Networking and Planning
Contents 1.
Network Building Requirements
2.
Networking and Planning
3.
Experimental Network Planning Examples
Contents 1.
Network Building Requirements
2.
Networking and Planning
3.
Experimental Network Planning Examples
Changes from 3G Network to LTE RAN
The EPC+eNB without RNCs is adopted, making the network structure flattened. The network evolves into an all IP network, with the uplink and downlink rates increasing greatly. The common rate is 30 Mbit/s and the maximum rate is 50 Mbit/s on a 3G network. The required bandwidth bandwidth is about 150 Mbit/s on an LTE network. The eNB capabilities have been boosted obviously in comparison with 2G/3G base stations. Dynamic connections need to be configured for interfaces on eNBs.
Service Requirements Service control layer RNC
2G/3G base stations
SR
Termination layer
VLAN IDs on the PON network
10
Backbone layer
OTN network
OLT
FTTH user
OLT
FTTH user
11
Distribution layer
10GE
OLT
FTTH user
QinQ
Key account services
10
Access layer
FTTH user
GE
Enterprise
NodeB
VPLS
10 11
NodeB
NodeB
10
Contents 1.
Network Building Requirements
2.
Networking and Planning
3.
Experimental Network Planning Examples
Hierarchical Architecture of METRO-E R FE
R
Access layer
NodeB
R
FE NodeB
PWE3
R R
GE POS
R
Distribution layer
Access layer R
RNC
Core layer R
R
GE POS
R
BTS
Access layer: accesses services from base stations. Numerous network nodes exist at this layer and the bandwidth pressure is small. Recommended network mode: ring, chain, or dual-uplink network Distribution layer: converges traffic and ports, with powerful dynamic scheduling capability. Many network nodes exist at this layer and the bandwidth pressure is relatively large. Recommended network mode: ring or dual-uplink network Core layer: accesses traffic from the distribution layer. It serves as the service system gateway and schedules entire traffic comprehensively. A few network nodes exist at this layer and the bandwidth pressure is large. Recommended network mode: dual-uplink, mesh, or rectangle-shape network
RNC
Network Topology Planning for the Access Layer of METRO-E
R845
R845
R860
R860
Dual-homing network
Ring network
R845
R860 R860
R845
Ring and chain network
Chain network
Staged Bandwidth Requirement Calculation for the Access Layer Assume that each access chain or ring contains 10 access points and each 3G node provides the access service for 3000-5000 users. It can be calculated that the capacity of each NodeB is 30 Mbit/s and will be expanded to 50 Mbit/s at the later stage based on the CS 16.4 kbit/s, CS 64 kbit/s, PS, and relevant overheads. The analysis is as follows:
Service Type
3G NodeB
2G BTS
Private Line (AG)
Ethernet
Capacity
30 (50) Mbit/s
4 Mbit/s
20 Mbit/s
30 Mbit/s
Interface
1FE+2E1s
2E1s
1FE
1FE
Convergence
N/A
N/A
N/A
50%
Early stage: Each access point is connected to only one 3G NodeB. The bandwidth usage is 30%
(that is, 30 x 10/GE) and the interfaces are 1FE+2E1s. Middle stage: One 3G NodeB, one 2G BTS (4 Mbi t/s), and 2 private lines (or NGN AGs) will be connected. The bandwidth usage is 74%, that is, (30M + 4 + 20 x 2) x 10/1 GE, and the interfaces are 3FEs+4E1s. Later stage: The network will be expanded to a 10G ring network, which covers four 3G NodeBs, two 2G BTSs, three private line services, and three Ethernet services. The bandwidth usage is 33%, that is, (50 x 4 + 2 x 4 + 3 x 20 + 3 x 30M x 50%) x 10/10G, and the interfaces are 10 FEs+12E1s.
Network Topology Planning for the Distribution Layer of METRO-E
Dual-homed core devices will be used for network building at the later stage if the capital resource is sufficient. The purpose is to reduce the network load and enhance network security.
Staged Bandwidth Requirement Calculation for the Distribution Layer Assume that each convergence ring accommodates a maximum of six nodes. Service
Stage I
Convergence Ratio
Total Bandwidth (Gbit/s)
GE ring
10
1
0.5
5
Ethernet private line (EPL)
FE
6
0.1
0.5
0.3
OLT
2 * GE
6
2
0.1
1.2
Total bandwidth
6.5
Quantity
Bandwidth (Gbit/s)
Convergence Ratio
Total Bandwidth (Gbit/s)
Access ring
GE ring
10
1
0.5
5
EPL
FE
12
0.1
0.5
0.6
OLT
2 * GE
12
2
Service Stage 3
Bandwidth (Gbit/s)
Access ring
Service Stage 2
Quantity
Access ring
10GE ring
0.1
2.4
Total bandwidth
8
Quantity
Bandwidth (Gbit/s)
Convergence Ratio
Total Bandwidth (Gbit/s)
4
10
0.5
20
EPL
FE
12
0.1
0.5
0.6
OLT
10GE
12
10
0.1
12
Total bandwidth
32.6
Bandwidth usage: 65%
Bandwidth usage: 80%
Bandwidth usage: 50%
Network Topology Planning for the Core Layer
RNC
Dual-homing network RNC
Ring
RNC
MESH The ring networking mode is recommended at the initial stage. The network can be upgraded to a mesh network based on the optical fiber laying status to enhance the network robustness and security. It is recommended that the distribution layer and core layer be integrated into one layer to form a mesh or rectangle-shape network if conditions permit.
Basic Principles for IP Address Planning Uniqueness
Continuity The counterclockwise allocation mode and the principle of allocating IP addresses from the core layer to the access layer ensure the continuity and aggregation of IP addresses.
Meaningfulness IP address allocation conforms to certain principles and useful information can be obtained from IP addresses.
No address can be duplicated with other addresses on the same network.
Expansibility Certain addresses need to be reserved for future device expansion.
Economization IP addresses must be fully used based on the minimum use principle to avoid wastes. For example, interface interconnection addresses can use 30-bit mask addresses.
IP Address Classification in Network Planning
Loopback addresses The system administrator creates one loopback interface for each router and allocates a separate IP address for the interface as the management address to facilitate management.
Interconnection addresses Interconnection addresses refer to the addresses used by interfaces for connecting two or more network devices.
Service addresses Service addresses refer to the addresses used by connected servers and hosts on the Ethernet and gateway addresses.
Example of Device Address Allocation (Loopback Addresses)
Allocate IP addresses to devices based on the network hierarchy, for example, allocate IP addresses to devices at the core layer, distribution layer, and access layer from small to large. Allocate addresses by ring number (ring 1, ring 2, ...) and allocate addresses in counterclockwise direction in rings. Adopt the principle of rings first and then chains. Make reservations during address allocation.
Use 32-bit masks for device addresses.
In principle, device addresses are determined during network design planning.
10.229.0.1/32
R1
R4
10.229.0.4/32
Core layer 10.229.0.2/32
R2
R3
10.229.0.3/32
Distribution layer
10.229.1.1/32
10.229.2.1/32
R5
R6
R7
10.229.1.2/32
R10 10.229.2.4/32
Access layer 10.229.2.2/32
R8
R9 10.229.2.3/32
Example of Interconnection Address Allocation
Allocate interconnection addresses based
10.254.0.1/30
addresses to interconnection interfaces
10.254.0.2/30
direction in rings. Adopt the principle of rings first and then chains.
Make reservations during address
R4 10.254.0.10/30
Core layer R3
10.254.2.1/30
10.254.2.10/30
Distribution layer
10.254.2.2/30
R5
addresses.
10.254.2.9/30 10.254.8.18/30 10.254.8.17/30
10.254.8.2/30
R7
R10
10.254.8.5/30
10.254.8.14/30
Access layer
10.254.8.6/30
Use 30-bit masks for interconnection
R6
10.254.8.1/30
allocation.
10.254.0.9/30
R2
Allocate addresses by ring number and allocate addresses in counterclockwise
10.254.0.13/30
R1
on the network hierarchy and allocate IP
from small to large.
10.254.0.14/30
10.254.8.13/30
R8
10.254.8.9/30
R9 10.254.8.10/30
Example of Service Address Allocation RNC
RNC addresses and NodeB
172.21.202.5/30
R1
172.21.202.6/30
R4
addresses must be in different IP
Core layer
address network segments in the R2
service address allocation.
Use 30-bit masks for service
Distribution layer
addresses.
R3
R5
R6
In principle, service addresses are provided by the service side.
R7
R10
Access layer 172.21.209.26/30
R8
172.21.209.25/30
172.21.209.30/30
R9 172.21.209.29/30
Node B
Service Application Provisioning and Planning Service Planning
Service IP Address Planning
—
IP: 20.1.1.1/29
L3 throughout the network
IP: 10.1.1.6/30 IP: 10.1.1.10/30
IP: 10.1.1.2/30 IP: 10.1.1.5/30 IP: 10.1.1.1/30
IP: 10.1.1.9/30
Each base station uses one independent network segment
IP address planning: Allocate IP addresses by ring and follow the principle of rings first and then chains. Allocate IP addresses in counterclockwise direction in rings and adopt the mode of odds up and evens down, odds on the left and evens on the right for address allocation in rings. Increase IP addresses from the near to the distant in tributary chains. Use 30-bit masks for IP addresses of ports (minimum subnet). Make reservations during address allocation. The principles of IP address allocation in a single ring are as follows: Allocate address blocks to loopback interfaces. Allocate 30-bit IP addresses in counterclockwise direction in each ring. Ensure the continuity of IP addresses for route convergence under address conservation. Allocate at least one network segment and two IP addresses to each base station.
Example of IP Address Allocation in NE Management R2
R1
12.2.254.1
12.1.254.1
Core layer R4
R3
R5 12.4.254.1
12.3.254.1
12.5.254.1
R9
Distribution layer
12.6.254.1
12.9.254.1
R10
12.8.254.1
R7 12.10.254.1 12.7.254.1
R12 12.6.1.1
Access layer
R8
R6
R14
12.8.3.1
12.6.2.1
R20 12.10.3.1
R18
12.6.3.1
R13
12.11.254.1
R17 R16
R11
12.8.2.1
12.10.1.1
R15
R19 12.10.2.1
12.8.1.1
IP Routing Protocol Planning Principles Objective: To make the network hierarchy clear for network convergence and ease of network operation.
ISIS 1000 ISIS 1 ISIS 100
ISIS160
The core layer and distribution layer are configured as the backbone area (L2 router). Each ring at the access layer is configured as an AS domain.
SMART CEN PROJECT
BTS Services Using TDM Access layer
End to End PWE3
—
Distribution layer
Convergence at the backbone layer
E1 BSC
BTS
GE
10GE
E1 TDM Data
Service bearer
L2VPN Ethernet Header Tunnel Label PW Label Control Word RTP Header (optional) TDM Data
OAM Protection technologies
MPLS-TP OAM LSP1:1 + PW FRR
TDM Data
Service Bearer on Ethernet NodeBs 3G NodeBs —
Access layer
Distribution layer
Convergence at the backbone layer
FE RNC GE
NodeB
10GE
FE
L3VPN Service bearer
L3VPN Payload
Payload
Ethernet header
VRF label VP label Ethernet header
Protection technologies
VPN FRR+LDP FRR/CRLSP 1:1+(TE FRR)
VPN FRR+LDP FRR/ CR-LSP 1:1+(TE FRR)
VRRP/IP FRR
LSP 1:1 Protection Active path
Standby path
Subnet
Subnet a) Normal working status
Active path Subnet
Standby path
Subnet b) Fault status
Services are transmitted through the active LSP in normal cases. When the active LSP is faulty, services are switched to the standby LSP for transmission. BFD or MPLS OAM is used as the fault detection mechanism. LSP protection can be understood as one group of bidirectional protection composed of two groups of unidirectional protection.
PW Protection BFD is used to detect PW faults quickly to implement OAM mapping between PWs and ACs. In this way, when a PW or PE is faulty, CEs can switch services to the standby path to enable end-to-end fault detection for the PW and implement PW backup, greatly enhancing the reliability of the L2 VPN.
Active Backbone network
Standby
Contents 1.
Network Building Requirements
2.
Networking and Planning
3.
Experimental Network Planning Examples
Network Topology Planning Experimental Network Shiju 2 Gaode
Topology of the
—
Yingbinju 2
Xiyuan 2
Xinqiu
Xiyuan 1 Shiju 1
Qinghemen
Yingbinju 1
South ring 10GE
North Central 10GE dongyuan 2 Shibei
Dianchang Shixian ring 10GE
Zhanqian Dual-node Dual-node Fuxinxian Zhangwu ring ring
A total of 19 ring systems are built in this project, including three 10GE rings at the distribution layer and 16 GE rings at the access layer. Some tributary chains are built and no more than 2 nodes exist on each tributary chain.
Networking idea for the core distribution layer: Three core equipment rooms and nine convergence equipment rooms form three 1 0GE core convergence rings, to improve the bandwidth usage of rings and enhance network security.
Networking idea for the access layer: The access layer is connected to the core convergence rings in dual-uplink mode to enhance the network security. Reorganize the network if the line routes and logic structure are unreasonable.
Quantity
Core Distribution Layer CiTRANS R860
Access Layer CiTRANS R845
17
114
IP Address and Routing Protocol Planning Allocation
IP Address
—
Configuration requirements of service IP addresses: 1. Service-side IP addresses of devices are allocated by the access side system. 2. It is recommended that 30-bit mask addresses be used as interconnection addresses. 3. Loopback addresses are allocated by carriers.
City City A
IP Address
Number of IP
Number of Physical Sites
Segment
Addresses
on the Live Network
10.229.0.0/17
32768
391
Allocate one address segment 10.229.0.0/17 to loopback interfaces on network devices in the Fuxin office. Allocate the smallest four address segments 10.229.0.0/22 to the core layer and distribution layer. Use the IP addresses 10.229.0.0/30 to 10.229.1.255/30. Reserve the IP addresses 10.229.2.0/30 to 10.229.3.255/30. Allocate four address segments to each converged access device. Use the IP addresses 10.229.4.0/22 to 10.229.124.0/22. Reserve the IP addresses 10.229.125.0 to 10.229.127.0.
IP Address and Routing Protocol Planning Example —
Node B1
B
Shiju
IP Address Planning
C
Core distribution layer: Allocation: 10.229.0.0-10.229.1.255/30 Reservation: 10.229.2.0-10.229.3.255 Loopback address: 10.229.0.0/32 Yingbinju office: 10.229.0.1/32 Xiyuan office: 10.229.0.2/32 Qinghemen office: 10.229.0.3/32 ...
Gaode Dianchang
Zhanqian
Xiyuan
Routing protocol planning
Access layer: 10.229.127.0/22 Loopback address: 10.229.62.0/2210.229.127.0/22 Access point A: 10.229.4.1/32 Access point B: 10.229.4.2/32 Access point C: 10.229.4.3/32 ...
RNC2 Yingbinju
Node B2
•
Allocation: 10.229.4.0/22-10.229.61.0/22 Reservation: 10.229.62.0/22-
RNC1 Xinqiu
•
A
IGP: OSPF
•
Domain division: core distribution layer: Area 0; access
layer: Area 1/2/3/4... Router ID/Cost/Priority/Loopback/CIDR
EGP: BGP: RSVP: LDP
•
Use loopback addresses to set up the BGP neighbor
relationship. Set RR on bridge nodes such as Xinqiu and Gaode.