Green Traffic Engineering Solution For

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Journal of Theoretical and Applied Information Technology

30th November 2023. Vol.101. No 22


© 2023 Little Lion Scientific

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

GREEN TRAFFIC ENGINEERING : SOLUTION FOR


REDUCING ENERGY CONSUMPTION IN SDN NETWORK
WITH SEGMENT ROUTING
FARAH LAMHARRAS1 , NAJIB ELKAMOUN2 ,OUIDAD LABOUIDYA3 ,HIND SOUNNI4
1
STIC Laboratory,Faculty of science Chouaib Doukalli University ,Eljadida ,Morocco
2
STIC Laboratory,Faculty of science Chouaib Doukalli University ,Eljadida ,Morocco
3
STIC Laboratory,Faculty of science Chouaib Doukalli University ,Eljadida ,Morocco
4
STIC Laboratory,Faculty of science Chouaib Doukalli University ,Eljadida ,Morocco
E-mail: 1farah.lamharras@gmail.com, 2elkamoun@gmail.com, 3Labouidya.o@ucd.ac.ma,
4
hind.sounni@gmail.com

ABSTRACT

Thanks to its flexibility of handling and processing due to the separation of control and data planes, The
new Software Defined Network (SDN) paradigm has attracted a lot of attention from researchers. On the
other hand, energy consumption in ICTs has become an important area of research, given the high energy
dissipation due to the manufacture of high-tech equipment and the direct use of electricity. The aim of this
study is to improve the energy efficiency of networks by switching off a subset of the links, using an SDN
(Software Defined Network) approach.We dynamically adapt the number of network links that are
switched on according to the network load.Our solution is based on the Segment Routing protocol, which
we will demonstrate that is better than MPLS.The algorithms were implemented and evaluated using the
OMNET++ simulator.Experimental results show that the number of links on can be reduced by a very
interesting percentage while maintaining a high quality of service and network stability.the proposed
solution presents a promising results to address the environmental and energy challenges faced by modern
networks.
Keywords: SDN, Green,Traffic Engineering,Energy saved,Segment Routing

1. INTRODUCTION operators a new network management paradigm. It


consists of a set of SDN switch or v-switches and
In recent years, studies [1] have shown a shift in one or more controllers. A controller provides a
the energy footprint of the ICT industry. This global view of a network. It helps an operator to
increase in energy consumption is caused mainly by optimize network performance, such as maximum
the proliferation and widespread use of wireless link utilization and energy savings, where the
broadband access, and the massive migration of bandwidth of current networks can satisfy both
services to the Cloud. Traditional networking peak and off-peak traffic demands [3].
functionality is mainly implemented in one or more [4,5] study the problem of optimizing energy in
switches/routers and application software for the SDN by rerouting traffic, using a routing protocol
end-user communication process. This represents a offering poor quality of service. Celenlioglu [6]
difficult task when it comes to programming a focuses on SDN using MPLS as a data transport
large, scalable network. With new technologies protocol based on label switching, which is
such as SDN, the control and data plane are introduced at the entrance to the MPLS network
decoupled, centralizing all network intelligence in and removed at its exit. Assefa [7] proposes to
one place, the SDN controller. replace under-utilized links with the next shortest
Energy efficiency in SDN can be simplified path, and with the next path in the direction of the
compared to traditional networks, since the control maximum link utility. All these studies address
plane is separated from the data plane, and energy- network energy saving using the idle method. [7,8]
efficient solutions can be implemented easily in an present a overview of different energy-efficient
SDN network [2]. In this paper our contribution is methods in SDN networks and carried out a
based on optimizing the energy consumed in SDN comparative study to demonstrate the usefulness of
networks. Software-defined network (SDN) offer this method over others.

7266
Journal of Theoretical and Applied Information Technology
30th November 2023. Vol.101. No 22
© 2023 Little Lion Scientific

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195


This article's main idea is to turn off under- Celenlioglu [6] focuses on SDN with
utilized link to save energy consumed by the MPLS. The work assumes that a set of MPLS paths
network, since even during peak consumption, links are pre-calculated (pre-established) and installed.
are rarely used beyond 50% of their capacity [9]. The controller performs admission control. The
The use of routing protocols can only be used to authors propose an algorithm for placing flows on
implement fairly basic rules based only on the active paths that aggregates traffic in order to put
destination address, and do not offer advanced equipment in standby mode, as well as an algorithm
traffic engineering capabilities. Unlike other for distributing the load of congested paths over
solutions which implement MPLS or dynamic several flows.The challenge is to decide which path
variation of IP metrics to reroute traffic, in this combinations can be actively used to ensure that the
paper, we use Segment Routing SR managed by an maximum number of switches are in standby mode.
SDN controller. We adopted Multi-controller To solve this equation the PLSP Selection
architecture to improve traffic rerouting [10]. To Algorithm (PSA) is used, which takes into account
calculate the new traffic routes, algorithms are set both the capacity and energy consumption of each
up to manage the entire network traffic. PLSP when activating or deactivating PLSPs (Pre-
The rest of the paper is structured as follows. established Label Switching Paths), MPLS is used
Section 2 present related works. Section 3 describes for large-scale networks.
in detail the advantages of Segment Routing over [6] chose to work with MPLS but, in our
MPLS, especially in SDN networks. Section 4 solution we use Segment Routing.We demonstrate,
introduces the proposed algorithm. Section 5 in section 3 and section 5 experimentation and
provides the simulation results and discussion, results, that it is more better than [6].
while Section 6 concludes the paper and offers Assefa [12] propose replacing
suggestions on future researches. underutilized links with the next shortest path, and
the next path in the direction of the maximum link
2. RELATED WORK
utility. They introduce an IP formulation for traffic-
The traffic approach in SDN is inspired by and energy-sensitive routing problems, based on
the fact that many network elements are under- link utility information, and evaluate the algorithms
utilized, given the redundancy of network paths. using real-time low, medium and high traffic
The idea is to set these components to standby volume traces and network topologies.but they only
when no packets need to be transmitted, in order to manage to save 30% of the energy consumed,
save energy. This technique is known as whereas our work achieves up to 50% of the energy
Linksleeping LS [11]. Energy efficient techniques saved.
utilize the approach of sleeping or turning off In [13], the proposed algorithm allows the
unused switches or links when the traffic volume is controller to configure remote routes and link rates
low and turn them on when the traffific volume by updating network device flow tables and setting
increases. port rates. This algorithm approximately solves the
Wang et al [4] study the problem of energy following SDN routing problem: given the network
optimization in SDN through traffic re-routing, in a topology, a set of data flows and their traffic
similar way to the Green-TE approach, by demands, the discrete operating rates available from
formulating a programmable integer model putting each link, searches for the integral routing path of
the under-utilized elements (idle) in sum under the the data flows and the operating rates of all links
constraints of link utilization and packet delay. Two that minimize network energy consumption.
algorithms are proposed to solve this problem: A more recent Green-TE approach using
alternative greedy algorithm and global greedy SR-based SDN is [24]. It is based on an algorithm
algorithm. Same authors in [5], define energy- that removes the least used link from the topology.
efficient routing to optimize control and data traffic In addition, it is guaranteed that a previously
in SDN, this approach adds a control traffic defined upper limit for the MLU is not exceeded.
constraint to the traditional energy-saving routing Before a link is removed, a centralized SDN
problem statement, They propose a heuristic controller calculates the respective SR
algorithm using dynamic weighting to compute the configurations to release that link from traffic. All
energy-sensitive routing path. this is done in an automated and dynamic way,
But their equations are still too complex to solve where the SDN controller constantly monitors the
using the usual optimisation techniques.[11] network and cuts the links when possible. But this
solution has not proven a scalability of the network
to demonstrate its stability.

7267
Journal of Theoretical and Applied Information Technology
30th November 2023. Vol.101. No 22
© 2023 Little Lion Scientific

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195


 By adding the header [b, L1], node h forces
Table 1: Comparative Table Of The State Of The Art And the packet to go to node a using the L1 link
Our Solution By %Energy Saved And Complexity Of The (Figure 2).
Algorithm And Protocol  In Figure 3, if h wants to send a packet to a via
Complexity/Protocol % energy f, it will add the header [f, a]. The packet will
[4] Yes Not then take the shortest path to f, then continue
defined to a.
[5] no Not Segment Routing is the ideal choice for
defined dynamic rerouting, both with SDN and
[6] No/MPLS 46% independently.
[12] NO 30%
[13] Yes/MPLS Up to 50%
[24] Yes/Segment Not
Routing defined
Our solution No/SegmentRouting Up to 50%

3. SEGMENT ROUTING AND MPLS

MPLS-TE was necessary to achieve a


higher level of traffic management, but its Figure 1:Segment identifier
complexity and lack of scalability limit its use.
Segment Routing, a standardized architecture
within the IETF SPRING group, offers an
innovative and simple approach to traffic
engineering [14,15-16], and to network
programming while resisting the scaling factor.
Organizations are increasingly adopting this
architecture. The use of MPLS as the initial data
plane has accelerated implementations and Figure 2: Forced passage through a link
deployments. The transition to Segment Routing
becomes easier as no additional protocols are
required [17][25]. Ongoing work on IPv6 (SRv6)
[18] will increase the number of uses, since it will
be possible to program any IP architecture using the
same principles.
SR use the same label switching concept
as MPLS, and can use the MPLS data plan without
modification. However, the control plane requires a Figure 3 : Forced passage through a node.
complete overhaul. For example, labels called
Segment Identifiers (SIDs) have a global scope,
4. PROPOSED ALGORITHM
unlike MPLS where their scope is local.
 A nodal identifier is unique for the entire It is clear from the previous overview that
network and identifies a node (a, b, c, ..., h). a large number of studies have been carried out on
 An adjacency identifier is local to a node and the subject of energy saving in SDN networks,
identifies an outgoing interface : in the case of particularly in terms of traffic management and the
node b, outgoing interfaces are identified by rerouting of flows between the various network
L1, L2, L3, L4 and L5). Unlike MPLS, label devices, and the use of low-energy modes. Data
distribution is done using an IGP extension rerouting techniques on alternative routes to reduce
rather than special protocols (LDP/RSVP-TE). the load on links are possible, as ISPs are over-
Once the network has been discovered, dimensioning and making networks more redundant
sending a packet to node a using the shortest in order to cope with consumption peaks and to
path boils down to adding the label a to its offer a high quality of service. However, even
header. during peak periods, links are rarely operated at
However, it is possible to manage traffic over 50% of capacity [19].
more precisely:

7268
Journal of Theoretical and Applied Information Technology
30th November 2023. Vol.101. No 22
© 2023 Little Lion Scientific

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195


In this paper, we address the energy The main objective is not to lose
consumption issue by changing the state of the connectivity. In the distributed case, nodes must
router ports at each link end. Once a link is no synchronize to avoid turning off a link that could
longer required to transfer data, the components are break the connectivity constraint. In the centralized
set to standby mode, and returned to operational version, if no network failure occurs, this risk does
mode when required: link on, link off. We assume not exist [10].
that switching off a link corresponds to turning off
a part of the integrated circuit, for example for a : 4.1 Select ON/OFF
100 Gbps = 2 (135 + 150) = 570 W (1) We implement a policy based on deleting
10 Gbps = 2 (10 + 50) = 120 W (2) less congested links, since this approach is the
Figure 4 shows variation of link energy commonly used in the literature for evaluating
consumption according to speed. results. This step consists in classifying links
according to their current transmission rate, and
then disabling the nodes that transmit the lowest
Energy consumed (W) amount of data. The challenge is not to lose
400 connectivity even when links are deactivated.
300 Regarding link reactivation, we search for
the utilization of the most-used node, and if its load
200
is greater than 75%, all links are reactivated and the
100 shortest paths are recalculated. Figure 6 below
0 shows the flowchart of the ON/OFF link selection
10 40 100 400 algorithm, which outputs a list of links that will be
switched on or off.
port card transponder

Figure 4 : Router port consumption according to port


speed (Gbps)
There is a significant increase in the
energy saved when port speeds increase.
The proposed solution can be divided into
3 steps:
 Select the links to be switched on or off.
 Re-route paths according to ON/OFF links.
 Send newly calculated routes to network
devices to switch links on or off.
The main algorithm can be described as
shown in the following flow chart (Figure 5) :

Figure 6 : Select ON/OFF flow chart

4.2 New_Route
In a traditional infrastructure, the network
may not be able to route all flows when selected
links are down. Each node doesn't always have to
recalculate the shortest paths through the network,
since congestion can still occur: congestion can
occur even when a link is active, for example if it is
one of the shortest paths for a large network.
The introduction of SDN technology into
the network is very promising, and solves the above
Figure 5 : Main algorithm flowchart
problem as long as the SDN controller has global
knowledge of network traffic, and knows the

7269
Journal of Theoretical and Applied Information Technology
30th November 2023. Vol.101. No 22
© 2023 Little Lion Scientific

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195


bandwidth used between all ends : the traffic matrix labels or segments for a particular flow along a path
is taken from [20]. or part of a path.
Flow statistics should be collected by the Figures 9 and 10 show the average
SDN switches in the routing area. It should be network throughput with mpls and with SR.
noted that in this step, the algorithm can also Network throughput is defined as the amount of
calculate alternative paths. A technique is bandwidth that is successfully delivered based on
implemented to calculate the shortest path between the total amount of requested bandwidth arriving on
each pair of network nodes and to test the ability to the network. Throughput results are very promising
route all data without congestion. The calculation using SR.
equation for the shortest path is as follows: A higher rejection rate results in less link
V·(E+V·logV) (3) utilization because most requests are not routed.It is
E : The number of links clearly demonstrated that the average throughput as
V : The number of nodes. a function of the number of responses for a network
This calculation is quite complicated, however, it using SR is higher than that using MPLS, so the use
can be optimized using optimization techniques of SR in an energy-efficient SDN network will be
[21], and using Dijkstra's algorithm at node level. beneficial.

4.3 Main Algorithm:


AVERAGE THROUGHPUT(Gb/s)
The main algorithm of our solution which 80
includes Select ON/OFF and New-Route is the Average throughput 60
following : 40
20
Inputs: list of links (output of SELECT
ON/OFF algorithm) 0
BEGIN: 300 400 500 600 700 800 900
1.New path provided by the controller number of requests
2.Request to switch off links mpls
3.Sends ACK acknowledgements to
controllers Figure 9 : Average Throughput in Gb/s
4.Controller sends broadcast request
5.Update routing tables and create new
AVERAGE LINK UTILISATION NUMBER
routes
OF REQUESTS
Average link utilization

6.The nodes in question stop sending


initialization messages. 100
7.Links enter standby mode
0
This solution is cost-effective compared to 100 200300 400500 600700 800 900
distributed solutions, because in the distributed
Number of request
version, nodes have to synchronize to avoid a link
going out and breaking the connectivity constraint.
In a centralized network using SDN, this issue does mpls segment routing
not exist.

5. EXPERIMENTATION AND RESULTS Figure 10 : Average Link Utilisation

To simulate the networks, we used the Figures 11 and 12 present network load
omnet++ 5.5.1 simulator refering to Pratap [23] and variation based on time over 24 hours and
Inet 3.6.6. Figure 7,8 present the simulated network percentage of activated links when applying the
topologies (Geant and Germany 50) are real algorithm and without it. Rerouting data over paths
topologies taken from topology zoo [22], for each considered not to be the shortest leads to additional
network we simulated one day of use. costs in terms of network load. The additional cost
SR simplifies the control plane used in of our solution is illustrated in Figures 9 and 10,
MPLS by removing RSVP. However, similar to the with a maximum value of around 17%. Despite this
MPLS data plane, the traffic source uses predefined overhead, the network converges. Even if the links

7270
Journal of Theoretical and Applied Information Technology
30th November 2023. Vol.101. No 22
© 2023 Little Lion Scientific

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195


are deactivated, there is no great influence on On the Germany 50 network (Figure 13),
network connectivity. we can see that from 3 a.m. to 9 p.m. the number of
For the energy section, we calculate the links switched off remains stable, so that for most
percentage of deactivated links, using a of the day 45% of the links are switched off, which
mathematical formula to determine the percentage also represents 45% of the energy saved.
of energy saved. The percentage corresponds to the On Figure 14, which represents the Géant network
number of deactivated links/total number of links. from 3am to 11pm, the average number of links
switched off is 50%.
Figures 13 and 14 clearly show that during
network load or number of packets exchanged peak hours the percentage of links switched off is
over 24 hours
almost constant, which means that energy
consumption is also constant and can be saved by
up to 50% for the 2 network topologies studied.
60 80%
The proposed solution guarantees
50 connectivity even when network load increases,
60%
40 which explains the stability applied to real-life core
30 40% networks such as Germany50 and Géant. The
generic nature of the approach means that it can be
20
20% progressively improved by simply modifying the
10 component algorithms.
0 0%
0 3 6 9 12 15 18 21 23 24
ENERGY SAVED
network load or number of packets exchanged 80% 80%

60% 60%
network load or number of packets exchanged
without algo 40% 40%
% links on 20% 20%

0% 0%
Figure 11 : Network load in Geant network 0 3 6 9 12 15 18 21 23 24

% links off % ennergy


network load or number of packets exchanged
over 24 hours

800 120% Figure 13: Percentage of energy saved and links


switched off as a function of time (Géant network)
100%
600
80%
ENERGY SAVED
400 60%
50% 50%
40%
200 40% 40%
20% 30% 30%
0 0% 20% 20%
0 3 6 9 12 15 18 21 23 24 10% 10%
network load or number of packets 0% 0%
exchanged without algo 0 3 6 9 12 15 18 21 23 24
network load or number of packets % links off % energy
exchanged
% links on
Figure 14: Percentage of energy saved and links
switched off as a function of time (Germany 50 network)
Figure 12 : Network load in Germany 50 network

7271
Journal of Theoretical and Applied Information Technology
30th November 2023. Vol.101. No 22
© 2023 Little Lion Scientific

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195


6. CONCLUSION AND PERSPECTIVE (WoWMoM), 2014 IEEE 15th International
Symposium on a, pages 1–6, June 2014.
In this article, we presented a new [6] Celenlioglu, M.R., S.B. Goger et H.A. Mantar.
technique for traffic engineering designed to save 2014. « An SDN-based energy-aware routing
energy consumed by SDN networks based on model for intra-domain networks ». In
Segement Routing protocol. The aim is to reduce Proceeding of 22nd International IEEE
the energy consumption of ICTs while maintaining Conference on Telecommunications and
network quality of service. We used the OMNET++ Computer Networks (SoftCOM). p. 61-66.
simulator to simulate differents approaches for [7] F Lamharras, elkamoun N, O
traffic management and flow rerouting in SDN labouidya,‘Energy Saved
networks, we compare network results using SR Approaches in Software Defined
and MPLS. Results show that SR gives better Networks: State of the Art’NISS19, March
performance than MPLS for the different scenarios. 27– 29, 2019, Rabat, Morocco © 2019
Our solution is stable respecting Association for Computing Machinery.
connectivity and applied on real networks. ACM ISBN 978-1- 4503-6645-
This study opens up a number of 8/19/03…$15.00 DOI
interesting prospects for future research into energy https://doi.org/10.1145/3320326.3320340
saving in SDN networks :1) Extension to other [8] F Lamharras, elkamoun N, O
topologies to assess its robustness and effectiveness labouidya,IJCSNS ‘Comparative
in a variety of network environments.2) Integration Analysis of Energy Saved Approaches
of other metrics such as latency, response time and in Software Defined Networks ’International
quality of service for overall network Journal of Computer Science and Network
optimisation.3) Managing the mobility of SDN Security, VOL.19 No.1, January 2019
networks and studying how to adapt our solution to [9] Hassidim (A.), Raz (D.), Segalov (M.) et
effectively manage the mobility of connected Shaqed (A.). – Network utilization :
devices. The flflow view. – In INFOCOM,
2013 Proceedings IEEE, pp. 1429–1437,
REFERENCES: April 2013.
[10] Alejandro Ruiz-Rivera, Kwan-Wu Chin,
and Sieteng Soh, GreCo: An Energy Aware
[1] The Climate Group,“Smart 2020: Enabling the
Controller Association Algorithm for
low carbon economy in the
Software Defined Networks, IEEE
information age,” 2008.
COMMUNICATIONS LETTERS,
[2] Adriana Fernandez-Fernandez, Cristina
VOL. 19, NO. 4, APRIL 2015
Cervello- Pastor, Leonardo Ochoa-Aday,
[11] Maruti, Gupta et Suresh Singh. 2003. «
Achieving Energy Efficiency: An Energy-
Greening of the Internet ». In Proc.
Aware Approach in SDN, 978-1-
ACMConference on Applications,
5090-1328- 9/16/$31.00 ©2016 IEEE.
Technologies, Architectures, and
[3] W. Fisher, M. Suchara, and J. Rexford,
Protocols for Computer
‘‘Greening backbone networks: Reducing
Communications (SIGCOMM 2003), p.
energy consumption by shutting off
19-26.
cables in bundled links,’’ in Proc.
[12] Beakal Gizachew Assefa and Oznur
ACM SIGCOMM, New Delhi, India, Aug.
Ozkasap, RESDN: A Novel Metric and Method
2010, pp. 29–34.
for Energy Efficient Routing in Software
[4] Zhanwei Wu, Ying Wang, Xingyu Chen,
DefinedNetworks,DOI10.1109/TNSM.2020.29
Xuesong, An Energy-Aware Routing for
73621, IEEE Transactions on Network and
Optimizing Control and Data Traffic in SDN
Service Management
Qiu4th Asia-Pacific Conference on
[13] Mohamad Khattar Awad, Yousef Rafique,
Computer Aided System Engineering –
Rym A. M’Hallah, Energy-aware Routing for
APCASE 2017
Software-defined Networks with Discrete Link
[5] R. Wang, Z. Jiang, S. Gao, W. Yang, Y. Xia,
Rates: A Benders Decomposition-based
and M. Zhu. Energy-aware routing
Heuristic
algorithms in software-defined
Approach,<![CDATA[Sustainable
networks. In World of Wireless, Mobile
Computing: Informatics and Systems]]>
and Multimedia Networks

7272
Journal of Theoretical and Applied Information Technology
30th November 2023. Vol.101. No 22
© 2023 Little Lion Scientific

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195


(2016),http://dx.doi.org/10.1016/j.suscom.2016.
11.003
[14] C. Filsfils, et al., Segment Routing Part I,
2017.
[15] P. L. Ventre et al., “Segment routing:
Acomprehensive survey of research rctivities,
standardization efforts, and implementation
results,” IEEE Commu. Surveys and Tutorials,
pp. 182–221, Jan. 2021.
[16] C. Filsfils, et al., Segment Routing Part II,
2019.
[17] C. -H. Lung and H. ElBakoury,
"Exploiting Segment Routing and SDN
Features for Green Traffic Engineering,"
2022 IEEE 8th International
Conference on Network Softwarization
(NetSoft), Milan, Italy, 2022, pp. 49-54,
doi: 10.1109/NetSoft54395.2022.9844091.
[18] Previdi, S., Filsfils, C., Vyncke, E., Raza,
K., Leddy, J., Field, B., . . . Raszuk, R. (2017).
draft- ietf- 6mansegment-routing-
header-07 - IPv6 Segment
Routing Header (SRH).
[19] Hassidim (A.), Raz (D.), Segalov (M.) et
Shaqed (A.). – Network utilization : The flflow
view. – In INFOCOM, 2013 Proceedings IEEE,
pp. 1429 1437, April 2013.
[20] http://sndlib.zib.de
[21] Bonetto (E.), Chiaraviglio (L.), Idzikowski
(F.) et Le Rouzic (E.). – Algorithms for the
multiperiod power-aware logical topology
design with reconfifiguration costs. Optical
Communications and Networking, IEEE/OSA
Journal of, vol. 5, n5, May 2013, pp. 394–410.
[22] http://www.topology-zoo.org/dataset.html
[23] Let, G.S., Pratap, C.B., Jagannath, D.J. et
al. Software-Defined Networking Routing
Algorithms: Issues, QoS and
Models. Wireless Pers
Commun 131, 1631–1661 (2023).
https://doi.org/10.1007/s11277-023-10516-y
[24] Otten, D., Ilsen, M., Chimani, M., &
Aschenbruck, N. (2023). Green Traffic
Engineering by Line Card
Minimization. arXiv preprint
arXiv:2306.00726.https://doi.org/10.48550/arX
2306.00726
[25] Gopal, K.P., Sambath, M., Geetha, A. et
al. Modelling of pigeon inspired optimization
based segment routing in QoS aware software
defined networks. Int. j. inf. tecnol. 15, 2499–
2508 (2023). https://doi.org/10.1007/s41870-
023-01300-z

7273
Journal of Theoretical and Applied Information Technology
30th November 2023. Vol.101. No 22
© 2023 Little Lion Scientific

ISSN: 1992-8645 www.jatit.org E-ISSN: 1817-3195

Figure 7 : Germany 50 network

Figure 8 : Geant Network

7274

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy