Beruflich Dokumente
Kultur Dokumente
Web: http://5gopenplatform.org
1
Abstract
This document contains an overall description of 5G K-NetSim (5G Korea-Network Simulator) and how
to use it for users. There is a list of features of 5G K-NetSim, a structure, and parameter descriptions in
this document. 5G K-NetSim has been developed with a modular and flexible architecture, users can
easily adjust the simulator to evaluate their algorithms or functions. This simulator is released under an
academic, non-commercial use license.
2
Contents
1. Introduction ·········································································································· 7
1.1 Purpose ········································································································· 7
1.2 License ········································································································· 7
1.3 Contact points ································································································· 7
1.4 Development environment ·················································································· 8
1.5 Download ······································································································ 8
1.6 Definition, symbols and abbreviations ···································································· 8
1.6.1 Definition ································································································· 8
1.6.2 Symbols ··································································································· 8
1.6.3 Abbreviations ····························································································· 8
2. List of features ·····································································································10
2.1 Performance metrics ·························································································10
2.2 Simulation parameters for configuration ·································································10
2.2.1 General parameters ·····················································································10
2.2.2 5G NR NSA ·····························································································11
2.2.3 5G NR Channel ·························································································11
2.2.4 5G Core ··································································································12
3. Network Simulator Structure·····················································································13
3.1 Overview ······································································································13
3.2 Introduction to ns-3 ··························································································13
3.3 Directory structure ···························································································14
3.4 Simulation methodology ····················································································14
3.5 Main components in 5G K-SimNet ·······································································15
3.5.1 Dual radio protocol architecture ······································································16
3.5.2 NR radio protocol ·······················································································17
3.5.3 SDN/NFV ································································································17
3.5.4 5G Core
4. Parameter description ·····························································································18
4.1 General parameters ··························································································18
4.2 5G NR NSA ··································································································19
4.3 5G NR SA ····································································································20
4.4 SDN/NFV ·····································································································23
5. Test scenarios ······································································································26
5.1 Overview ······································································································26
5.2 Parameter setting ·····························································································27
5.3 Results ·········································································································29
3
6. How to use ··········································································································31
6.1 How to install 5G NR NSA simulator ····································································31
6.2 Workspace ····································································································31
6.3 How to install 5G SDN/NFV simulator ··································································32
6.4 How to run your own simulation ··········································································34
7. Change Log ·········································································································36
8. References ··········································································································37
4
List of Figures
5
List of Tables
6
1. Introduction
This document describes 5G network simulator. 5G-K network simulator is implemented based on ns-3
network simulator. It is hard to develop the new network simulator because network simulator should
contain all the protocol stacks with related to end-to-end communication. Even if we develop the new
network simulator, it is difficult to prove the reliability of our network simulator because of massive
characteristics of network simulator. Hence, we develop the 5G-K network simulator based on ns-3
simulator. ns-3 network simulator is one of representative network simulators in communication and
computer network research. ns-3 models network entities and structs complete set of the network protocol
stack, so it can predict network behaviors more accurately. However, conventional ns-3 simulator lacks of
consideration about virtualization effects or cutting edge of SDN technologies, or support of dual
connectivity of LTE and 5G networks. For simulating 5G network system, it is necessary to see virtualized
effects or SDN functions, or test dual connectivity scenario for deployment of 5G network. Our 5G-K
network simulator includes OpenFlow controller, OpenFlow switches, modules for evaluating
virtualization effects and dual connectivity of LTE and 5G network. Users who want to simulator 5G
network system can evaluate their own network system topology with our 5G-K network simulator.
1.1 Purpose
This document is intended for use by software engineers working directly on the network simulator. In
the paper we explain about two main components of the simulator (DC and SDN/NFV), parameters and
metrics of the simulator, design considerations, simulator architecture, and how to use the simulator.
1.2 License
TBD
Tel: +82-42-350-5472
E-mail: jtyim@kaist.ac.kr
7
1.4 Development environment
OS: Linux Ubuntu 14.04
Language: C++ (gnuplot for GUI, some python codes in ns-3)
Compiler: Waf (python-based build system)
Setting: ns-3 version 3.26 based simulator (pre-configuration is required. Please refer to
https://www.nsnam.org/docs/release/3.26/tutorial/html/getting-started.html)
1.5 Download
Download source code from the 5G web site(http://5gopenplatform.org)
1.6.2 Symbol
Symbols Descriptions
1.6.3 Abbreviations
Abbreviations Descriptions
5G NR 5th generation new radio
DC dual connectivity
NSA non-standalone
SA standalone
SDN software-defined network
NFV network function virtualization
8
9
2. List of features
2.1 Performance metrics
User throughput
– User throughput is an important metric for network simulators, while it can show general
network performance for each user device.
– User RTT can show timely change of packet transfer performance. If RTT increase/decrease
sharply in short period of time, there might be special reasons such as VNF scaling, and so
on.
– With TCP congestion window size, user can verify the operation of network simulator and
changes of RTT or throughput.
Scaling delay
– With SDN controller and switches, user can engineer traffic flowing the simulation
networks. SDN switch throughput per port can show results of traffic engineering.
2.2.3 5G NR Channel
Operating band parameters (src/mmwave/model$ mmwave-phy-mac-common.cc)
(~/GigaKoreaProject/mmwave_new_handover/ns3-mmwave/src/mmwave/helper$ mmwave-
helper.cc)
– Scenario (m_scenario)
11
– Optional Nlos (m_optionNlosEnabled)
– Shadowing (m_shadowingEnabled)
– Tx power (m_txPower)
– TxPower (m_txPower)
12
UE antenna parameters (src/mmwave/model$ mmwave-ue-net-device.cc)
13
2.2.4 5G Core
AMF-related parameters (virt-5gc.cc, virt-5gc-node.cc, topology input file (user input))
14
– UPF bandwidth (Virt5gcNode::bwSize)
Data center parameters (virt-5gc-node.cc, virt-5gc-vm.cc, topology input file (user input) )
Node ID (Virt5gcNode::node_id)
Software-Defined Networking (SDN) parameters (User input file (scratch file), ovs-point-to-
point-epc-helper.cc, qos-controller.cc )
5G K-SimNet
src
lte
scratch
mmwave
virt-5gc
...
All the source codes are implemented in ‘src’ folder. For example, LTE module can be found in
‘src/lte’. Each module has sub-modules, helper and model. Protocols, algorithms, etc. are implemented
in ‘model’. Helper makes the configuration and installing interfaces easier.
16
Figure 2 Simulator procedures
Simulation will proceed as described in above figure, topology definition, model, configuration,
execution, and output analysis. In the first stage, topology definition, the nodes needed for experiment
and interfaces between nodes are defined. In model phase, protocol stacks, such as TCP, LTE and Wi-
Fi, and applications are installed in nodes. Attributes, environmental variables, and command line
arguments are configured in the following stage. The scratch code includes topology definition, model
and configuration. The simulation runs with command line execution in Linux OS. The experimental
results are logged in text format or specific format, e.g., Wireshark. You can study the result with
gnuplot, MATLAB, or Wireshark.
17
Figure 3 represents the simulator architecture of 5G K-SimNet. The 5G K-SimNet is being
expanded from ns-3, and 5G K-SimNet currently supports LTE-NR (New Radio) dual connectivity
(DC) and 5G core module including SDN/NFV (Software-Defined Network/Network Function
Virtualization) for end to end performance.
LTE-NR dual connectivity module supports one of 5G's main features to compensate for the
instability of NR link. To support dual connectivity, 5G K-SimNet remodified the existing LTE and
NR modules to develop a new LTE-NR dual module. Many layer layers have changed, but NAS,
PDCP, RRC, and RLC layers have been modified. Through the LTE-NR dual module, incoming
packets through the eNB are distributed by X2-inteface and are transmitted to the UE, are reordering.
5G core network functions consist of AMF, SMF and UPF. These functions are essential
components for simulating 5G data transmission. Control signal procedures between AMF, SMF and
UPF are implemented, while less important parts are omitted such as security, policy, and so on. And
SDN/NFV module is an essential function for the 5G core networks. The green modules of SDN/NFV
and Virtualization delay blocks are implemented for new 5G features. So far, the SDN / NFV module
has been developed based on the existing 4G EPC network, but the 5G core entity will be developed
and applied through continuous updating.
This model includes the LTE radio protocol stack and the 5G new radio (NR) protocol stack. These
entities reside entirely within the UE and the LTE eNB/NR eNB nodes. This model is based on dual
18
connectivity, providing the cooperating network architecture between 4G LTE and 5G NR in the
transition period from 4G to 5G. Figure 5 shows the block diagram of LTE-NR dual connectivity for
user plane in downlink case. The LTE eNB is deployed as a C-plane anchor node (master eNB, shortly
MeNB) and the NR eNB is deployed for boosting the user throughput or balancing load between LTE
eNB and NR eNB (secondary eNB, shortly SeNB). The downlink traffic splits at the PDCP entity at
LTE eNB and routing to either the RLC entity at LTE eNB or that at NR eNB. In order to enjoy the
dual connectivity, the traffic split function is deployed at MeNB. The traffic split entity is defined for
all traffic, i.e., split bearer, individually. The PDCP TX entity, split layer, should perform the packet
sequencing for the PDCP RX entity to re-order the split packets. The PDCP RX entity, aggregation
layer, should perform the packet reordering to guarantee in-sequence deliver of the received packet to
upper layer. We have developed the dual radio protocol stack, the packet sequencing, packet reordering,
and simple traffic split algorithm exploiting the LTE model, developed by LENA, and the mmWave
radio model, developed by New York university (NYU). We also have developed the radio protocol
and the channel model for NR operating at the frequency below 6 GHz, presented in next chapter.
In this scenario, we can evaluate the performance of the protocol for LTE-NR dual connectivity.
For example, the following features can be evaluated: traffic split/routing algorithm, RLC queue
management scheme, MAC scheduling algorithm, and so on. These algorithms and schemes can be
modified or developed by user. The simulation results are as follows: end-to-end performance, such
as TCP/UDP throughput, round trip time, and congestion window size, specific protocol performance,
such as PDCP packet drop rate, RLC queue size/delay, PHY SINR values, and so forth.
20
Figure 5. SDN/NFV simulation process
The simulator reflects such side effects by following simulation process of figure 6. First, users
should set their own simulation topology and parameters related to SDN/NFV operation. Once they
(user codes) are provided to the simulator, it runs in two ways, virtualization modules and SDN
modules.
Virtualization modules calculate virtualization-related delays such as scaling delay or provisioning
delay. First, simulation nodes are placed depending on the topology of user codes. Dynamic workloads
(Simulation timely changes of workloads) of core NFs (MME, P/S-GW) are generated by using static
workloads of parameters of user codes. VNF delays are calculated after configurating scaling
thresholds and analyzing the topology, VNF policy.
SDN modules also places simulation nodes first. They configure OpenFlow switches and a
controller. After then they run OpenFlow application such as QoS bandwidth controller. The final
simulation results come out by merging results of virtualization modules and SDN modules.
The 5G architecture is defined as service-based and the interaction between network functions is
represented in two ways. In this paper, we will describe 5G core architecture with reference point
representation.
21
NSSF AUSF N13 UDM
N14 N15
N1 N2 N4
UE (R)AN N3 UPF N6 DN
N9
As above figure shows, the 5G System architecture consists of the following network functions
(NF).
Authentication Server Function (AUSF), Core Access and Mobility Management Function (AMF),
Data network (DN), Unstructured Data Storage network function (UDSF), Network Exposure
Function (NEF), NF Repository Function (NRF), Network Slice Selection Function (NSSF), Policy
Control function (PCF), Session Management Function (SMF), Unified Data Management (UDM),
Unified Data Repository (UDR), User plane Function (UPF), Application Function (AF), User
Equipment (UE), (Radio) Access Network ((R)AN), 5G-Equipment Identity Register (5G-EIR)
However, for simulation purpose, our simulator includes very specific and important parts of NFs,
UE, RAN, AMF, SMF and UPF. And only AMF, SMF and UPF are the NFs in 5G core. Role of each
NF is defined as follows.
The Access and Mobility Management function (AMF) includes the following functionality. Some
or all of the AMF functionalities may be supported in a single instance of an AMF:
- Registration management.
- Connection management.
- Mobility Management.
The Session Management function (SMF) includes the following functionality. Some or all of the
SMF functionalities may be supported in a single instance of a SMF:
22
- Session Management e.g. Session establishment, modify and release, including tunnel
maintain between UPF and AN node.
- UE IP address allocation & management (incl optional Authorization).
- Selection and control of UP function.
- Configures traffic steering at UPF to route traffic to proper destination.
- Control and coordination of charging data collection at UPF.
- Downlink Data Notification.
The User plane function (UPF) includes the following functionality. Some or all of the UPF
functionalities may be supported in a single instance of a UPF:
- Anchor point for Intra-/Inter-RAT mobility (when applicable).
- External PDU session point of interconnect to Data Network.
- Packet routing & forwarding.
- Packet inspection and User plane part of Policy rule enforcement.
- Traffic usage reporting.
- Downlink packet buffering and downlink data notification triggering.
4. Parameter description
4.1 General parameters
X2 interface settings
– X2 link data rate: configure data rate of X2 link, link between base stations (eNB, gNB).
– S1AP link data rate: configure data rate of S1AP link, link between MME and base stations
23
(eNB, gNB).
– MTU of S1AP link: configure maximum transmission unit (byte) of S1AP link.
– S1-U link data rate: configure data rate of S1-U link, link between S-GW and base stations
(eNB, gNB).
– MTU of S1-U link: configure maximum transmission unit (byte) of S1-U link.
4.2 5G NR NSA
Type of dual connectivity: select dual connectivity alternative, 1A or 3C
– Alternative 1A
– Alternative 3C
{0, 20, 40, 60, 80, 100, 120, 140, 160, 180, 200, 220, 240, 260, 280, 300, 500, 750} ms
24
Type of splitting algorithms: for 3C, select way of traffic splitting.
– Splitting to single eNB: MeNB steers the traffic (PDCP PDU) to own RLC or SeNB’s RLC
– Alternative splitting scheme to two eNB: MeNB distributes PDCP PDU to own RLC or
SeNB’s RLC one by one.
– Assistant info-based splitting scheme: MeNB can calculate split ratio based on assistant
information, such as individual link performance and so on.
4.3 5G NR SA
Operating band parameters
– NrOperatingBand1: NR operating bands in FR1. For now, only TDD mode bands are
supported.
25
NR Uplink (UL) operating band Downlink (DL) operating Duplex Parameter
Operating BS receive band Mode Value
Band UE transmit BS transmit
UE receive
FUL_low – FUL_high FDL_low – FDL_high
n38 2570 MHz – 2620 MHz 2570 MHz – 2620 MHz TDD 38
n41 2496 MHz – 2690 MHz 2496 MHz – 2690 MHz TDD 41
n50 1432 MHz – 1517 MHz 1432 MHz – 1517 MHz TDD 50
n51 1427 MHz – 1432 MHz 1427 MHz – 1432 MHz TDD 51
n78 3300 MHz – 3800 MHz 3300 MHz – 3800 MHz TDD 78
n77 3300 MHz – 4200 MHz 3300 MHz – 4200 MHz TDD 77
n79 4400 MHz – 5000 MHz 4400 MHz – 5000 MHz TDD 79
Channel parameters
– Scenario: The available channel scenarios are 'RMa', 'UMa', 'UMi-StreetCanyon', 'InH-
OfficeMixed', 'InH-OfficeOpen', 'InH-ShoppingMall'.
– ChannelCondition: 'l' for LOS, 'n' for NLOS, 'a' for all.
– NumerologyIndex: In this B6NR module, multiple types of SCS are available. In the table
below, µ indicates this parameter. The numerology index from 0 to 2 represent 15kHz,
30kHz, 60kHz SCS each. Only from 0 to 2 value of index are supported in FR1. As shown
in the table below, the number of subframes per frame is fixed at 10, and the number of
slots per subframe is determined by this value, so the number of slots per frame is also
assigned. Additionally, OFDM symbol duration and cyclic prefix duration values are
calculated by the input value of numerology index.
[38.211] Table 3: Number of OFDM symbols per slot, slots per frame, and slots per
subframe for normal
26
µ slot
N symb frame,µ
N slot subframe,µ
N slot
0 14 10 1
1 14 20 2
2 14 40 4
3 14 80 8
4 14 160 16
5 14 320 32
– ChannelBandwidth: In the FR1 band, the channel bandwidth has various options as shown
below. The number of RBs is determined by SCS and channel bandwidth. The channel
bandwidth for each NR band is limited as shown in the 2nd table below.
SCS 5MHz 10MHz 15MHz 30 MHz 20 MHz 25 MHz 40 MHz 50MHz 60 MHz 70 MHz 80 MHz 90 MHz 100 MHz
[kHz]
NRB NRB NRB NRB NRB NRB NRB NRB NRB NRB NRB NRB NRB
15 25 52 79 [160] 106 133 216 270 N.A N.A N.A N.A N.A
30 11 24 38 [78] 51 65 106 133 162 [189] 217 [245] 273
60 N.A 11 18 [38] 24 31 51 65 79 [93] 107 [121] 135
– SymbolsPerSlot: Currently, this parameter has a fixed value of 14. In the next version, it
27
may include less OFDM symbols in the case of a mini-slot.
– SlotFormat: There are a lot of formats defined in 38.211. This would be similar to LTE-
TDD subframe configuration. In NR slot format, however, DL and UL assignment vary
along to the symbol level, not to subframe level and there are much more diverse formats
for various requirements. Currently, DL and UL symbols are supported and flexible symbol
needs to be further developed.
0 D D D D D D D D D D D D D D
1 U U U U U U U U U U U U U U
gNB-side parameters
UE-side parameters
Decide memory size of virtual machine(s) running AMF(s). Memory capacity affects
migration delay or scaling delay, and so on.
– AMF bandwidth
Decide link bandwidth of virtual machine(s) running AMF(s). Bandwidth affects almost
all virtualization delays such as scaling in/out delays.
– AMF workloads
28
Decide expected amount of computational, memory, disk, and bandwidth loads which
will be taken to the AMF(s).
– AMF x axis
– AMF y axis
– SMF bandwidth
– SMF workloads
– SMF x axis
– SMF y axis
– UPF bandwidth
– UPF workloads
– UPF x axis
– UPF y axis
Virtualization parameters
29
– Scaling policy (memory split ratio)
– VM provisioning delay
Decide the delay of provisioning new VM. This parameter can be replaced later by
introducing scheduling parameters.
Node ID
Node ToR ID
Describe ToR ID of each node. Each node is under ToR switches, and this parameter
indicates which ToR switch the node is connected.
Describe physical machine ID of each node. In the data center, a physical machine can
operate several virtual machines. In the simulation, node is a VM. This parameter
indicates which physical machine the VM node is operated.
Link bandwidth
30
– remote host-GW data rate
– In-port, out-port
– QoS weight
5. Test scenarios
5.1 Overview
31
Dual connectivity
– In the test scenario, we verify the setup procedure of LTE-NR dual connectivity, traffic
splitting in PDCP TX entity, and traffic merging in PDCP RX entity.
– To this end, we deploy a LTE eNB as master node, a NR gNB as secondary node, and UE
equipping the dual radio interface of LTE and NR.
– The PDCP TX entity splits the traffic based on split ratio calculated by using assistance
information from LTE and NR RLC entities, and routes split traffic onto RLC TX entities
in eNB and gNB. The PDCP TX entity allocates the sequence number to the spilt traffic in
order to the PDCP RX entity can merge and deliver the split traffic in sequence.
– The PDCP RX entity receives the data from LTE-NR dual inks and merges it in sequence
to deliver it to App layer.
5G NR
– In the test scenario, we check out the SNR values of received transport blocks (TBs)
whether the building exists or not.
– A gNB is fixed on the y-axis, and its height is 21.5 m and a UE moves at a speed of 60
km/h in a direction parallel to the y-axis.
– Downlink packets for the UE are continuously generated during the simulation time, which
is set to 6 s.
– We set analog beamforming vector update period of which value is denoted by T, to 160
ms.
– In the test scenario, additional delays (scaling-in delay, scaling-out delay, migration delay,
etc.) from virtualization or traffic control using SDN are verified.
– Virtualization effects are calculated by virt-5gc modules considering topology inputs and
VM inputs of user code.
X2 DataRate 10Gb/s
X2 LinkDelay 20ms
S1uLinkDelay 1ms
S1apLinkDelay 0s
33
Simulation time 20s
FrequencyRange 1
CenterFrequency 4700
Operating band parameters
NrOperationBand1 79
NrOperationBand2 n/a
PathlossModel Nr3gppPropagationLossModel
Scenario UMa
Channel parameters
ChannelCondition 1
OptionalNlos true
NumerologyIndex: 2
ChannelBandwidth 100
Frame structure and physical
resources parameters SymbolsPerSlot 14
SlotFormat 0
34
Mem [MB]
505/512
(Workload/Capacity
Disk [GB]
1900/3000
(Workload/Capacity)
Bandwidth (Mbps)
9500/10000
(Workload/Capacity)
Light load UPF & SMF CPU [ea]
170/210
(Workload/Capacity)
Mem [MB]
860/916
(Workload/Capacity
Disk [GB]
2400/4000
(Workload/Capacity)
Bandwidth (Mbps)
9500/10000
(Workload/Capacity)
AMF CPU [ea]
160/200
(Workload/Capacity)
Mem [MB]
460/512
(Workload/Capacity
Disk [GB]
1300/3000
(Workload/Capacity)
Bandwidth (Mbps)
9500/10000
(Workload/Capacity)
5.3 Results
Figure 6 is the collection of graphs, which show the results of our NR non-standalone scenario simulation.
The figure 6 is plotted by gnuplot tool for depicting our simulation data.
Above figure consists of 4 graphs: the variation of RLC TX queue size, TCP RTT, RLC Path throughput
and TCP Cwnd. In all graphs, x axis represents simulation time.
As you know, because LTE link capacity is only 75Mbps, we cannot achieve more than 100Mbps
throughput. However, using 5G NR (mmWave) , we get more than 100Mbps throughput. You can find RLC
path throughput graph from above figures, and total throughput (LTE + mmWave) is 300Mbps that LTE
link cannot get. Therefore, by aggregating both bandwidths (LTE + 5G NR) in the dual connectivity
structure, we are able to increase UE’s throughput a lot.
One another benefit would be short latency due to UE’s connection to 2 eNBs. When downloading big file
or web page from server, dividing traffic into 2 eNBs would reduce overall latency a lot, compared with
using single eNB.
35
Moreover, in LTE-NR dual connectivity scenario, a mobility algorithm can be evaluated. Figure 8 shows
the results of mobility performances in LTE-NR dual connectivity. The left figure describes the RLC queue
backlog of each gNB in Byte with respect to time (sec). From this figure we can observe that gNB handover
is well performed with respect to UE mobility. The right figure shows the RLC throughput at the UE with
respect to time (sec). From this figure, we can observe that our NR handover procedure and algorithm is
suitable to support 5G UE.
36
Figure 8 Mobility performance in LTE-NR dual connectivity
Using the simulator for SDN/NFV module, user can get following graphs. In the scenario, an UE
communicates with a remote host, and core networks are virtualized. There are two conditions, heavy load
and light load. In heavy load condition, core network entities such as P/S-GW or MME are suffer from high
workloads. In light load condition, core network entities have lighter workloads than those of heavy load
condition.
Figure 9 User RTT in light load condition and heavy load condition
We can find out that with heavy load condition, user experienced more out-lied RTT compared to
light load condition. The out-lied points mean that auto-scaling is performed by NFV platform, so user
experience high RTT because packets are stalled during auto-scaling of VNFs.
37
(a) Without building (b) With building
Figure 10(a) shows the SNR values of received transport blocks (TBs) when the building does not
exist. In an ideal case, where we know the channel matrix during the simulation, the highest SNR values
can be obtained by the best analog bemaforming vectors. We set the analog beamforming vector update
period T=160 ms to reflect the beam tracking procedure. We observe that the SNR values are lower than
those of the ideal case. This is because the gNB and the UE transmit and receive data using analog
beamforming vectors based on past channel information. Figure 10(b) represents the SNR values of
received TBs when the building exists. In both the ideal and non-ideal cases, we observe the SNR decreases
sharply when the channel condition is changed to the non-line-of-sight (NLOS) due to the building. When
the T is applied, if the channel condition is changed, i.e., from LOS to NLOS, vice versa, the SNR value is
zero. Therefore, we do not represent the minus infinity values on a log scale in Figure 10(b).
6. How to use
6.1 How to install 5G NR NSA simulator
Source code is in the server, 147.248.57.22:10022 at the folder, GigaKoreaProject
Configure
cd /GigaKoreaProject/mmwave_new_handover/ns3-mmwave/
$ ./build.py
$ cd ns-3.26
./waf configure
38
Build
$ ./waf –run scratch/mc-downlink
6.2 Workspace
Run program only in the scratch folder.
If you want to make your own simulation, you should move your code file to ~/ns3-mmwave/scratch/.
And when you run program, you need not write file extension (.cc)
Example
./waf --run scratch/example.cc
./waf --run example.cc
6.3 How to install 5G SDN/NFV simulator
5G SDN/NFV simulator is based on the ns-3 network simulator running on the linux operating system,
so we suppose that users who want to download source code and use it have experience on using linux.
39
$ sudo apt-get install build-essential gcc g++ python git mercurial unzip cmake
$ sudo apt-get install libpcap-dev libxerces-c-dev libpcre3-dev flex bison
$ sudo apt-get install pkg-config autoconf libtool libboost-dev
Type “Y” when the system asks you to install the packages
40
3.2.3. Link to ns-3
$ hg clone http://code.nsnam.org/ns-3.26
$ cd ns-3.26
$ hg clone https://bitbucket.org/ljerezchaves/ofswitch13-module src/ofswitch13
$ cd src/ofswitch13
$ hg update 3.1.0
$ cd ../../
$ patch -p1 < src/ofswitch13/utils/ofswitch13-src-3_26.patch
$ patch -p1 < src/ofswitch13/utils/ofswitch13-doc-3_26.patch
41
Users can modify the baseline scratch file for their customized simulation.
And if users just want to change the input parameter, “/input/virt-5gc-vmimput-
heavy.input” and “/input/virt-5gc-toposample.input” could be the keys.
42
The first one includes VM-related parameters such as cpu, memory, disk, bandwidth
capacities and workloads.
43
The second one includes topology parameters such as x, y axis of nodes, type of nodes,
and VM ID which node is placed on.
44
7. Change Log
v1.1, 2019-02-17
v1.2, 2019-04-15
45
8. References
[1] 3GPP, “Study on Architecture for Next Generation System,” 3rd Generation
Partnership Project (3GPP), TR 23.799 V14.0.0, Dec. 2016.
[2] 3GPP, "System Architecture for the 5G system", 3rd Generation Partnership Project
(3GPP), TS 23.501 V15.1.0, Mar. 2018.
[3] 3GPP, "Procedure for the 5G system", 3rd Generation Partnership Project (3GPP),
TS 23.502 V15.1.0, Mar. 2018.
[4] 3GPP, “Evolved Universal Terrestrial Radio Access (E-UTRA) and Evolved
Universal Terrestrial Radio Access Network (E-UTRAN),” 3rd Generation
Partnership Project (3GPP), TS 36.300 V15.0.0, Dec. 2017.
[5] 3GPP, “Packet Data Convergence Protocol (PDCP) specification,” 3rd Generation
Partnership Project (3GPP), TS 36.323 V14.2.0, Mar. 2017.
[6] 3GPP, “Radio Resource Control (RRC),” 3rd Generation Partnership Project (3GPP),
TS 36.331 V15.0.1, Jan. 2018.
[7] 3GPP, “Study on channel model for frequencies from 0.5 to 100 GHz,” 3rd
Generation Partnership Project (3GPP), TR 38.901 V14.3.0, Dec. 2017.
[8] 3GPP, “Multi-connectivity,” 3rd Generation Partnership Project (3GPP), TS 37.340
V15.0.0, Dec. 2017.
[9] ns-3 [online]. Available: https://www.nsnam.org
[10] Polese, Michele, et al. "Improved handover through dual connectivity in 5G
mmWave mobile networks." IEEE Journal on Selected Areas in Communications
35.9 (2017): 2069-2084.
[11] Calheiros, Rodrigo N., et al. "CloudSim: a toolkit for modeling and
simulation of cloud computing environments and evaluation of resource provisioning
algorithms." Software: Practice and experience 41.1 (2011): 23-50.
[12] Chaves, Luciano Jerez, Islene Calciolari Garcia, and Edmundo Roberto
Mauro Madeira. "Ofswitch13: Enhancing ns-3 with openflow 1.3 support."
Proceedings of the Workshop on ns-3. ACM, 2016.
46