Improvement of Data Transmission Speed and Fault Tolerance Over

Improvement of Data Transmission Speed and Fault Tolerance Over
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 8, No. 3, 2017
Improvement of Data Transmission Speed and Fault
Tolerance over Software Defined Networking
SM Shamim
Mohammad Badrul Alam Miah
Nazrul Islam
Dept. of Information and
Communication and Technology
Mawlana Bhashani Science and
Technology
University
Dhaka, Bangladesh
Dept. of Information and
Communication and Technology
Mawlana Bhashani Science and
Technology University
Dhaka, Bangladesh
Dept. of Information and
Communication and Technology
Mawlana Bhashani Science and
Technology
University
Dhaka, Bangladesh
Abstract—Software Defined Networking (SDN) is a new
networking paradigm where control plane is decoupled from the
forwarding plane. Nowadays, for the development of information
technology large number of data traffic has been added in the
global network each day. Due to proliferation of the Internet, ecommerce, video content and personalized cloud-based services
higher channel bandwidth required to deliver larger data from
one center to others. Lower data communication speed and fault
tolerance are major factors for SDN which degrades network
performance. This paper presents enhancement of data
communication speed and fault tolerance over SDN using Link
Aggregation Control Protocol (LACP. The result of this paper
shows network performance has been improved by increasing
approximately 31% data transmission speed over SDN using
LACP. Moreover, this paper shows fault tolerance have been
improved by LACP which prevents failure of any single
component link from leading to breakdown of the entire
communications.
Keywords—Fault Tolerance; Link Aggregation Control
Protocol (LACP); OpenFlow; Mininet Emulator; Software Defined
Networking (SDN)
I.
INTRODUCTION
Software Defined Networking (SDN) [1, 2, 3] is a new
approach for managing, building and designing computer
networks which decouple the network’s control plane from the
forwarding planes. It has emerged as new paradigm in
networking which has the possibility to enable ongoing
network innovation and enable the network as a programmable,
pluggable component of the larger cloud architecture [4]. SDN
is being strongly considered as the next promising networking
platform. In recent years, SDN has been developing
tremendously in different organizations [5]. In order to reduce
operational costs and strengthen network architecture different
companies are planning or deploying SDN in their network [6].
In the next five years, SDN will be considered one of the most
advanced information technologies over the world [7, 8].
About US $2 billion has been estimated to invest in SDN for
knowledge discovery [9].
In order to handle the larger data high configuration router
and switch are needed. Server and storage resources are
interconnected via switches and routers [10]. Adding more
switch and router will increase operational cost which reduces
the network performance. In addition, network path failure one
of the major problem which reduce network efficiency. An
efficient routing, sever load balancing, access control and
traffic monitoring system has to be designed to overcome these
limitation. One of the possible solutions of these problems is
Link Aggregation control protocol where two or more ports in
an Ethernet switch are combined together to operate as a single
virtual port. It increases available bandwidth by aggregating
two or more links between network devices.
Due to programmability of Software Defined Networking,
standard mechanisms needed for achieving higher data
transmission speed and fault tolerance. Different researchers
propose few techniques [11, 12, 13, 14 and 15] to improve data
transmission speed and fault tolerance over Software Defined
Networking. In this consequence, the paper [11] analyzed
Bandwidth and latency aware routing using OpenFlow over
SDN which improve network performance. Paper [12]
proposed a novel architecture BRAS (Broadband Remote
Access Server) which could enhance data transmission speed
according to users’ preference in specific applications.
In [13] authors presents Software Defined Networking
based on OpenFlow can be used to build efficient solutions in
order to handle fault-tolerant multicast in substation
environments. Their implementation handles single link failure
and also indicates how their approach can be expanded to
handle multiple link or node faults. Fault tolerance issue with
systematically review the existing methods has been proposed
in [14] which are useful in failure recovery. In [15] proposes
new architecture to strengthen the reliability and fault-tolerance
over SDN in terms of network operations and management.
After studying related works realized that several researchers
improve data transmission speed and fault tolerance over SDN
separately with different technique.
However, there needed further studies to improve data
transmission speed and fault tolerance over SDN in order to
increase network performance. Yet, there has been lack of
studies, which can enhance both of these two major facts at
same time. Though Link Aggregation control protocol is
known for traditional network architecture, no implementation
has been done over Software Defined Networking. Link
Aggregation [16] is a technology defined in IEEE802.1AX2008, which is a method of combining multiple physical lines
to be used as a logical link. It increases capacity and
352 | P a g e
www.ijacsa.thesai.org
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 8, No. 3, 2017
availability of the network between specific devices (both
switches and end stations) with the help existing Fast Ethernet
and Gigabit Ethernet technology. It has higher potential
transmission speed and higher accessibility in contrast to
conventional connections using an individual cable. The
purpose of this paper is to enhance data communication speed
and fault tolerance over SDN at same time using Link
Aggregation control protocol (LACP).
The rest of the paper is organized as follows. Section II
detail describes Link Aggregation with the types in
background. Section III describes the research methodology.
Section IV presents the experiment setup in details.
Experimental results and discussion are evaluated in Section V.
Section VI concludes the paper and deliberates the future
perspectives.
II.
BACKGROUND STUDY
In Link Aggregation, multiple parallel physical links are
combined together between two devices in order to form single
logical link. This function also provides load balancing where
the processing and communications activity distributed over
multiple links to avoid single link overwhelmed. The
architecture diagram of Link Aggregation function is shown on
Fig-1.
appear that open-source network simulator Mininet has good
potential for simulating Software Defined Networking. In order
to evaluate designed network topology with OpenFlow virtual
switch Mininet has been installed over Ubuntu 14.04. Its
installation and configuration is easy and straightforward than
other simulators. Virtual Software Defined Networking can be
designed using Mininet which consists of OpenFlow [25]
controller, OpenFlow-enabled Ethernet switches and multiple
hosts connected to those switches. OpenFlow is one of the
most widely deployed SDN communications standards
protocols. This protocol used in order to communicate between
controller and other networking devices i.e. switch, router etc.
A component based Software Defined Networking framework
Ryu has been used as OpenFlow controller. Ryu Controller
managed and maintained by open Ryu which is written in
Python [26]. After careful study of the experiment different
network analysis graph has been plotted which shows the
expected results.
IV.
EXPERIMENT SETUP
Custom network topology has been designed using Mininet
API which is shown on Fig-2. Designed network consist of one
OpenFlow switches, an OpenFlow Ryu controller and three
hosts. All the host h1, h2 and h3 are connected with the
OpenFlow switch s1. Link Aggregation function has been
implemented between OpenFlow switches s1 and host h1.
All of the hosts have assigned unique IP address and MAC
address. The IP address and MAC address for host h1 are
’10.0.0.1/24’ and ’00:00:00:00:00:01’. For all the other host
corresponding IP and MAC address is also assigned i.e. host h2
(’IP=10.0.0.2/24’ and MAC=’00:00:00:00:00:02’) , host h3
(IP=’10.0.0.3/24’ and MAC=’00:00:00:00:00:03’) , and host
h4 (IP=’10.0.0.4/24’ and MAC=’00:00:00:00:00:04’).
Fig. 1. Link Aggregation Architecture Diagram [17]
Two Edges switch are connected to distribution switch and
distribution switch connected to server using link aggregation.
In order to form a higher data transmission capacity several
link from server may connect with different switch ports. Link
Aggregation provides higher link availability, increase link
capacity and arrogates replace upgrading over conventional
network [17].
III.
RESEARCH METHODOLOGY
A literature review has been performed in Software
Defined Networking research scope and challenges. After the
review, fault tolerance and data communication speed option
arrived over SDN. A handy simulation tool was needed to
analyze the SDN. Different experimental studies have been
performed among OMNET++ [18], EstiNet [19], OFNet [20],
Maxinet [21], NS-3 [22] and Mininet [23, 24]. The studies
Fig. 2. Designed Network Topology
The Linux bonding driver [27] provides a method for
combining more than one network interface controllers (NICs)
into single logical bonded interface. Initially, bonding driver
module has been loaded in host h1 to perform link aggregation.
There are two interfaces in host h1 which are h1-eth0 and h1eth1. These two interfaces are bond together in order to form
one logical interface, i.e. bond0One of the commonly used
network analysis tool iperf has been used to measure the
performance. Network analysis tool Wireshark formerly known
353 | P a g e
www.ijacsa.thesai.org
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 8, No. 3, 2017
as Ethereal has been used which captures packets in real time
and display in human-readable format.
sender and receiver. The size of average packet and average
Bytes in per second are approximately 82.699 and 82701.
Two scenarios have been executed to evaluate the
performance where deigned network topology has been
executed without LACP implementation and secondly
topology executed with LACP implementation. The
corresponding result of each execution has been captured by
Wireshark. For each corresponding result three performance
analysis graphs throughput graph, time sequence graph and
round trip time graph has been drawn. Details comparison
among these graphs has been shown in result section.
Without LACP SDN throughput begins from the lower
value approximately 10,100 after that it keeps on increasing
and decreasing within some specific range. At the same time,
the LACP SDN throughput graph begins with large value
approximately 16,000. After some time it decreases sharply
and again increase sharply within some specific range. A
details comparison between two throughput graphs has been
shown on table-1.
TABLE I.
V.
EXPERIMENT RESULT
A. Throughput Graph
In data communication network throughput refers to
average rate of successful message delivery over a
transmission channel which measured in bits per second or in
data packets per second or data packets per time slot. Data may
be delivered over a physical or logical link, or pass through a
certain network node. Figure-3 shows throughput graph with
implementation of LACP over SDN and Figure-4 shows
another throughput graph for without implementation of LACP
over SDN. Throughput graph is valuable in understanding endto-end performance.
COMPARISON OF THROUGHPUT RESULT BETWEEN LACP SDN
AND WITHOUT LACP SDN
Features
LACP SDN
Total Packets
Time Duration Between First
And Last(sec)
Avg. Packets/Sec (bytes)
4202
Without LACP
SDN
2896
54.864
35.019
108.765
82.699
Avg Packet Size
1506.591
1511.572
Bytes
6330694
4377512
Avg Bytes/Sec
163864.50
125004.87
Avg Mbit/Sec
1.311
1.00
From the Table-1, average megabyte per second for the
without LACP SDN is 1.00MB and for the LACP SDN is
1.311 MB which is approximately 31% higher. The simulation
graph and data table-1 are shown that LACP SDN has higher
rates of throughputs than without LACP SDN. Data
communication speed has been improved approximately 31%
for LACP SDN compare to without LACP SDN.
B. TCP Time Sequence Graph
Time-Sequence graphs visualize TCP-based traffic. In an
ideal situation, the graph plots from the lower left corner to the
upper right corner in a smooth diagonal line.
Fig. 3. With LACP SDN Throughput Graph
From the With LACP SDN throughput graph Fig-3, highest
throughput in bytes approximately 20250 bytes and lowest is
16000 bytes. There are about 4202 packets has been
transmitted between sender and receiver. The size of average
packet and average Bytes per second are 108.765 and
125004.869.
Fig. 5. TCP Time Sequence Graph
Fig. 4. Without LACP SDN Throughput Graph
From the Without LACP SDN throughput graph Fig-4,
there are about 2896 packets has been exchanged between
Fig-5 shows comparison of two time sequence graph where
LACP SDN time sequence graph identified by blue lines and
without LACP SDN time sequence graph indicated by red
lines. The Y axis defines the TCP sequence numbers and X
axis defines the simulation time. The slope of the line would be
the theoretical bandwidth of the pipe. The steeper the line, the
354 | P a g e
www.ijacsa.thesai.org
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 8, No. 3, 2017
higher the throughput. From the Fig-5 more packets has been
transmitted for LACP SDN compare to without LACP SDN.
C. Round Trip Time Graph
Round-trip time (RTT) is the length of time takes a data
packet to be sent plus length of time takes for acknowledgment
to be received of that packet to be received. RTT Graph depicts
round trip time from a data packet to corresponding ACK
packet. The Y axis is created based on the highest round trip
latency time. Latency times are calculated as the time between
a TCP data packet and the related acknowledgment.
nodes. Now check connectivity between host h2 and host h1
and captured the corresponding result using Wireshark. The
result shows in Fig-7, host h2 still communicate with host h1
using port s1-eth2 and port h1-eth1.
Fig. 7. Ping Result from host h2 to host h1
If we remove port h1-eth1 from the aggregation, h2 will
still communicate to the host h1 using port s1-eth1 in the
switch and port h1-eth0 in the host h1. Instead of failure occurs
one links, Link Aggregation function able to check and
automatically recover the communication using other links.
VI.
Fig. 6. TCP Round Trip Time Graph
The Y axis defines the Round Trip Time (RTT) in seconds
and X axis defines TCP Sequence Numbers. Fig-6 shows
latency times are very high at few points in the trace file and
there are specific moments when the traffic is bursty in nature.
For without LACP SDN RTT graph, some packet has higher
latency time compare to average latency time. There are also
some packets for LACP SDN RTT graph has higher latency
time. Lower round trip time for the corresponding sequence
number always expected. After comparing two graphs, LACP
SDN RTT graph shows better output which has less round trip
time than without LACP SDN RTT graph.
D. Improvement of Fault Tolerance
Fault tolerance is the ability of the system to perform its
function even in the presence of one of multiple link failures.
Fault tolerance setup or configuration prevent computer
network device from failing in the event of an unexpected
problem or error among connected devices. It is one of the
major problems for networking application which can be
improved by using Link Aggregation function. Designed
network topology in Fig-2, two aggregated link are available
between host h1 and switch s1 where each side link has
separated port numbers (h1-eth0, h1-eth1, s1-eth0, and s1eth1). All of the host (h2, h3 ad h4) can communicate with the
host h1 using either port s1-eth1 or port h1-eth0. Each of the
host can also communicate by using port s1-eth2 in switch and
port h1-eth1 in host h1. Now disable one communication
channel from aggregated group where disabled channel port
number h1-eth0 which is opposite interface of s1-eth1.
After separating one channel, test the network connectivity
by using ping command which sends ICMP echo request
message and wait for corresponding reply between defined
CONCLUSION
This paper presents improvement of data transmission
speed and fault tolerance over Software Defined Networking.
The result obtained after extensive simulation study which has
evaluated by TCP time sequence graph, throughput graph and
round trip time graph. Throughput graph shows data
communication speed has improved approximately 31% over
SDN by using Link Aggregation control protocol. Simulation
result in TCP time sequence graph and RTT graph shows
network performance also improved for LACP SDN. LACP
ensure failure safety systems which are crucial for every
network administrator. The automatic configuration protocol
LACP provides redundancy with dynamic switching to the
standby link in case the active link fails. Moreover, it can be
implemented in SDN using existing hardware which decreases
the operational cost for upgrading the performance and
resiliency of a system. Our future works involves improvement
of fault tolerance and data transmission speed over Software
Defined Wireless Networking (SDWN).
[1]
[2]
[3]
[4]
[5]
REFERENCES
B. Nunes, M. Mendonca, X.-N. Nguyen, K. Obraczka, T. Turletti, et al.,
“A Survey of Software-Defined Networking: Past, Present, and Future
of Programmable Networks,” Communications Surveys & Tutorials,
IEEE, vol. 16, no. 3, pp. 1617–1634, 2014.
Y. Jarraya, T. Madi, and M. Debbabi, “A Survey and A Layered
Taxonomy of Software Defined Networking,” Communications Surveys
& Tutorials, IEEE, vol. 16, no. 4, pp. 1955–1980, 2014.
D. Kreutz, F. M. Ramos, P. Esteves Verissimo, C. Esteve Rothenberg, S.
Azodolmolky, and S. Uhlig, “Software-Defined Networking: A
Comprehensive Survey,” Proceedings of the IEEE, vol. 103, no. 1, pp.
14–76, 2015.
M. Kobayashi, S. Seetharaman, G. Parulkar, G. Appenzeller, J. Little, J.
Van Reijendam, P. Weissmann, and N. McKeown, “Maturing Of
OpenFlow and Software-Defined Networking through Deployments,”
Computer Networks, vol. 61, pp. 151–175, 2014.
B. Leng, L. Huang, X. Wang, H. Xu, and Y. Zhang, “A Mechanism For
Reducing Flow Tables in Software Defined Network,” IEEE
International Conference on Communications (ICC), pp. 5302–5307,
IEEE, 2015.
355 | P a g e
www.ijacsa.thesai.org
(IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. 8, No. 3, 2017
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
C. Yoon, T. Park, S. Lee, H. Kang, S. Shin, and Z. Zhang, “Enabling
Security Functions With SDN: A Feasibility Study,” Computer
Networks, 2015.
“Network world, gartner: 10 critical it trends for the next five years.”
Online
Available:
http://www.networkworld.com/news/2012/102212gartner-trends-263594.html. Accessed: 201511-20.
M.t. review, 10 emerging technologies: Tr10: software-defined
networking.”
Online
Available:
http://www2.technologyreview.com/article/412194/tr10software-defined
-networking. Accessed: 2015-11-20.
“Enterprise networking, IDC: SDN a 2 billion market by 2016,” Online
Available:
http://www.enterprisen-etworkingplanet.com/datacenter/idcsdn-a-2-billion-market-by-2016.html. Accessed: 2016-10-25.
TBitar, N., Gringeri, S., & Xia, T. J. “Technologies and protocols for
data center and cloud networking”. IEEE Communications Magazine,
51(9), 24-31. (2013).
Pongsakorn, U., Kohei, I., Putchong, U., Susumu, D., & Hirotake, A.
(2014). “Designing of SDN-Assisted Bandwidth and Latency Aware
Route Allocation”. Journal of Information Processing Society of Japan,
2014(2), 1-7.
Chicago
Li, K., Guo, W., Zhang, W., Wen, Y., Li, C., & Hu, W.
(2014, May). “QoE-based bandwidth allocation with SDN in FTTH
networks”. In 2014 IEEE Network Operations and Management
Symposium (NOMS) (pp. 1-8). IEEE.
Pfeiffenberger, T., Du, J. L., Arruda, P. B., & Anzaloni, A. (2015, July).
“Reliable and flexible communications for power systems: fault-tolerant
multicast with SDN/OpenFlow”. 7th International Conference on New
Technologies, Mobility and Security (NTMS) in 2015 (pp. 1-6). IEEE.
Chen, J., Chen, J., Xu, F., Yin, M., & Zhang, W. (2015, November).
“When Software Defined Networks Meet Fault Tolerance: A Survey”.
In International Conference on Algorithms and Architectures for
Parallel Processing (pp. 351-368). Springer International Publishing.
Kim, D., & Gil, J. M. (2015). “Reliable and Fault-Tolerant SoftwareDefined Network Operations Scheme for Remote 3D Printing”. Journal
of Electronic Materials, 44(3), 804-814.
[16] “Link Aggregation Control Protocol (LACP) (802.3ad) for Gigabit
Interfaces.”
Online
Available:
http://www.cisco.com/c/en/us/td/docs/ios/12_2sb/feature/guide/gigeth.ht
ml. Accessed: 2016 − 08 − 27.
[17] “link aggregation according to IEEE standard 802.3ad online.” Online
Available:https://www.google.com.bd/url?sa=trct=jq=esrc=ssource=webcd=3cad=rjauact=8ved=0ahUKEwj
Ek5GQ7aJ AhV XCI4KHR7pAjUQF gg3MAIurl = http, Accessed:
2015 − 11 − 20.
[18] “OMNeT++ Discrete Event Simulator”
Online Available:
https://omnetpp.org/, Accessed: 2016-09-25
[19] “EstiNet
9.0
|
Simulator”,
Online
Available:
http://www.estinet.com/ns/?page_id=21140, Accessed: 2016-09-30
[20] “OFNet- Quick User Guide ”Online Available: http://sdninsights.org/,
Accessed: 2016-10-19
[21] “MaxiNet: Distributed Network Emulation”, Online Available:
https://maxinet.github.io/, Accessed: 2016-11-08
[22] “ NS-3”, Online Available: https://www.nsnam.org/, Accessed: 201611-08
[23] “Mininet: An Instant Virtual Network on your Laptop (or other PC) Mininet” Online Available: http://mininet.org/. Accessed: 2016-06-15.
[24] F. Keti and S. Askar, “Emulation of Software Defined Networks Using
Mininet in Different Simulation Environments,” 6th International
Conference on in Intelligent Systems, Modeling and Simulation (ISMS),
2015, pp. 205–210, IEEE, 2015.
[25] N. McKeown, T. Anderson, H. Balakrishnan, G. Parulkar, L. Peterson, J.
Rexford, S. Shenker, and J. Turner, “OpenFlow: Enabling Innovation in
Campus Networks,” ACM SIGCOMM Computer Communication
Review, vol. 38, no. 2, pp. 69–74, 2008.
[26] “osrg/ryu . github.” Online Available: https://github-.com/osrg/ryu.
Accessed: 2015-11-01
[27] Bonding,
Online
Available:
http://www.linuxfoundation.org/collaborate/workgroups/-networking/bonding, Accessed: 2016-0615.
356 | P a g e
www.ijacsa.thesai.org
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement