SCTP-MANET NEW EXTENSION OF SCTP PROTOCOL FOR THE OPTIMIZATION OF MANET PERFORMANCES

International Journal of Wireless & Mobile Networks (IJWMN) Vol. 8, No. 5, October 2016
DOI: 10.5121/ijwmn.2016.8504 53
SCTP-MANET NEW EXTENSION OF SCTP
PROTOCOL FOR THE OPTIMIZATION OF MANET
PERFORMANCES
SADOUNI Salheddine, 1
Benslama Malek 1
and Beylot André-Luc2
1
Electromagnetism and telecommunication laboratory, University of Constantine
2
IRIT, Université de Toulouse
ABSTRACT
Ad Hoc mobile networks are constituted of nodes that move freely without a centralized administration.
These nodes contribute in the routing of data packets that are sent by a source. This happens when the
latter is not capable of reaching its destination. On the other hand, their mobility causes recurrent
breakdowns of the routing paths notably with sparse MANET. In order to optimize the performance of such
networks, we suggest a new extension of protocols: Stream Control Transmission Protocols (SCTP) named
SCTP-MANET. Their main function is therefore to improve the availability of the links in sparse MANET
protocols. This could be achieved by a better integration of Multihoming. With this aim in mind, this new
extension is based on a cross-layer interface between transport and routing layers as well as the use of
specific messages.
KEYWORDS
SCTP, AODV, Cross layer interface, Multihoming, Throughput
1. INTRODUCTION
The main purpose behind establishing Ad Hoc mobile networks is to widen the notions of
mobility. Since they may contain mobile platforms – in other words nodes – which are free to
move without constraints. These networks represent an autonomous system, capable of
organizing itself without any previous infrastructure.
Radio transmissions being limited, every node is likely to be involved in the routing process in
order to forward the packets of a node that could not reach its destination. This involves the use of
three types of protocols: proactive, reactive, and hybrid. Proactive protocols try to maintain the
best paths available in order to reach every possible destination. The routes are saved even when
not used. On the contrary, Reactive protocols create and maintain the routes according to the
needs. This could be achieved by the launch of a global discovery procedure aiming at finding out
a path between the source node, and the destination one. As far as hybrid protocols are concerned,
a proactive technique is involved when recognizing the topology locally. However, the routes
leading to more remote nodes are attained through a reactive technique [1]. In this article, we will
focus on the reactive routing category, more precisely on Ad Hoc on Demand Vector Routing
protocols.

54
However, the main disadvantage of MANET is the perpetual instability of its topology as well as
the number of the nodes participating in the data routing. A sparse MANET undergoes more link
interruptions between source and destination. The intermediary nodes that do not have at disposal
alternative routes, will thus delete the data packets. This in return shows weak network
performances [2]. For such reasons, we suggest in this article a new extension of SCTP protocol.
We call it: SCTP-MANET. Its main function would be to define the best method of integrating
Multihoming to MANET. This should be done while taking into account its characteristics. The
previously mentioned concept will focus on the creation of multiple paths between source and
destination node, and ensuring the transparent passage of a broken-down node to one that works.
In order to validate the contribution of this new extension to network performances in this article,
we will be based upon a comparative study of the average throughput between SCTP-MANET
application and the standard SCTP version in different MANET topologies. This could be done
by using Riverbed Modeler software which represents a new version of Openet Modeler.
In order to do this, section two (2) will be devoted to tackle AODV routing protocol and its link
failure detection mechanism. Section three (3) will broach SCTP transport protocol and how it
handles Multihoming. The Cross-layer interface and SCTP-MANET new extension will be
presented in section four (4). As far as section five (5) is concerned, we will present the
simulation results obtained by the simulator Riverbed Modeler for which we have suggested an
SCTP setting and some of the extensions tackled in this paper. Finally, section 6 will allow us to
conclude and highlight the perspectives of this work.
2. AODV ROUTING PROTOCOLS
Ad Hoc on demand distance vector routing belongs to the category of reactive protocols. The
latter create and maintain the routes according to the needs. They use a broadcast mechanism in
order to discover the valid routes [3]. Therefore, two operations are defined.
2.1. Routes Discovery
Two types of messages are used [4]
• RREQ (Route REQuest )
• RREP (Route REPly)
A node initiates a process of route discovery when it needs a route to a new destination, or a
destination to which the route has become invalid. The sender broadcasts the RREQ message in
the network to find a path. The message is first received by the closest neighboring nodes which
will in return spread the message until reaching the destination node. Each intermediary node
stores in its routing table the address of the node that transmitted the RREQ message. Therefore,
the reverse path is established.
Once the message reaches the destination node (or a node familiar with a path leading to this
one), it sends a RREP message to the source using the reverse path. RREP travels along the path
which is followed by the route request in the opposite direction, and modifies the routing tables of
the nodes through which it passes.
2.2. Routes Maintenance
To repare a broken-down route, AODV uses the message [4]
• Route ERRor (RERP)

55
During this operation, this protocol predicts a mechanism that will repair the route locally and
save routing message transmissions. In addition, it anticipates optimizations to repair a route
locally.
2.3. Link Failure Detection
AODV protocol suggests two methods to detect links state: The transmission of the message
(Hello) at regular intervals [5], and the awareness of the information coming directly from MAC
sublayer [6].
3. SCTP TRANSPORT PROTOCOL
SCTP is a transport layer protocol which was designed by Internet Engineering Task Force
(IETF) in order to provide a reliable and secure transport of a variety of applications. One of these
most important applications is network signaling [7]. For the design of SCTP, two major
functionalities have been introduced, depending on the stream and Multihoming [8].
Stream refers to a sequence of messages within the same SCTP association (chunk). As far as
Multihoming is concerned, SCTP defines it as being the capacity of a terminal point to contain
multiple interfaces with different IP addresses. Consequently, in the case of a multi-homed
network connection, numerous paths could be established for each source interface. SCTP
chooses a primary address that will be used as a destination for every data transfer. Thus, the
other addresses are considered to be alternatives [8]. In case of primary address breakdown,
SCTP uses the alternative addresses to forward the lost data until the link is re-established with
the former one.
3.1. SCTP Association
An association is established by the exchange of four messages: INIT, INIT-ACK, COOKIE-
ECHO and COOKIE-ACK. To initiate an association, the sender emits an INIT message to the
receiver which answers it in return with an INIT-ACK message containing configuration
information to constitute a COOKIE. The receiver analyzes the information that the received
COOKIE-ECHO packet contains, in order to ensure that it is the original sender of the COOKIE
message. If the key is valid, the receiver resends a COOKIE-ACK to the sender and considers the
association to be established [9].
3.2. SCTP ADDIP Dynamic Address Addition
SCTP ADDIP is an SCTP extension defined in RFC 5061. It allows a dynamic update of the
association address list. It is known as Mobile SCTP (mSCTP) [10]. It permits to a mobile that
moves away of its home network to obtain a new address. The latter is integrated automatically to
the association, and used to pursue a transparent transfer of data.
ADDIP adds two new types of control messages: Address Configuration Change (ASCONF) and
Address Configuration Acknowledgment (ASCONF-ACK). These messages enable the addition,
the suppression or the modification of the Address on the whole of IP addresses of the two
association extremities [10].

56
3.3. SCTP Multiple-Route Method
During an SCTP association, the protocol verifies the validity of each communication route
between two nodes. The latter’s destination transport address is declared unoccupied if not used
during a pre-defined period higher than or equal to an HB interval (30 seconds by default). A
HEARTBEAT message is associated to a source transport address and to a destination one. It
answers with a HEARTBEAT-ACK message [11]. A source node increments ERROR
COUNTER parameter associated to a transport destination address when it does not receive a
retransmission timeout.
According to the value of error counter parametre, SCTP allows to detect a route’s rupture (error
counter passess the threshold Path.Max.Retrans) or an unreachable destination node (the sum of
error counter associated to the different destination node addresses is higher than the value of
Association.Max.Retrans). When the primary route is marked inactive, the sender switches
automatically its data traffic to an active alternative route without closing the association. On the
other hand, it will continue to send Heartbeat chunks regularly to the primary route in order to
verify its state. When the primary route becomes active again, SCTP protocol uses it to send the
traffic over [12].
3.4. SCTP and AODV Limits
The integration of Multihoming to sparse MANET increases their performances by switching the
stream of information from a broken-down route to one that works. That is why the use of SCTP
that takes in charge Multihoming, combined with AODV routing protocol seems to be the most
appropriate. Except that two situations arise in the present case:
1. AODV is a Unipath protocol that constitutes merely one route on demand between source
and destination [13]. Thus, when this route is broken down, the intermediary nodes suppress the
data packets, because they do not have an alternative route.
2. In addition, if on the level of transport, SCTP protocol has an alternative route, AODV
routing protocol will not be aware of the existence of a second route when the primary one breaks
down. Consequently, the source node launches a new route discovery operation [14].
For such reasons, we ought to suggest SCTP extensions that will be combined in a new protocol
SCTP-MANET. The principles of the latter will be tackled in the section to come.
4. THE DESCRIPTION OF THE NEW SCTP-MANET PROTOCOL
In order to optimize the performances of sparse MANET by using a quick recovery of the broken
down routes, it is crucial to compensate for its previously described shortcomings. Therefore,
SCTP-MANET extension would be based upon two main elements:
1. Cross layer interface.
2. The use of specific messages by adding additional parameters.
These enhancements will thus focus on not only the architecture, but also on the messages of the
SCTP protocol itself.

57
4.1. Cross Layer Interface
The current architecture of network protocols lies on a set of modules that are bound to provide a
particular service in an autonomous way. Due to their hierarchical organization in a stack, each
module finds it difficult to communicate with the other modules.
In this paper, we will use a transport and routing cross layer interface that is based on the model
of the interaction with an intermediary entity type [15]. This cross layer interface performs its
data sharing by a service of recovery, storage and data sending between the transport layer and
the routing one [16]. The data concerned with the recovery and the sharing are: RREQ, RREP,
and RRER of AODV routing protocol. As well as the messages establishing an association of
SCTP transport protocol: INIT and INIT-ACK.
In the case of primary path breakdown, the standard SCTP version launches a route discovery
procedure to look for a new route. Whereas the use of the information of transport layer 4 by the
routing layer 3, SCTP-MANET sends directly the packets through a second path. A primary
route recovery procedure will be launched thereafter.
In order to optimize network resources and avoid signaling bursts signal peaks that are relative to
the creation of multiple paths through every network node, we will implant CLI cross layers at the
level of source nodes (SCLI) (as shown in figure 1) and destination nodes (DCLI) only (as shown
in figure 2)
Figure 1. Cross Layer Interface Architecture – source node
Figure 2. Cross Layer Interface Architecture – destination node

58
4.2. The Use of Specific Messages
In addition to the use of a cross layer interface, SCTP-MANET, in order to function properly, is
based upon specific messages by adding additional parameters. In the first place, INIT and INIT-
ACK messages are necessary for the initialization of an SCTP association. In the second place,
RREQ and RREP routing messages are used by AODV protocol in route search and discovery.
SCTP-MANET starts by the creation of a primary route numbered 1. Then two other secondary
routes are numbered successively 2 and 3. These numbers will be implemented using binary code
in the flag field INIT and INIT-ACK. The eight flag bits are not used in control messages.
Consequently, INIT 1 brings to the flag field the value 00000001; INIT 2 brings 00000010, and
so on for the rest of the messages INIT3 and INIT-ACK from 1 to 3. (Figure 3)
Figure 3. The Format of a Message INIT 2
As far as RREQ and RREP routing messages are concerned, their numbering is similar to that of
INIT and INIT-ACK messages; from 1 to 3. However, they are to be implemented in the fields
RESERVED. Constituted of 9 bits to 0 for RREP and from 11 bits to 0 for RREQ successively,
these fields are ignored once received [17]. Their numbering is binary coded. RREQ 1 will have
the value 00000000001 and RREP 1 000000001 and so on for the rest of RREQ and RREP
messages.
Figure 4. The Format of a Message RREQ 3
SCTP-MANET has been designed to operate simultaneously with three active paths: a direct
route, and two rescue ones in case the first one breaks down. It is possible to schedule this

59
extension to ceaselessly look for alternative paths. However, such a move could lead to routing
signaling peaks which, in return, could result in saturating the network.
Along the same lines, the numbering of routing messages involves the sequential creation of
routes in order of importance (primary route, the first alternative route, then finally the second
alternative one). Particularly when knowing that AODV routing protocol uses a flooding-based
route discovery [18].
Finally, a node which is requested by several routes simultaneously involves their interruption in
case it moves. That is why disjointed routes should be created. In order to do this, a numbered
RREP marks its number in the routing table of every node when passing to a destination node.
Afterwards, once a new route discovery operation is launched, the numbered RREQ ignores the
marked nodes with a number different than its own.
4.3. SCTP-MANET Association Establishment Steps
The objective of an SCTP-MANET association is to continuously operate with three paths (one
primary path, and two secondary ones). In case of a primary path breakdown, SCTP-MANET
automatically switches the data sending to the first secondary path without any disconnection if it
is available. If not, it switches to a second alternative path. In order to do so, it achieves the
following operations:
1. Over a first phase, the source node attempts to initialize a transport SCTP-MANET
association. That is why it sends a message INIT1 to SCLI interface. The latter stores the
INIT1 message, and asks AODV protocol to create a path to the destination node.
2. Therefore, AODV sends a RREQ1 using a diffusion based process. Once it reaches the
destination node, DCLI stores it. Then, it answers with a RREP1.
3. On the way back, RREP1performs the same tasks as an ordinary RREP [4] (it changes the
intermediary node routing tables). Except that this RREP also marks the nodes with its
number as it passes.
4. Once RREP1 reaches the source node, SCLI sends INIT1 to the path that is previously
established by AODV. The destination node resends an INIT-ACK1
5. Then, the SCTP association establishment procedure is initialized between source and
destination node with the four messages: INIT1, INIT-ACK1, COOKIE-ECHO and
COOKIE-ACK. The last two messages are not numbered.
6. After establishing an association between source and destination node, SCTP-MANET
considers this path to be primary and relies on it to send the data packets.
7. Meanwhile, SCTP-MANET initializes the same route discovery operation with numbered
2 INIT, then with numbered 3 INIT.
8. That is why SCLI and DCLI take over and accomplish the same previously described
steps. Except that a numbered RREQ ignores the marked nodes with another number
different of its own one.
9. The return of a numbered RREP means the discovery of a secondary path to the
destination. The corresponding message INIT-ACK carries the IP addresses participating
in the association, and ADDIP takes over to include this new path.
10. The message COOKIE-ECHO will not be sent because the association already exists
[19].
11. SCTP-MANET uses an active route surveillance mechanism thereafter, by sending
HEARTBEAT chunks [11].

60
Eventually, when the primary route breaks down, SCTP-MANET seeks re-establishing it with the
sending of RREQ1, but only after ten (10) RREQ1 sending attempts fail. Therefore SCTP-
MANET eliminates this route, and realizes a gap in its active route list.
Figure 5. Source and Destination Communication with SCTP-MANET
5. SIMULATION
5.1. Simulation Framework
Riverbed Modeler Academic Edition 17.5 is a highly sophisticated simulation software package
that enables developers to model communications networks and distributed systems and allows
them to analyse the behaviour and performance of modelled systems through Discrete Event
Simulations (DES) [20] This framework is the new version of Opnet Modeler network simulator,
from which we could import network models, as well as scenarios to be executed in it. This
software is one of the most popular, accurate and applicable in the real world in the field of
network simulation and is recognized for its high reliability. It allows for the simulation of
different scenarios for a specific project and uses a project and scenario approach to modeling
networks. The project approaches a collection of related network scenarios in which each
explores a different aspect of network design. It contains at least one scenario, that is, a single
instance of a network. Simulating a scenario can overcome constraints of proprietary hardware
and software such as lack of development tools. The Riverbed Modeler offers to its user a
Graphic User Interface (GUI), standards-based LAN and WAN performance modeling and
detailed library model for most protocols and devices [21].
5.2. Performance Evaluation
The collected statistic for this comparative study is Throughput, in communication of various
types of networks, such as MANET, network throughput is the average rate of successful

61
message delivery over a communication channel. The throughput is usually measured in bits per
second (bit/s or bps) or in data packets per second or data packets per time slot. These data may
be delivered over a physical or logical link, or pass through a certain network node. The system
throughput or aggregate throughput is the sum of the data rates that are delivered to all terminals
in a network. [22] Throughput analysis is one of several important parameter of network
performance. A weak throughput implies that the transmission medium is scarcely requested by
stations, or that the medium capacity is badly used by the access technique.
5.3. Simulation Parameters
In order to validate a qualitative contribution of our SCTP-MANET extension, we considered
three different scenarios. Each one of them contains two variants:
1. The use of the standard version of SCTP protocol.
2. The use of the new extension SCTP-MANET.
Regarding the first variant, AODV protocol is activated in the third routing layer. The former
already exists in the simulation framework. This applies to all the network nodes.
On the other hand, SCTP transport layer was not implanted in the simulation tool. We have set it
up in its basic version. Same goes to the previously suggested extensions. In addition, it is
noteworthy that this important work includes modifications of the automatons generally used in
the transport layer. Particularly by adding relative states of COOKIE-ECHO and COOKIE-ACK
message sending. These modifications apply only to the source and destination nodes.
Regarding the second variant, SCTP-MANET extension is designed to operate in the fourth layer.
In order to do so, a Find-new state is added to the functioning of SCTP. The role of the latter will
be to define the second and the third paths after SCTP-MANET association has been activated.
As far as cross-layer interfaces are concerned, two new processors are created: SCLI for the
source node and DCLI for the destination node. They are situated between the layers three and
four and joint by packets stream. These modifications only apply to source and destination nodes.
Finally, regarding the routing, AODV protocol is activated in all the network nodes. Numbered
RREP and RREQ routing messages are taken into charge by SCLI and DCLI processors.
Table 1. Simulation model parameters used for SCTP-MANET
Receive buffer (bytes) 65535
Send buffer size unlimited
Maximum (INIT1, COOKIE) ACK delay (sec)
for SCTP-MANET
0.2
Maximum (INIT2, INIT 3) ACK delay (sec) for
SCTP-MANET
0.4
Maximum (INIT, COOKIE) ACK delay (sec) for
SCTP
0.2
Maximum ACK Segments 2
Duplicate ACK Threshold 3
Initial RTO (sec) 1

62
Minimum RTO (sec) 0.5
Maximum RTO (sec) 64
5.4. Simulation Scenarios:
In a MANET, every node is likely to contribute in the data routing. Thus, the number of the nodes
that constitute a MANET influences directly the network performances. In order to observe this,
we realize three scenarios:
1. The first scenario: a dense network of thirty (30) nodes.
2. The second scenario: a medium-density network of twenty (20) nodes.
3. The third scenario: a sparse network of ten (10) nodes.
Table 2. Simulation model parameters used for the three scenarios
Data rate Mbps 1
number of wireless communication
interfaces
4
mobility model MANET_Up_right
MANET_down_left
Route active time (sec) 3
Hello Interval (sec) uniform(1. 1.1)
Allowed hello loss 2
Net surface (km²) 20
Node traversal time (sec) 0.04
Route error rate limit (pkts/sec) 10
Timeout buffer (sec) 2
AODV Local repair Enable
Time of simulation (hour) 1
Figure 6. The Topology of a Thirty-Node MANET

63
5.5. The Results
Figure 7. First scenarios thirty nodes MANET average throughput
Figure 7 demonstrates the average throughput of the first scenarios with (30 nodes). For these two
(2) variants, we notice an important gap between the two (2) curves; because when SCTP-
MANET reaches a maximum value of 30 000 bits/sec, SCTP has a maximum value of 22 000
bits/sec.
Figure 8. Second scenario twenty nodes MANET average throughput

64
Figure 8 shows the average throughput of the second scenario (20 nodes). Compared to the first
scenario, we notice a decrease of the throughput of the two variants. On the other hand, the gap
between the average throughput of the two variants SCTP and SCTP-MANET rises.
Figure 9. Third scenario ten nodes MANET average throughput
Finally, figure 9 illustrates the third scenario of a sparse network of only ten (10 nodes). The same
could be said here: a decrease of the average throughput for the two variants of this scenario.
Only an important gap grows between SCTP average throughput which reaches 3 900 bits/sec
and SCTP-MANET which has a maximum value of 9200 bits/sec that is almost three times
higher.
5.6. Results Discussion
A comparative analysis of these three scenarios reveals that the decrease of the number of
MANET nodes causes a fall in the average throughput. However, the use of a new extension
SCTP-MANET shows even better average throughput compared to SCTP, with a gap that rapidly
develops between these two variants (SCTP-MANET and SCTP) in sparse networks.
This could be attributed to the decrease of the number of the nodes participating in the routing,
which involves a weak availability of the links. This directly results in recurrent transport path
breakdowns. Compared to SCTP basic protocol, the extension SCTP-MANET manages better the
recovery of the broken-down paths in sparse MANET; thanks to its anticipating method of
detecting two alternative paths, as well as to its transparent switching from a broken down link to
a functional one.
6. CONCLUSION
Compared to the standard SCTP version, the use of SCTP-MANET optimizes significantly the
performances of sparse MANET which suggest weak link availability. The enhancement of such
network performances is due to the fact that SCTP-MANET prevents transport path breakdowns.
This extension works with a minimum of three active paths. In addition, the use of a direct

65
transport and routing cross layer interface and specific SCTP-MANET messages suggests a better
integration of Multihoming. The latter takes in charge realizing a transparent passage between the
paths without data loss or network disconnection.
Finally, in order to enhance the contribution of this new extension to network performances, we
suggest that future work in this area study the use of a similarity counter. Incremented when paths
pass by the same nodes, this counter decides upon the worth of a secondary path.
REFERENCES
[1] C.S.R. Murthy, B. Manoj, Ad-hoc Wireless Networks: Architectures and Protocols, Prentice Hall,
New Jersey, USA, 2004
[2] G. Angione, P. Bellavista, A. Corradi, E. Magistretti, A k-hop clustering protocol for dense mobile
ad-hoc networks, in: International Conference on Distributed Computing Systems Workshops, vol.
01, 2006. doi:http://dx.doi.org/10.1109/ICDCSW.2006
[3] Marina M.K., Das S.R. “Ad hoc on-demand multipath distance vector routing”. WC& MC, Vol. 6,
Wireless Ad hoc Networks: Technology and Challenges,2006
[4] Thomas Kunz, "Implementation vs. simulation: Evaluating a MANET multicast protocol", in
Proceedings of the Global Mobile Congress 2004, Shanghai, China, 129-134, 2004, Delson Group
2004
[5] A. Nilsson, C.E. Perkins, A.J. Touminen, R. Wakikawa, J.T. Malinen, AODV and IPv6 Internet
access for ad hoc networks, Mobile Computer and Communications Review, 102–103, 2013.
[6] J. Yu, P. Chong, A survey of clustering schemes for mobile ad hoc networks, IEEE Commun. Surv.
Tutor. 32–48, 2005
[7] Stewart R “Stream Control transmission Protocol” RFC 4960, IETF, Septembre 2007.
[8] R. Stewart, Q. Xie, M. Tuexen, S. Maruyama, M. Kozuka, Stream Control Transmission Protocol
(SCTP) Dynamic Address Reconfiguration, RFC 5061, 2007.
[9] R. Stewart, Q. Xie, K. Morneault, C. Sharp, H. Schwarzbauer, T. Taylor, I. Rytina, M. Kalla, L.
Zhang, V. Paxson, Stream Control Transmission Protocol, RFC 4960, 2007.
[10] Kelly , G.m. Muntean , P. Perry , J. Murphy , Delay-centric handover in SCTP over WLAN, Trans.
Autom. Control Comput. Sci. 1–6, 2004.
[11] J.R. Iyengar, P.D. Amer, R. Stewart, Concurrent multipath transfer using SCTP multihoming over
independent end-to-end paths, IEEE ACM Trans. Netw. 951–964, 2006.
[12] T. Stegel, J. Sterle, J. Bešter, A. Kos, SCTP association between multi-homedendpoints over NAT
using NSLP, Electrotech. Rev, 277–284. 2008
[13] A. Nasipuri, R. Castaneda, S.R. Das, Performance of multipath routing for on-demand protocols in
mobile ad hoc networks, Mobile Networks and Applications,339–349. 2001
[14] P. Goudarzi , Scalable video transmission over multi-hop wireless networks with enhanced quality of
experience using Swarm in- telligence, Signal Process. Image Commun.
[15] Charoenpanyasak S.,Paillassa B.,“SCTP multihoming with Cross Layer Interface in Ad Hoc
Multihomed Networks” IEEE Wi-Mob, New York USA, 2007.
[16] Kim K.,“A Cross Layered Approach for Multihoming on SCTP in mobile Ad hoc Networks”,
TENCON (Taipei International Convention Center), IEEE Region 10 Conference,2007.
[17] G. RAJKUMAR, DR. K. DURAISAMY, “A Review of Ad Hoc On-Demand Distance Vector
Routing Protocol for Mobile Ad HocNetworks”, in Journal of Theoretical and Applied Information
Technology, Vol. 36 No.1, 2012
[18] R. Oliveira, M. Luis, L. Bernardo, R. Dinis, Towards reliable broadcast in ad hoc networks, IEEE
Communications Letters 314– 317. 2012
[19] Md. Nurul Islam, A. Kara, Throughput analysis of SCTP over a multi-homed association, Computer
and Information Technology, CIT ’06, in: The Sixth IEEE International Conference on Computer and
Information Technology, Seoul, Korea, 110–110. 2006
[20] Introduction to OPNET Modeler. Networks Competence Area, COM. DTU, Retrieved from:
[Online].Availableat:http://oldwww.com.dtu.dk/staff/vbi/EuroFGI/doc/EuroFGI_Intro2 Modeler.pdf.

66
[21] Riverbed Technology, The Riverbed Splash Community, OPNET Modeler Space, the RSVP Refresh
Reduction Extensions, 2013. <https://splash.riverbed.com/docs/DOC-3202>.
[22] Koltsidas G., Palvidou F., Kuladinithi K., Timm-Giel A., Gorg C., “Investigating the performance of
multipath protocol for ad-hoc networks” The 1th Annual IEEE PIMRC, Septembre 2007.

SCTP-MANET NEW EXTENSION OF SCTP PROTOCOL FOR THE OPTIMIZATION OF MANET PERFORMANCES

More Related Content

What's hot (19)

Similar to SCTP-MANET NEW EXTENSION OF SCTP PROTOCOL FOR THE OPTIMIZATION OF MANET PERFORMANCES (20)

Recently uploaded (20)

SCTP-MANET NEW EXTENSION OF SCTP PROTOCOL FOR THE OPTIMIZATION OF MANET PERFORMANCES