E NE20E-S4 NE20E-S16/S16A NE40E-X3/X3A NE40E-X8/X8A NE40E-X16/X16A NE5000E-X16/X16A

NE9000-20/8
NE5000E B2B/2+8 Cluster

10G Uplink 480G/Chassis 480G/1T per Slot B2B/2+8 Cluster Petabit-Level

3
 Huawei NE40E

NE40E/CX600-X3 NE40E/CX600-X8 NE40E/CX600-X16 NE40E/CX600-X3A NE40E/CX600-X8A NE40E/CX600-X16A

1.08T 7.08T 12.58T 2.79T 25.16T 50.32T

360 Mpps 2880 Mpps 5760 Mpps 900 Mpps 5760 Mpps 11520 Mpps

4
Agenda

5
 NE40E-X8A

Intra-device hot backup Intra-board hot backup

System-level hot backup
Power Module
#1 Chip#1

LPU Chip#2
Master SRU Slave SRU Power Module
SFU Chip#N
#N

FAN modules Power modules: N+1 backup

N+1 backup Management plane: 1:1 backup
SRU
Clock A Chip#1
1:1 backup
Chip#2
Clock B
Chip#N
SFU
3+1 backup
Clock modules: 1:1 backup
SFU SFU
Power Module LPU
#1 #4
DC: N+1 backup
AC: N+N backup Chip A1 Chip B1

High-speed data buses between boards Chip An Chip Bn

N+M backup
High-speed data buses between chips
N+M backup

6
 NE40E-X16A
Power Module Item NE40E-X16A
Power entry module
Switching capacity 50.32 Tbps (LPUI-1T)
Air intake frame
Fan Module Slot Capacity 800Gbps
MPU * 2 Port Density 128*100GE; 768*10GE

40RU Slots Number 16 LPU, 2 MPU, 4 SFU

SFU * 4
Dimension (W*D*H) 1778 mm x 442 mm x 650 mm (40U)

•7720 W (fully configured with LPUF-240s)

LPU * 9 Typical power consumption •9040 W (fully configured with LPUF-480s)
•12390 W (fully configured with LPUI-1Ts)
Fan Module
Redundant MPUs 1+1
Front View Rear View Redundant Switch fabrics 3+1

•[400G bundle] Four fan assemblies.

Redundant fans
•[1T bundle] Six fan assemblies
Slot Name Slot Quantity Slot ID
•[400G bundle] DC: 6+1
LPU 16 1 to 16 •[400G bundle] AC: 5+5
Redundant power supply
•[1T bundle] DC: 9+1
MPU 2 17 to 18 •[1T bundle] AC: 7+7

SFU 4 19 to 22

7
 NE40E-X8A
Item NE40E-X8A
Switching capacity 25.16 Tbps (LPUI-1T)
Air intake frame Slot Capacity 800Gbps

Fan Module Port Density 64*100GE; 384*10GE

SRU * 2 Slots Number 8 LPU, 2 SRU, 2 SFU

21RU
SFU * 2 Dimension (W*D*H) 930 mm x 442 mm x 650 mm (21U)

LPU * 8 •4110 W (fully configured with LPUF-240s)

Typical power consumption •4770 W (fully configured with LPUF-480s)
•6520 W (fully configured with LPUI-1Ts)
Power Module Redundant SRUs 1+1
Power entry module Redundant Switch fabrics 3+1
Front View Rear View •[400G bundle] Two fan assemblies.
Redundant fans
•[1T bundle] Three fan assemblies

•[400G bundle] DC: 3+1

Slot Name Slot Quantity Slot ID
•[400G bundle] AC: 3+3
Redundant power supply
LPU 8 1 to 8 •[1T bundle] DC: 5+1
•[1T bundle] AC: 4+4
SRU 2 9 to 10

SFU 2 11 to 12

8
 NE40E-X3A
Item NE40E-X3A
Air intake frame
Switching capacity 2.79 Tbps (LPUI-200)

Fan Module Slot Capacity 200Gbps

6RU

LPU * 3 Port Density 6*100GE; 60*10GE

MPU * 2 Slots Number 3 LPU, 2 MPU

Front View
Dimension (W*D*H) 264 mm x 442 mm x 710 mm (6U)

Air exhaust vent Typical power consumption 1120 W (fully configured with LPUF-200s)

Redundant MPUs 1+1

Power entry module
Redundant Switch fabrics None

Rear View The device can work properly for a short

Redundant fans
time at 40ºC if a single fan fails.
LPU 1 Slot Name Slot Quantity Slot ID
Redundant power supply 1+1
LPU 2
LPU 3 1 to 3
LPU 3
MPU 4 MPU 5 MPU 2 4 to 5

9
 NE40E-X16A
PM technical specifications

PM PEM
Item Parameter
Rated input voltage -48 V/-60 V
Input voltage range -40 VDC to -72 VDC
Front view Rear view Maximum input current 63 A
Maximum output power 2200 W
Typical power consumption 148.5 W
Mappings between PEMs and PMs DC

Typical heat dissipation 481.8 BTU/hour

PEM PM
A1, B1 1
 PMs work in N+1 backup mode; The number of PMs can
A2, B2 2
be configured based on the board power consumption.
A3, B3 3
 NE40E also support 240V and 380V High Voltage DC
...... ......
A16, B16 16 and 180V AC to 264V AC power supply.

10
 NE40E-X8A
PM technical specifications
Item Parameter
Rated input voltage -48 V/-60 V
PM PEM Input voltage range -40 VDC to -72 VDC

Front view Rear view Maximum input current 63 A

Maximum output power 2200 W
Mappings between PEMs and PMs DC
Typical power consumption 148.5 W
PEM PM
Typical heat dissipation 481.8 BTU/hour
A1, B1 1

A2, B2 2  PMs work in N+1 backup mode; The number of PMs can

A3, B3 3 be configured based on the board power consumption.

 NE40E also support 240V and 380V High Voltage DC
...... ......
and 180V AC to 264V AC power supply.
A8, B8 8

11
 NE40E-X3A

PEM technical specifications

The PEM is a 48V DC power entry module
Item Parameter

Rated input voltage -48 V/-60 V

Input voltage range -40 VDC to -72 VDC

Maximum input current 63 A

Diagram of the power supply system Typical power consumption 1.0 W

Typical heat dissipation 3.2 BTU/hour

 The PEM has one straight-through power input that

provides the following functions: Surge protection, filtering,
• The NE40E-X3A supports either DC or AC power supply.
• Power modules convert the input voltage into -48 V DC voltage to short circuit protection, and alarm function
supply power for the entire system.  NE40E-X3A also support AC power system with 200-240 V
• The power supply system consists of two AC power modules or two DC
power modules working in 1+1 backup mode. AC power.

12
 Power Distribution Box NE40E
 This product is a 6U (263.9 mm) power distribution box that
complies with the IEC297 standard.
 It can convert 4 DC input lines into 12 DC output lines or
2DC to 12DC using grounding copper bar.
 Easy install and cabling, flexible power expansion in future;
 Support full slots of 480G;
 PDU must independently install in another rack;
 4 NE40E-X8A in one site can share one PDU rack.

PDU technical specifications Typical power inputs for NE40E bundle

Item Parameter
Number of Number of
Dimensions (H x W x D) 264 mm x 442 mm x 639 mm (6U) Bundle Type
Powers Power Inputs
Installation platform Huawei N68E cabinet NE40E-X16A(480G DC Bundel) 7 DC 14
DC input Rated voltage -48 V
oltage NE40E-X16A(1T DC Bundel) 10 DC 20
Maximum voltage range -72V — 40V
Input Maximum current 300A
NE40E-X8A(480G DC Bundel) 4 DC 8

Output Maximum current 12×100 NE40E-X8A(1T DC Bundel) 6 DC 12

13
 NE40E-X16A/X8A/X3A

Air Exhaust Vent Design

NE40E-X16A/X8A: The air exhaust vent
draws air from the front and exhausts air
from the back to achieve the optimal heat
dissipation, facilitating cabinet layout and
equipment room planning.

Air Channel Design

NE40E-X16A/X8A: uses U-shaped air
channels that preferentially ensure heat
U-shaped dissipation for optical modules that have low
Air Channels power and require low temperature to
achieve a heat dissipation balance within
boards and ports.

Easy-to-maintain Design
NE40E-X16A/X8A: The fan assembly is
placed in the rear of the chassis, which is
easy to remove and maintain.

14
А я ш NE40E-X8A
SRU Control Plane CPUs in SRUs
(1:1 redundancy) and LPUs and control path
CPU across whole router
Route calculation
BKP Outband communication between
CPU CPU boards
PCI PCI
LPU LPU Device management and
maintenance
PFE TM TM PFE
Data configuration
CPU CPU Data storage
PCI PCI
LPU LPU
SFU
PFE TM TM PFE
(3+1 Redundancy)
Data Forwarding Plane(PFE and
CPU CPU TM in LPUs and SFUs)
PCI PCI
LPU LPU Forwarding packets
PFE TM TM PFE Editing packets
B B Applying QoS policy
Heat dissipation system Power supply system
(1+1 redundancy)
K K
(N*+N redundancy) Applying security policy
P P

PFE: Packet Forwarding Engine

TM: Traffic Manager Control Bus
BKP：Back Plane
Data Bus

15
 А

Acronyms
MPU
PIC: physical interface controller  The PIC performs data link layer
NP: network processor
TM: traffic manager
Packets
Host packets FIC: fabric interface controller process encapsulates packets into
( local attack defense)
Ethernet frames and sends the frames
Uplink board

Uplink NP Uplink TM Uplink FIC

to the uplink NP.
SFU cell
Optical or electrical

CAR traffic policing

Frames Packets
Physical link

Inbound interface

for warding
processing

classification
Traffic
Packets

Route-based
 The NP looks up forwarding table,

Frag mentation
avoidance

Traffic shaping
Congestion

Congestion
manage ment
Uplink
signals

PIC
performs traffic classification, traffic
policing and the traffic management
performs congestion management and
SFU
Downlink board
avoidance.
Downlink NP Downlink TM Downlink FIC
electrical signals

Packets Packets SFU cell

Then the FIC fragments the packet into

Multicast replication
Frames 
infor mation
processing

policing

classification
Traffic

avoidance
CA R traffic

Data link layer

Congestion
Outbound interface

Traffic shaping

Congestion
manage ment
Physical link

Optical or

Downlink

processing
Packet
PIC
cells with fixed length for high speed
switching.

16
А NE40E-X16A
Channels: Each SFU provides N * pairs SFU Redundancy: 3+1
of high rate serial links for each LPU. Control plane

MPU MPU 128 pairs of

SerDes link SFUI-1T
SFU SFU
per LPU
.. ..
LPU LPU
LPU LPU
The rate of each pair of SerDes links is 12.672Gbps.
SFU SFU
SFU

e.g.NE40E-X16A with 1T Line Card

SFUI-1T SFUI-1T SFUI-1T SFUI-1T • Full-mesh interconnection between SFU and LPU
• One link means 32 pairs of SerDes
• Switching capacity = slot number * SerDes
number per slot * SerDes bandwidth * (64B / 66B
overhead) * 2 (Bidirectional)
• For example: The Switching Capacity of NE40E-
X16A is (16*128*12.67*(64/66)*2) = 50.32Tbps
LPU1 LPU2 LPU3 LPU4
... LPU13 LPU14 LPU15 LPU16

17
А NE40E-X8A
Channels: Each SFU provides N * pairs SRU＝MPU＋SFU SFU Redundancy: 3+1
of high rate serial links for each LPU. Control plane

128 pairs of
SerDes link SFUI-1T
SRU SRU
per LPU
.. ..
LPU LPU
LPU LPU
The rate of each pair of SerDes links is 12.672Gbps.
SFU SFU
SFU

e.g.NE40E-X8A with 1T Line Card

SFUI-1T SFUI-1T SFUI-1T SFUI-1T • Full-mesh interconnection between SFU and LPU
• One link means 32 pairs of SerDers
• Switching capacity = slot number * SerDes
number per slot * SerDes bandwidth * (64B / 66B
overhead) * 2 (Bidirectional)
• For example: The Switching Capacity of NE40E-
X8A is (8*128*12.67*(64/66)*2) = 25.16Tbps
LPU1 LPU2 LPU3 LPU4 LPU5 LPU6 LPU7 LPU8

18
А NE40E-X3A

MPU Control plane

(1:1 backup)
Fan
32 pairs of 200G LPU
SerDes link
per LPU .. ..
200G LPU 200G LPU
LPU LPU
The rate of each pair of SerDes links is 15Gbps.

LPU
Data plane

• Full-mesh interconnect among 3 LPUs

• One interconnect link up to 32 pairs of SerDers between two LPUs
• X3A Switching capacity = 3 (slot number) * SerDes number per slot * SerDes bandwidth * (64B / 66B overhead) * 2 (Bidirectional)
• For example: The Switching Capacity of NE40E-X3A is (3*32*15*(64/66)*2) = 2.79Tbps

19
Solar
Starting Catch up Leading Leading

20G 50G 240G 1T

SOLAR 1.0 SOLAR 2.0 SOLAR 3.0 SOLAR 5.0

Multiple cores ,multiple threads，provide high parallel processing capability

High Architecture Optimization, 1T throughput design for solar 5.0
Solar Chip performance Smart Memory
Highlights
Fully programmable，support multiservice IPv4/IPV6/MPLS/VPN
Flexibility Great code instruction space，massive table size
Hardware and software collaborate, shorten TTM

Auto frequency modulation

Green Design ATOM group auto-actived
Interface serdes auto adjusting

ATOM Fault Detection and Isolation

High reliability Thread Fault Detection and Isolation
Function Block Fault Detection and Isolation/Reset
20
 NE40E
800G LPU 480G LPU 240G LPU

1 * 100GE
24 * 10GE

12 * 10GE
2 * 100GE
2 * 100GE 1 * 100GE
48 * 10GE 4 * 100GE 2 * 100GE 24 * 10GE
8 * 100GE + 24 * 10GE + 12 * 10GE
LPUI-1T LPUI-480 LPUF-480 LPUI-240 LPUF-240

120G LPU 50G LPU

24 * GE 8 * STM-1 8 * GE
6 * 10GE 12 * 10GE

2 * 10GE 4 * STM-1 2 * 10GE

1 * 40GE 8 * STM-16

2 * STM-64 2 * STM-1 8 * E1/T1

1 * 100GE 2 * STM-64
5 * 10GE 48 * GE 2 * 10GE
12 * 10GE 1 * 100GE 2 * 40GE
+ 24 * GE
LPUI-120 LPUF-120 LPUI-51 LPUF-51 LPUF-50

21
 Card Type Sub-card Bras L3VPN High Queue

LPUF-XX-E * Flexible YES YES YES

Flexible line card
LPUF-XX Flexible No YES YES

LPUF-XX-B Flexible No No YES

Partly Yes
LPUI-XX-E Integrated YES YES
(Except LPUI-51-E)
Integrated line card
LPUI-XX Integrated No YES No

LPUI-XX-B Integrated No No No

* LPUF-XX-E should be used with –E sub-cards in BRAS scenario to achieve best user access performance.

22
 (MPU) NE40E-X16A

Interface Interface
Main Processing Unit B5 (16G Memory) Connector Type
Name Type
FSP0, Serial Reserved. It functions as an interface for fast chassis
RJ45
FSP1 interface switchovers.
GE/10GE GE/10GE SFP+/SFP Reserved cascading interface.
MGMT- ETH(10M/10 It connects to an NMS and can work in half-duplex or
RJ45
ETH 0M/1000M) full-duplex mode.
Console RS-232 It connects to the console for on-site system
serial RJ45 configuration. Baud rate: 9600 bit/s (default value),
interface which is configurable.

Item Specification CLK/TOD0, External Used to input or output 2-Mbit/s clock signals, 2-MHz
CLK/TOD1 synchronizat RJ45 clock signals, 1pps+ASCII time signals, or two channels
Processing unit Quad-core 2.0GHz ion interface of DCLS time signals.
SDRAM 8 GB x 2，25M BGP routes CLK/1PPS External Used to input or output 2-Mbit/s clock signals, 2-MHz
Flash 16 MB synchronizat SMB clock signals, or 1 PPS signals.
ion interface
Storage SSD card:8 GB
CLK/Serial External Used to input or output 2-Mbit/s clock signals, 2-MHz
1:1 backup, synchronizat SMB clock signals, or RS-232 signals.
Reliable ion interface
Non-Stop Routing

23
 (SRU) NE40E-X8A

Switch and Route Processing Unit A9(16G Memory) Interface Interface

Connector Type
Name Type
FSP0, Serial Reserved. It functions as an interface for fast chassis
RJ45
FSP1 interface switchovers.
GE/10GE GE/10GE SFP+/SFP Reserved cascading interface.
MGMT- ETH(10M/10 It connects to an NMS and can work in half-duplex or
RJ45
ETH 0M/1000M) full-duplex mode.
RS-232 It connects to the console for on-site system
Console serial RJ45 configuration. Baud rate: 9600 bit/s (default value),
interface which is configurable.
Used to input or output 2-Mbit/s clock signals, 2-MHz
Item Specification CLK/TOD0,
RJ45 clock signals, 1pps+ASCII time signals, or two channels
CLK/TOD1
Processing unit Quad-core 2.0GHz of DCLS time signals.
External
synchronizat Used to input or output 2-Mbit/s clock signals, 2-MHz
SDRAM 8 GB x 2，25M BGP routes CLK/1PPS SMB
ion interface clock signals, or 1 PPS signals.
Flash 16 MB Used to input or output 2-Mbit/s clock signals, 2-MHz
CLK/Serial SMB
Storage SSD card:8 GB clock signals, or RS-232 signals.
External
1:1 backup,
Reliable SMB clock SMB
Non-Stop Routing interface

24
 (MPU) NE40E-X3A

Main Processing Unit D4(16G Memory) Interface Interface

Connector Type
Name Type
FSP0, Serial Reserved. It functions as an interface for fast chassis
RJ45
FSP1 interface switchovers.
GE/10GE GE/10GE SFP+/SFP Reserved cascading interface.
MGMT- ETH(10M/10 It connects to an NMS and can work in half-duplex or
RJ45
ETH 0M/1000M) full-duplex mode.
RS-232 It connects to the console for on-site system
Console serial RJ45 configuration. Baud rate: 9600 bit/s (default value),
interface which is configurable.
Used to input or output 2-Mbit/s clock signals, 2-MHz
Item Specification CLK/TOD0,
RJ45 clock signals, 1pps+ASCII time signals, or two channels
CLK/TOD1
Processing unit Quad-core 2.0GHz of DCLS time signals.
External
synchronizat Used to input or output 2-Mbit/s clock signals, 2-MHz
SDRAM 8 GB x 2，25M BGP routes CLK/1PPS SMB
ion interface clock signals, or 1 PPS signals.
Flash 16 MB Used to input or output 2-Mbit/s clock signals, 2-MHz
CLK/Serial SMB
Storage SSD card:8 GB clock signals, or RS-232 signals.
External
1:1 backup,
Reliable SMB clock SMB
Non-Stop Routing interface

25
 LPUI-1T 8x100G

 4M IPv4 FIB Entries or 2M IPv6 FIB Entries (Shared)

 2M MAC
 256K ARP and 64K ND (Separated)
 128K ACL or 16K ACLv6 (Shared)
 Power Consumption: < 1W/G
 Packet Buffer: 100ms

LPUI-1T
8-Port 100GBase-CFP2

Attribute Description
Optical type supported 100Gbps CFP2 optical module
Working mode Full-duplex
Compliant standard IEEE 802.3
Frame format Ethernet_II, Ethernet_SAP and
Ethernet_SNAP

26
 480G

LPUF-480 LPUI-480

48*10GE 2*100GE + 24*10GE

24*10GE 2*100GE 4*100GE

Attribute Description  4M IPv4 FIB Entries or 2M IPv6 FIB Entries (Shared)

Optical type supported 10Gbps SFP+ optical module  2M MAC
10Gbps SFP+ BIDI optical module  256K ARP and 64K ND (Separated)
100Gbps CFP2 optical module  256K ACL or 32K ACLv6 (Shared)
 Packet Buffer: 100ms
Working mode Full-duplex
 64K/128K PPPoX/IPoX users/slot, 1M users/chassis
Compliant standard IEEE 802.3  Power Consumption: < 1W/G

27
 240G

LPUF
LPUF-240 (1*100GE) (12*10GE) (3*40GE)

LPUI
LPUI-240(2*100GE) LPUI-240(24*10GE) LPUI-240(1*100GE+12*10GE) LPUI-240 (6x40GE)

Attribute Description  4M IPv4 FIB Entries or 2M IPv6 FIB Entries (Shared)

Optical type supported 10Gbps SFP+ optical module  2M MAC
10Gbps SFP+ BIDI optical module  256K ARP and 64K ND (Separated)
100Gbps CFP2 optical module  256K ACL or 32K ACLv6 (Shared)
 Packet Buffer: 200ms
Working mode Full-duplex
 64K/128K PPPoX/IPoX users/slot, 1M users/chassis
Compliant standard IEEE 802.3  Power Consumption: < 1.5W/G

28
 200G
LPUF-200 LPUI-200

LPUI-200(2*100GE) LPUI-200(20*10GE)

1*100 GE CFP2 10*10GE SFP+ LPUI-200(1*100GE+10*10GE)

Attribute Description
 4M IPv4 FIB Entries or 2M IPv6 FIB Entries (Shared)
Optical type supported 10Gbps SFP+ optical module  2M MAC
10Gbps SFP+ BIDI optical module
 256K ARP and 64K ND (Separated)
100Gbps CFP2 optical module
 256K ACL or 32K ACLv6 (Shared)
Working mode Full-duplex  Packet Buffer: 100ms
Compliant standard IEEE 802.3  Power Consumption: < 1.5W/G

*Only supported in NE40E-X3A/X8A/X16A, not supported in NE40E-X3/X8/X16.

29
NE40E 120G Line Card

LPUF 1*100G 12*10G

6*10G 1*40G 8*2.5GPOS 24*1GE
LPUF-120 Only one half-width sub Only one half-width sub
card allowed card allowed

LPUI
LPUI-120 10*10G LPUI-120 2*40G LPUI-120 1*100G LPUI-120 12*10G

Attribute Description
Optical type supported 1.25Gbps eSFP optical module  4M IPv4 FIB Entries or 2M IPv6 FIB Entries (Shared)
1.25Gbps eSFP BIDI optical module  2M MAC
10Gbps SFP+ optical module  256K ARP and 64K ND (Separated)
10Gbps SFP+ BIDI optical module  128K ACL or 16K ACLv6 (Shared)
100Gbps CFP2 optical module
 Packet Buffer: 200ms
Working mode Full-duplex
 Power Consumption: < 1.7W/G
Compliant standard IEEE 802.3

30
 50G

LPUF
24*1GE 2*10GE 5*10GE 20*GE RJ45
LPUF-51-E 2-Port POS
(occupying 2 sub-slots)

LPUI
LPUI-52-E 5*10GE LPUI-52-E 24*1GE+2*10GE LPUI-51-E 48*1GE

Attribute Description
Optical type supported 10Gbps XFP optical module  4M IPv4 FIB Entries or 2M IPv6 FIB Entries (Shared)
1Gbps electrical module
1.25Gbps eSFP optical module  2M MAC
1.25Gbps eSFP BIDI optical module  256K ARP and 64K ND (Separated)
10Gbps SFP+ optical module
10Gbps SFP+ BIDI optical module  64K ACL or 16K ACLv6 (Shared)
Working mode Full-duplex  Packet Buffer: 200ms
Compliant standard IEEE 802.3

31
 50G (LPUF-50, 4 Flexible Sub-card)

8*155M POS 4*155M POS 2*155M POS 8*622M POS

POS

The LPUF-50
1*10G POS
Port sub card (occupying two sub-slots) 2*2.5G POS 4*2.5G POS

Ethernet 8GE/2*10GE

2*155M/4*155M/8*155M
POS 8*622M CPOS
2*2.5G/4*2.5G/1*10G 8*155M CPOS 4*155M CPOS 2*155M CPOS

2*155M
CPOS 4*155M
8*155M
ETH E1
E1 24-port E1
P10-A 8*1GE P50-A 2*10G 24-port E1

32
 VSUF-160/VSUF-80 CGN IPsec

Versatile Service Unit Flexible Cards (VSU) Industry Leading CGN and IPsec Specification

Item VSUF-80 VSUF-160

VSUF-80
80 Gbps @1024B 160 Gbps @1024B
CGN
50 Gbps @512B 100 Gbps @512B
Throughput
(with 1 sub card and 1 license) (with 2 sub cards and 2 license)

CGN 32M 64M

Session Count (fully configured sub cards) (fully configured sub cards)
SP-80
CGN Flow
1 M/s 2 M/s
Creation Speed
VSUF-160 16Gbps @512B 32Gbps @512B
IPsec 25.6Gbps @1024B 51.2Gbps @1024B
Throughput 32G @1400B 64G @1400B
(with SP80) (with 2x SP160)

IPsec
20000 20000
Tunnel Count
SP-160 Ipsec Tunnel 200 tunnel/s 200 tunnel/s
Creation Speed (with SP80 under typical scenario) (under typical scenario)

33
И breakout
Breakout Box Breakout Cable

 High density fiber scenario  Loose fiber scenario

 Long fiber scenario  Short fiber scenario

Breakout Box
MPO-LC Breakout Cable

MPO-MPO
Patch Cord

Only 100G optical Module needed *100G and 10G modules needed

 A breakout box supports the 4 * 100G interface for  A breakout fiber supports a 1 * 100GE interface for a 10
the 40 * 10GE interface * 10GE interface
 Up to 2 breakout boxes per board  Up to eight breakout fibers per board
 Patch cord supports 10m / 30m length  The length of the breakout fiber is 15m
34
Agenda

 BNG


 IP CORE
 DCI

35
 BNG

BNG

Version Evolution

V8R9C00: 1M session per chassis hot backup

480G

V8R9C10: up to 128k session per slot 240G

100G

40G
20G
10G
16 GB memory, a maximum of 1M users per chassis
…
2007 2008 2009 2012 2013

Solar 1.0 Solar 2.0 Solar 3.0 Number of

Solar chip 65 nm technique
0.13 um technique 40 nm technique
evolution 40G throughput
100G throughput 480G throughput transistors
400,000,000+

36
 CGN

И CGN NE40E

VSUF-160+
160G 2xLicense+2
xSP160

VSUF-160+
120G
2xLicense+
SP160

VSUF-80+ VSUF-160+
80G License+ 2xLicense
SP80
60G VSUF-160+
License

40G VSUF-80+
VSUF-160
License

20G
VSUF-80

VSUF-80 Flexible and configurable capacity, adapting to different phases VSUF-160

Item VSUF-80 VSUF-160

80 Gbps @1024B (with 1 sub card and 1 capacity license) 160 Gbps @1024B (with 2 sub cards and 2 capacity license)
Throughput
50 Gbps @512B (with 1 sub card and 1 capacity license) 100 Gbps @512B (with 2 sub cards and 2 capacity license)

Session count 32M (fully configured sub cards) 64M (fully configured sub cards)

Flow creation speed 1 M/s 2 M/s

37
 CGN

CGN NE40E

1:1 hot standby(board) Load balance(board) Load balance(chassis) 1:1 hot standby(chassis)

BNG+ BNG+ BNG+ BNG+

C C C C CGN board CGN board CGN board CGN board
G
Session G G G
N backup N N N
1 2 1 2
Backup
Switch fabric Switch fabric channel

Line Line
card card
OLT OLT

• Abundant high reliability technology ensure CGN service

• CGN with BRAS implement 50ms switchover，user service no interrupt

38
 IPsec

И IPsec NE40E
Appearance Performance Name Description
It has two sub card slots and can be inserted into the
VSUF-160 32G Flexible Card Versatile Service Unit (VSUF-160)
CX600/NE40E/ME60.
It has one sub card slot and can be inserted into the
VSUF-80 16G Flexible Card Versatile Service Unit (VSUF-80)
CX600/NE40E/ME60.

SP-160/80 16G Value-added service processing unit It can be used as a sub card for the VSUF-160/VSUF-80.

VSUF-160 2*SP-160
VSUF-160 SP-160
64G
SP-80
VSUF-160
48G
VSUF-80 OR

VSUF-80
32G

16G

39
Agenda

 BNG


 IP CORE
 DCI

Г я

40
 Seamless MPLS
Seamless MPLS inter-domain/AS HVPN

The RSG and ASBR can be merged into one equipment

L3VPN-LDP/RSVP TE L3VPN -Seamless MPLS LSP

IPRAN IP Core

AGG1 AS 2 RSG2 ASBR1 AS 1 PE1

PE

AGG2 ASBR2
RSG2 PE2

Seamless MPLS solution with high reliability

Seamless MPLS Base Line •OSPF Remote LFA • Load-Balance : BGP LSP load-balance on • X2/eX2 automatic deployment :
NE40E •RFC3107 •X2 interconnection through two TE LSP Big FIB(128K) and LSP(32K) for
•VPN o BGP LSP o TE/LDP default route •
Roadmap •Labeled BGP LSP FRR •VPN next hop separation quick
Simplification : Remove BFD for BGP LSP 10K CSG zero configuration and
through Tail-end Protection self-connection
•BFD for BGP LSP trigger VPN FRR brush

V800R006C00(14Q3) V800R008C00 (16Q1) V800R009C00(17Q2) V800R009C10 (17Q4)

41
 Segment Routing

MPLS Segment Routing

MPLS rethinks itself

Then why not cut off control plane ?

• What is essence? Forwarding Plane: Label Switch

• Who is bothering? Control Plane: LDP/RSVP Segment Routing

Segment Routing Network

P1 P2 PE2
IGP
ECMP
IGP  LDP/RSVP Control Plane Abandoned
Label
IGP allocated and Distribute label information via IGP
IGP
IGP propagated by
IGP PE3 IGP  MPLS Forwarding Plane Remained
PE1
IGP
Simple operation of label (push, swap and pop)
IGP
P3 P4  Traffic Engineering
Cooperate with SDN controller to provide traffic engineering
IGP has already known topology and bandwidth of the network
IGP could also distribute label information to replace LDP/RSVP

42
 Segment Routing

Ч Segment Routing?
1 2
Maintenance Simplified Easy to Extension

•Only IGP on Control Plane •Label Number Comparison: Node + Adjacency

•No Multi Protocols Inter-working (Segment Routing) vs Node ^2
•Shortest Path and ECMP Naturally Supported •Forwarding State only on Ingress Node
•No LDP Mapping or RSVP Soft Refresh

3 4
Fast Convergence Better SDN Integration
•SDN-based SR-TE Solution to Optimize Network
•Ti-LFA Protection in Any Topology
Bandwidth Utilization
•Automatic Backup Path Calculation and Install by IGP
•No Need to re-establish LSP in case of Path Change,
•Micro Loop Avoidance
Label Stack = Path

Basic functionality for Tecent： • Enhancement： • OSPF for SR-BE/SR-TE

• PCECC for SR-TE • SR-BE • OSPF for SR-BE TI-LFA
NE40E
• SID stitching • ISIS adjacency Segment • SR and LDP interworking(ISIS)
Roadmap • HSB for SR-TE • ISIS for IPV6 SR-TE
• SRGB
• TI-LFA • ISIS for IPV6 SR-TE HSB

V800R008C10 (16Q4) V800R009C10 (17Q4) V800R010C00(18Q2)

43
 IPFPM

IP FPM
Innovative
Industry's Only Performance IPTool
Monitoring FPM measurement
Based on Actual IP service flows uTraffic – E2E Performance Monitoring Analysis Tool

Packets colored  GIS: Automatically locates NEs and displays the NEs on a pre-
imported map.
 Traffic visualization: Visualizes bandwidth, traffic direction, and
traffic utilization and monitors service quality (packet loss rate,
delay, and jitter).
 Customized GUI: Map, monitored NEs, alarm thresholds,
monitored time range, and server locations
 One-click release of analysis reports for offline analysis
Actual service flow

Traffic/Utilization

70M/80% 1G/ 10%

BT

Actual service traffic Fast fault locating 10ms/0%

S
10ms/0%
Monitors actual service flows, Applicable to all service Automatic hop-by-hop fault
with 10-6 measurement locating in E2E mode RN
scenarios 10ms/0% Delay/PLR
precision
P2MP and MP2MP C

ECMP

44
Agenda

 BNG


 IP CORE
 DCI

45
 PCE+
PCE+/SDN-based TE
Traffic calendar Maintenance window Traffic Monitor
Explicit Path Path Optimization
Deployment
Service Simulation
Application Traffic Optimization

Agile Controller

BGP-LS/PCEP BGP-LS/PCEP BGP-LS/PCEP BGP-LS/PCEP BGP-LS/PCEP

Backhaul Metro IP Core Metro Backhaul

PCE+ Solution is Huawei’s comprehensive solution to optimize Tunnel path based on flexible requirement which is
based on SDN architecture.
PCE+ solution is mainly based on PCEP standards and also relies on other protocols such as BGP-LS, SNMP, etc.
PCE+ solution includes PCE,PCC, Traffic Analysis and other related elements
NE40E acts as PCC element

Global Real time Centralized Centralized Policy driven Customized

topology Synch computation provisioning optimization ….
46
 Seamless MPLS

Seamless MPLS IP CORE

Backbone
Region IBGP Region
IBGP IBGP

OSPF area X OSPF area 0 OSPF area Y

Edge-PE ABR ABR Service-PE
P P
(L/M) (XL) (XL) (L/M)

MPLS
E2E Service Provisioning
BGP_L BGP_L BGP_L
LSP_L LSP_L LSP_L LSP_L LSP_L

 Adopt Seamless MPLS architecture to build an E2E MPLS transport layer which support wide variety of different services on a
single MPLS platform fully integrating region and backbone networks, lays a solid fundamental infrastructure future oriented.
 Unified MPLS forwarding, BGP signaling
 In order to achieve better scalability and faster convergence, it is better to partition the network into several independent and
isolated IGP domains. IGP is used for intra-area connectivity only. No redistribution occurs between different IGP areas
 Inter-area connectivity is realized by using (labeled) BGP.

47
 Load Balancing

Tunnel (LSP) Load Balancing

Route Load Balancing Trunk (LAG) Load Balancing

 The number of load balancing is configurable, NE40E supports:  Tunnel table will handle the relationship between next-hops and
 Maximum 64 paths for no-recursive load balance like IGP; out interfaces. The out interfaces are not only include physical
 Maximum 256 paths for recursive load balance like BGP; interfaces, but also some logical interfaces such as Trunk

 Maximum 64 member ports in a trunk/LAG interface and 256 interfaces, TE tunnel interfaces, LSP, even LDP over TE tunnels.
trunk/LAG interface per chassis.  If the out interface is Trunk interface, there will be a second stage
load balancing between different member ports of a Trunk.

Huawei NE40E can provide load balancing using various Hash Keys and Hash Algorithm

48
Agenda

 BNG


 IP CORE
 DCI
Г я

49
 EVPN

EVPN – Next Generation Solution for Ethernet Service

Cloud DC era, VPLS has not adapted to the needs of cloud DCI

VPLS DCI-PE DCI-PE EVPN DCI-PE DCI-PE

DC-GW
ES1
DC-GW DC-GW DC-GW

extended
single-active ALL-active community

DCI-PE DCI-PE DCI-PE DCI-PE DC-GW

Data plane Data plane Data plane Data plane Control plane Data plane
Source MAC learning Source MAC learning Source MAC learning Source MAC learning MAC transmitted by BGP Source MAC learning
Same principal as L3VPN

 BUM flooding  Single-active  NO MAC Mobility

takes up a lot of bandwidth Link load is not balanced low efficiency to refresh MAC

50
 EVPN

EVPN Three Typical Applications

Load Balance Fast Convergence MAC Mobility
2
PE2 1 PE2 CE2
PE2
CE1 PE1 BGP –EVPN
（E-VPN Ethernet A-D Route CE2 CE1 PE1 BGP Withdraw CE1 PE1 BGP –EVPN
（E-VPN Ethernet A-D Route CE2 MAC Advertisement Route

ECMP PE3 PE3 PE3 CE3

ES ES

V8R7C00 V8R8C10 V8R8C11 V8R9C10 V8R10C00 (planning)

• EVPN E-LAN • PBB-EVPN E-LAN • EVPN IRB • Option C • EVPN E-LINE
• VLAN into EVPN • EVPN E-TREE
• VXLAN into EVPN • VLL into EVPN E-LAN
• VLL into EVPN E-LINE
* 5883 board support EVPN, such as LPUF/I-51-E, LPUF/I-120/120-E,LPUF/I-240, LPUF/I-480 and so on . • Option B
51
Agenda

52
YANG я
YANG
Challenges Flexible, Readable and Common Values
Data Model
NETCONF YANG Drives
HW Devices HW Devices
NETCONF YANG
Easy to operate
NETCONF configuration and YANG

Cisco Devices Network state data Cisco Devices Network

SNMP YANG
AL Devices Good readability ALL Devices

1 Different management systems/tools 1 Same management systems/tools

2 Hard to analyze, due to mass data Used in all layers 2 Easy to analyze using common tools
3 Not convenient to accelerate automatic of NETCONF 3 Convenient to accelerate automatic
process Protocol process
4 Obscure codes Framework 4 Friendly codes

53
 Huawei U2000

Operator
U2000

CE
PE PE
CE CPE

Simplified Service O&M Simplified Performance Collection Fast Locating of Massive Alarms
Visualization: During E2E management using the U2000, Engineering alarm marking: At the engineering
service creation and management are displayed in GUI, implementation stage, engineering alarms are marked for you
allowing O&M engineers to be more aware of services and Simplified management: Task-specific performance data
to focus on valid alarms.
lowering the technical requirements of O&M engineers. collection, wizard task creation, template-based
management of performance indicators, and periodical and Analysis of repeated intermittent alarms: Set a period of
High efficiency: The U2000 provides various batch service time during which repeated alarms are filtered out and only
batch task execution. A maximum of 150,000 performance
processing capabilities depending on service deployment one alarm is displayed.
collection tasks can be executed, simplifying network O&M.
characteristics, accelerating service launching. Alarm correlation analysis: The root alarm can be quickly
Facilitated browsing: display of TCA, historical
Integration: The alarm, performance, and O&M located among a large number of alarm records.
performance, and real-time performance data for each
capabilities are integrated in the E2E service management Alarm merging: Only one alarm is displayed on the U2000 for
performance collection task.
window, the alarms generated due to the same symptom, but the time
serving as a good platform for service-oriented O&M. and number of times for alarm occurrences are recorded.

54