This document describes setting up network slicing on a Mininet testbed using FlowVisor. It outlines creating an upper and lower network slice managed by separate SDN controllers. Flow spaces are configured on switches to delegate traffic on certain ports to each slice. Connectivity tests show the slices isolate hosts from each other while intra-slice communication is maintained.

1. 5G Network Slicing Using Mininet
Mohammed Abuibaid
Fall 2019
Email:
m.a.abuibaid@gmail.com

2. Outline
Background
Generic 5G Network Slicing Framework
FlowVisor
Demo – Network Slicing

3. Background - One Size Does Not Fit All
Multiple Applications 
Different QOS requirements 
Same authentication, mobility, reliability, delay, and QOS
 COMPROMISES 
4G Radio Acess Network 4G Evolved Packet Core Network

4. Generic Mobile Network Slicing Framework
 Heterogeneous set of infrastructure components
like data centers, routers and base stations.
E.g. Rogers Communications
 Creation of each network slice according to service
instance requests coming from the upper layer.
 Enabling Technologies: SDN and NFV
 NFs: Firewall, MMEs, S-GWs, Load balancer
Foukas, X.; Patounas, G.; Elmokashfi, A.; Marina, M. K. (2017). "Network Slicing in 5G: Survey and
Challenges" (PDF). IEEE Communications Magazine. 55 (5): 94–100.
 MVNOs: e.g. Fido, Chatr Mobile, and SimplyConnect.
 All shares the underlying physical network and it
provides a unified vision of the service requirements.

5. FlowVisor
 Special Purpose OpenFlow Controller
 Enables Network Virtualization
 Transparent proxy between OpenFlow switches
and multiple controllers
 Creates rich slice of network resources
 Delegates control of each slice to a different
controller
 It intercepts the messages
FlowVisor Implemented on OpenFlow
What is FlowVisor?

6. Demo – Network Slicing
Isolated Network
Slices
Physical Network Topology

7. Network Slicing: Step by Step
 In the main VM terminal, running the custom topology except operating the POX controller which will be
started after slice configurations.
sudo python Network_Topology.py
 In a new VM terminal (via Putty), making sure that the flowvisor is stopped.
sudo /etc/init.d/flowvisor stop
 Generating the flowvisor Config.json file.
fvconfig generate /etc/flowvisor/config.json

8. Network Slicing: Step by Step
 Starting flowvisor
sudo /etc/init.d/flowvisor start
 By using fvctl command, enabling the Flowvisor topology controller
fvctl -f /dev/null set-config --enable-topo-ctrl

9. Network Slicing: Step by Step
 Displaying the content of flowvisor Config file to make sure all switches dpid’s are listed in the fvadmin
field
fvctl get-config

10. Network Slicing: Step by Step
 In case the switches are not linked to the Flow visor, restarting the flowvisor to ensure all topology
switches get connected to it
sudo /etc/init.d/flowvisor restart
 Listing the existing slices to make sure nothing has been created previously.
fvctl list-slices

11. Network Slicing: Step by Step
 Listing the existing flow spaces to make sure nothing has been created previously
fvctl -f /dev/null list-flowspace
 Listing the existing data paths to the connected switches
fvctl -f /dev/null list-datapaths

12. Network Slicing: Step by Step
 Listing the existing links between all switches
fvctl -f /dev/null list-links

13. Network Slicing: Step by Step
 Upper and Lower Slices creation:
 Using the fvctl command, creating a slice called upper which will be managed by a separate controller
that control all the traffic in this slice. The “controller-url” is set “tcp:localhost:7777” and the admin email is
“adam@upperslice”
fvctl -f /dev/null add-slice upper tcp:localhost:7777 adam@upperslice
 Similarly, creating the lower slice:
fvctl -f /dev/null add-slice lower tcp:localhost:3333 aleen@lowerslice

14. Network Slicing: Step by Step
 Listing the existing slices to make sure that the upper and lower slices are correctly created.
fvctl list-slices
 Upper Slice Configurations:
 On switch s1:
Creating a flow space named dpid1-port1 (with priority value 1) that maps all the traffic on port 1 of switch
s1 to the upper slice, giving it all permissions (upper=7): DELEGATE, READ, and WRITE.
fvctl -f /dev/null add-flowspace dpid1-port1 1 1 in_port=1 upper=7
 Creating a flow space named dpid1-port3 (with priority value 1) that maps all the traffic on port 3 of switch
s1 to the upper slice, giving it all permissions (upper=7): DELEGATE, READ, and WRITE
fvctl -f /dev/null add-flowspace dpid1-port3 1 1 in_port=3 upper=7

15. Network Slicing: Step by Step
 On switch s2:
creating a flow space named dpid2 (with priority value 1) that maps all the traffic at switch s2 (match
value of any) to the upper slice, giving it all permissions (upper=7): DELEGATE, READ, and WRITE.
fvctl -f /dev/null add-flowspace dpid2 2 1 any upper=7
 On switch s4:
Similar to S1 configurations, creating a flow space named dpid4-port1 (with priority value 1) that maps all
the traffic on port 1 of switch s4 to the upper slice, giving it all permissions (upper=7): DELEGATE,
READ, and WRITE.
fvctl -f /dev/null add-flowspace dpid4-port1 4 1 in_port=1 upper=7
 Creating a flow space named dpid4-port3 (with priority value 1) that maps all the traffic on port 3 of switch
s4 to the upper slice, giving it all permissions (upper=7): DELEGATE, READ, and WRITE
fvctl -f /dev/null add-flowspace dpid4-port3 4 1 in_port=3 upper=7

16. Network Slicing: Step by Step
 Lower Slice Configurations:
 Similar to the upper flow spaces configurations, creating the following flow spaces on switches s1, s3,
and s4.
fvctl -f /dev/null add-flowspace dpid1-port2 1 1 in_port=2 lower=7
fvctl -f /dev/null add-flowspace dpid1-port4 1 1 in_port=4 lower=7
fvctl -f /dev/null add-flowspace dpid3 3 1 any lower=7
fvctl -f /dev/null add-flowspace dpid4-port2 4 1 in_port=2 lower=7
fvctl -f /dev/null add-flowspace dpid4-port4 4 1 in_port=4 lower=7

17. Network Slicing: Step by Step
 Ensuring that all upper and lower flow spaces are correctly configured.
fvctl -f /dev/null list-flowspace

18. Network Slicing: Step by Step
 In a new terminal (via Putty), running a POX controller for the upper slice listing on port 7777.
cd pox
./pox.py openflow.of_01 --port=7777 forwarding.l2_pairs
 Performing connectivity test (pingall) in the main VM terminal. As shown below, only hosts h1 and h3
reaches each other, isolating them from the other hosts h2 and h4.

19. Network Slicing: Step by Step
 In a new terminal (via Putty), running POX controller for the lower slice (listing to port 3333).
cd pox
./pox.py openflow.of_01 --port=3333 forwarding.l2_pairs
 Performing connectivity test (pingall) in the main VM terminal. As shown below, hosts h2 and h4
reaches each other, isolating them from the other hosts h2 and h4.

20. Network Slicing: Step by Step
 One can notice that the hosts can reaches only the other hosts in their slice and there is no connections
to the hosts in other slice.
 Testing the maximum achievable bandwidth between hosts h1 and h3 for the upper slice and between h2
and h4 for the lower slice using the command iperf.
iperf h1 h3
iperf h2 h4