Relevance of Cross Layer design on TCP/IP network model
Arpana Mysore Ananthan, University of Ottawa (amyso043@uottawa.ca)
Abstract—A network can be wired or wireless and each of these has its own set of advantages and disadvantages when it comes to the way it functions. A few common areas which are important for any network are the way routing is done, the method of resource allocation, TCP performance so on and so forth. Issues in the functionality of these areas results in a lesser efficient network. Thus, a solution to overcome these discrepencies gives rise to the analysis of what cross layering is and how this concept can be applied in mitigating issues in a few common wireless networks, mostly adhoc networks. The concept of cross layering aids in a better communication between the network layers and thus brings about a more wholesome interaction between them to better cater to the network requirements. In this article, a generalized idea of what cross layer networks are is discussed and further, networks where this concept can be applied to solve related issues has been touched upon.
Index Terms—Adhoc network, Cross layer design, QoS
I. INTRODUCTION
T
HE present architecture of internet is based on the five layered TCP/IP model[42]. It has the property of abstraction and thus each layer is unaware of the information contained in the other layers. Communication can occur only between two adjacent layers in the present TCP/IP model. However, with the introduction of cross layering, two non-adjacent layers can communicate with each other[1]. The information of one layer can be shared amongst all the five layers which converts the information hiding property into information sharing. This is essential for a layer to understand how it needs to perform based on the inputs it receives from the other layers. This makes the layers more dynamic and gets rid of the abstraction property to some extent while retaining the modularity for the most parts. Cross layer structure can be designed for any of the five layers, mostly for the network, data link and physical layers. The three main goals of cross layer design are security, Quality of service and mobility. The way these three goals are achieved are detailed in further sections. It is however interesting to note that at any given point it is easier to obtain a tradeoff between two of these goals and highly ideal to achieve all three goals. The most important areas where cross layer design can be implemented to yield better results are in wireless networks in terms of the way a network topology is and real time video streaming[45] which has a high bandwidth requirement and cannot tolerate even the slightest amount of delay[4], [5]. The information sharing as mentioned earlier can be done between the nodes and information is passed around based on sharing schemes which can be categorized based on how the information needs to be shared between the layers[6], [7], [8], [9] and by the organization of the network[6], [7], [8], [9] respectively as a)Manager method where the layers can directly communicate with each other b)Non-manager method where a concept of vertical plane is introduced and it acts as a central database through which the layers can communicate with each other ; c)Centralized method where a central controlling entity is hosted on a node which takes care of the information sharing. A tier structure is implemented in this case[1] d)Distributed method where the information is handled without the central entity.
The organization of this article is as follows. In Section II the details of the above classification of cross layer design have been detailed followed by the discussion of the goals in section III. Section IV details a few network scenarios where cross layer design have been implemented. Section V provides the conclusion.
II. CLASSIFICATION OF CROSS LAYER DESIGN
It is essential to understand how the concept of cross layering can be implemented to successfully obtain results that are expected. For example, it is good practice to introduce a QoS co-ordination plane to improve the capabilities in wireless networks[1]. This shall further be discussed in the next section while discussing the goals of cross layer design. Before implementing these features, certain questions need to be answered as to how the design needs to be deployed? If all the nodes in the wireless network need to have the cross-layering capability or just a selected few? Is an entirely new architecture needed or would the existing protocol architecture would suffice? In order to come up with the most suitable design for specific network needs, the cross-layer classification needs to be considered.
A. Non-Manager method
Fig.1 Non-manager method[1]
This method maintains the five-layer structure of the TCP/IP model while changing the functions of protocols in the necessary layers thus permitting the direct communication between the layers[1]. The introduction of cognitive networks (CN)[34],[39] can be considered as the development of a new type of data network. A cognitive network is a specific kind of network which has a unique way of identifying the current network conditions and based on these, makes decisions and acts accordingly[1],[36],[41] How this can be used in co-ordination with cross layer networks shall be seen further in the later sections. The representation of the non-manager method is given in fig.1. Consider the layered structure in fig.2 which is an example for the non-manager cognitive type of cross layer design. The four layers TCP, Data link, MC and Physical, communicate with each other; not necessarily all of them with each. The TCP layer communicates with the Data link layer and thus determining the frame size sent to it; with the MAC layer regarding the access decisions[1] and with the physical layer about the modulation and coding schemes. These communications are possible only when the cross-layer design is implemented.
Fig.2 Cognition based cross layer design[1]
B. Manager method
Fig.3 Manager method [1]
This differs from the previous method since this introduces the concept of a vertical plane which acts as a sharing entity from where the layers can share the data[1]. If any two layers need to communicate with each other, it has to be done via the vertical plane fig.3 unlike the non-manger method in which the layers can directly communicate with each other. There are various issues with the strict layered protocol architecture such as high error rate in wireless networks, increased QoS demands of a dynamic network and the unpredictability associated with it[1] which can be successfully eliminated by means of usage of a vertical plane. In [8] the concept of a cross layer manger has been introduced fig.4. This can be considered as a vertical plane as the layers share information with the manager and this cross-layer manager acts as a database of the state variables and shares it with the other layers when necessary.
Fig.4 Functionality of cross layer manager[1]
C. Centralized method
Fig.5 Centralized method [1]
This type of implementation of cross layer design involves one central node that makes the information sharing decisions. In a cellular network this can be a base station which in a hierarchical flow, acts as a centralized entity fig.5 [1]. Here the concept of a low priority function is introduced, which takes into consideration the state of the channel and the multi user variation. In[10] this concept has been used in a UMTS network to obtain good performance improvement. The general idea here would be the usage of the priority function to assess the channel information and state by the radio scheduler that decides the way transmissions must go about i.e. it is on the basis of this priority function that the transmissions are scheduled in a descending order for each user[10]. The priority function is dependent on the MAC layer and this affects the parameters of the transport and application layer such as delay and capacity. A user with a better channel becomes a high priority user and these users receive good channel conditions which results in a reduced power of transmission. The priority function in this example is determined by the base station and thus acting as a centralized entity.
D. Distributed method
Fig.6 Distributed method[1]
There is no usage of a centralized node in this method. Thus, information sharing is dependent mostly on the mesh connectivity fig.6 of one node to another. In other words, a node can be reached via multi-hop paths. Distributed cross layer design is less complex when compared to the complexity involved with centralized cross layer design[11]. The problem of overhead can be noticed here since the sharing of information in a multihop network proves to be a cause for this. Cross layer congestion control scheduling is a solution to this issue[11]. This compares the node and link centric information to come up with a potential solution to solve the congestion[11]. When nodes request for information to be sent across the layers, it is shared with them by means of a distributed scheduling algorithm[11]. This algorithms aids in cross layer information sharing between the layers. It allocates the required requested network resources to the nodes upon a request for cross layer information sharing. An example for a distributed cross layer design is shown in the fig.7. The goal here is to arrive at an achievable capacity of the network[1]; for which Adaptive modulation coding is used at the data link layer which is used to have varying modulation schemes so as to sync with the channel quality thus at all times maintaining a reasonably decent link rate[1]. The CQI (Channel Quality Indicator) provided by Adaptive modulation is utilized by the MAC layer in deducing frequencies that need to be allocated to various links.
Fig.7 Example of distributed cross layer design [1]
The MAC and network layer operate coordinating with each other by selecting traffic flows that might offer lesser probability of congestion such that it is most optimal. Further the Transport layer uses congestion distortion optimized scheduling to have control over the transmission and retransmission of packets[1][12]. At the topmost level, the application layer decides a suitable encoding rate to go through in yielding the most efficient streaming[1]. It is important here to observe that there is no centralized entity here which hosts the information sharing, but each node is responsible for communicating with the other node it deems necessary for a superior outcome.
III. STEPS AND GOALS OF CROSS-LAYER DESIGN
A. Steps involved in cross-layer design process[1]
1. In order to minimize the overhead while sharing information between layers, the layer abstraction computes the abstraction of parameters for each layer. The reason this is essential is that this helps in avoiding the transmission of excess parameters to the cross-layer optimizer[1]. Example of certain parameters are FEC, bit error rate, frame rate, modulation scheme, channel coding etc.[13].
2. Reconfiguration of the parameters to optimize special objective functions can be performed by optimization techniques.
3. The reconfigured parameters need to be distributed to the concerned layers such that the layers can perform their corresponding operation. This is done by layer reconfiguration[13].
The process of layer abstraction occurs at first followed by which the parameters obtain a revised value from the optimizer. These values are distributed then to all the five layers. This is particularly benefitable since the cross-layer optimizer reduces by a vast majority the strain on the vertical plane that is present in the manager method architecture of cross-layer design classification.
B. Goals of cross-layer design
There are three goals of cross-layer design which are Security, QoS and Mobility fig.8. These goals can be modelled as a co-ordination model[1] which conveys the functionality that can be supported by cross-layer designs[7] [8].
Fig.8 Cross layer co-ordination plane [1]
The co-ordination model comprises of three co-ordination plane viz. security plane, QoS plane and mobility plane. Each of these co-ordination planes have the information regarding the protocols and the modifications made to them as per the cross-layer design.
1. Security
Encryption standards such as SSH, SSL, Wi-Fi protected access can be deployed in a cross-layer design with the end goal of a secured way of communication. Examples for security in cross layer design can be seen in [14] which talks about the using the cross-layer design in wireless MAN by making use of a sub linear rekeying algorithm which provides utmost security; [15] the concept of network security management gathers the information from all layers which is then used to come up with and optimal security solution.
2. QoS
The main idea to achieve a QoS in a cross layer designed network is to have the upper layers such as the application layers communicate with the lower layers such a physical and data link layers[7]. This cross-layer communication helps in better servicing of the QoS requirements. For example, TCP performance can be increased by making use of a cross layer design based on explicit loss notification which can eliminate the unnecessary intermediate dropping of the packets.
3. Mobility
Wireless network comes with the problem of excessive packet drop and rapidly changing node topologies, it is the idea of certain cross-layer designs to provide uninterrupted and seamless communication to avoid the aforementioned issues[1]. An example for trying to increase the number of users who are serviced is to use TDMA and FDMA techniques. One of the methods to resolve the time slot waste issue is by using a cross layer design-based CDMA[1].
IV. CROSS LAYER DESIGNS FOR EFFICIENT ROUTING AND VIDEO TRANSMISSION
A. Cross layer design for Adhoc networks
Adhoc networks have the innate capability of a quick set-up of the network without actually the need for a fixed topology[2]. They can self-organize and are dynamic in nature which is an added advantage. The challenges faced in an adhoc network are mostly to meet energy requirements, QoS, security and scalability. Incidentally these are the very goals that the cross-layer design guarantees. Thus, it is of utmost importance to tweak the routing protocols by the use of cross layer design such that a better system performance ( normalized energy consumption) can be obtained[2]. In the previous sections, a more general view of the cross-layer design methodology has been detailed. Based on this, in this section, a detailed example of a cross layer design to cater to energy efficient routing is detailed for adhoc networks[35].
Implementing cross layer design for better performance of adhoc networks may sound ideal, however there are certain challenges associated with this. This is due to the time variant property of the wireless channels[2]. The signal in a wireless channel is susceptible to fading effects such as path loss, shadowing and multipath which is not the case in wired networks. This leads to performance degradation in higher layers. To avoid this, either the packet has to be retransmitted on the data link layer or the transmission power has to be modified accordingly to maintain the transmission without any hinderances[2]. This in turn may affect the functionality of other nodes by causing interference and also excess contention for channel accessing[2].
In the network layer, the existing route may fail or become invalid; a request for a new route might arrive; either way energy needs to be spent to perform route maintenance or for a route discovery procedure[2]. Thus, proper modifications need to be made for the protocol stack to be used at each layer to make up for the rapid variations[16].
Although the cross-layer design concept seems perfect, there are some disadvantages that come with it as well. These are the following:
a. The interactive nature of the layers in cross layer design results in creation of dependencies i.e. if any one of the layers is affected, not only that layer, but also the other layers associated with this layer get affected. This results in a propagation of error across all layers which is not favorable.
b. The complete changeover of the layered architecture to help cross layer design can result in a high cost of implementation[17].
Thus, the designing of cross layer networks must be done with utmost care so as to not lose the modularity completely and by ensuring the OSI layer rules be adhered to thus ensuring the advantages of layered structure[18], [19].
B. Cross layer design for energy efficient routing
Wireless Adhoc networks have no fixed topology and the nodes function based on the energy that they have. This energy is supplied to them via a battery which is limited, and thus efficient usage of energy is very crucial[38]. For better energy management, not all nodes transmit information. Only the nodes identified as relay, transmit the information. This results in a dynamic construction and tear down of the network as per requirement. This mobility causes to have various changing CQIs hence a considerable number of control messages needs to be exchanged between nodes that are involved in this operation[2]. This ultimately leads to a high energy consumption. To reduce the energy consumption at tall points thus keeping the nodes alive for the maximum possible duration the problem statement. Power aware routing protocols seem like a solution for a multi user network[20].
C. Cross layer routing design
1. The first consideration of cross layer design can be seen to be between the physical and network layer in the further discussion. In order to reduce the Bit error rate (BER) a new scheme which involves diverse error resilient forward error correction (FEC)[2] was introduce in [21]. With a low SNR, there can be an improvement in the transmission range. The fig.9 depicts an example for cross layer design involving the physical and network layer with the assistance of Multiple-Antenna Relay nodes {MA-RN) [2].
Fig. 9 Cross layer design involving physical and network layer[2]
The routing metric here is the number of hops however, the use of the MA-RN concept brings about a decreased number of hops from the source to the destination. The effect of MA-RNs can be seen on the bit transmission range and the performance of the FER by keeping the other parameters such as node density, mobile speed and the transmission range constant[2]. node transmits power[2]. It is the property of MA-RN to reduce the overall energy consumption Er. This parameter is number of MA-RNs(nMA) used, mobile speed, the number These parameters although are dynamically varying, it only makes sense to keep them constant such that they don’t have rapidly changing effects on the normalized energy consumption parameter is shown in the fig.10.
Fig.10 Normalized energy consumption equation[2]
where N is the number of nodes in the network and H is the number of hops in an established route and Lapp is the transmission range which all combined together yield a normalized energy consumption parameter[2]. Along with that, a near to perfection capacity achieving coding abstraction[2] and a coding scheme namely a cascaded IrCC-URC-STTC is deployed in the physical layer[22]. In the data link layer IEEE802.11b is used. In the network layer, there has been use of DYMO protocol for routing[23]. This protocol is highly flexible in high mobility environment which is very good for wireless networks and also since it exerts a low control load on the network[23]. In the transport layer, UDP is used[24]; a streaming technique called as CBR data streaming is used in the application layer[24]. The wireless channel is modelled by use of the Additive white gaussian noise (AWGN) channel which is subject to free space path loss and Rayleigh fading like most channels[2].
2. The second network scenario where cross layer design is implemented is for video transmissions which are high bandwidth requiring networks and always need to be serviced with utmost priority. A device in a running metro train, while streaming video, undergoes excessive handoffs due to the rapidly changing base stations while moving. This yields a heavily distorted video which is unfavorable.[3]. Thus, there is a need to have a cross layer design to optimize the handoff decisions[33] and the parameters of the application layer[3]. As considered in [3], for the purpose of understanding this, a metro passenger information system(PIS) is modeled by a finite state markov chain based on which the state information is obtained and a channel state transition matrix[3] of probability is deduced based on real time testing conducted in sample scenarios. A PIS fig.11 is nothing but a multimedia information system[43] which provides real time information to the passengers based on their interests
Fig.11 PIS[3]
such as bus schedules if commuting is what they are looking for, live news updates, natural disasters so on and so forth. All of these are delay sensitive applications and cannot afford to have distortion or delay in them. The PIS considered for analysis in [3] is a metro PIS which is a train used by passengers for commuting purposes. This system deploys a train ground video communication network[3] which ensures the delivery of fine quality video. Although WLAN is the most used technology for this system[32],[37] due to its readiness and availability, it is not ideally suited to be used under a high-volume traffic since it is not built to handle such high data constrained traffic. This in turn affects the QoS and QoE since the communication quality is poor in highly mobile environments as these[3]. A handoff procedure needs to be performed at a very rapid rate each time the device moves from the access point of a WLAN to the access point of another[3]. This handoff can incur considerable latency which can cause multiple interruptions to the video transmission. In addition to that, traffic management is also another issue that needs to be considered since such video transmissions often require high bandwidth, however it may not always be available which makes it necessary to compress the large video by video compression standards viz. H.263 and MPEG4[3].
To improve video transmission quality over such lossy conditions, there has been a considerable amount of work done in the area. Forward error correction (FEC) and Automatic repeat request (ARQ) have been used to correct bit errors in the compressed videos[25], [26]. [27] introduces the concept of adaptive sub-band coding which helps solve the lossy channel problem. [3-8] discusses the scheduling and adaptive framework for mobile worldwide interoperability for WiMAX systems. A combined source channel rate-distortion model is proposed in [28], [3] which can, with a minimum delay determine the channel error distortion that was induced. A concept called fountain codes[29] has been used for
protection from error caused due to lossy packet in the network. The solutions to optimize the video transmission are all mostly to cater to the lossy channel due to fading and other effects. The low quality of video occurring due to handoffs in rapidly mobile networks is equally important. This has been addressed in [3] and detailed. The way a metro PIS is designed is crucial. What is the right instant for a handoff to occur and what changes need to be made for the application layer policies are some key aspects that needs to be considered which in a way affect the performance of the network. On the basis of the channel conditions, the wireless controller deployed on the train makes handoff decisions and combines the parameters of the application layer as well to better serve the QoS[40] that is expected.
A. Metro PIS- A simplified view
Consider the metro PIS detailed in fig.11. The system comprises of the following subsystems:
a) A control center
b) Station
c) Train
d) Network
The control center is responsible for transmission of the real time data to the station subsystem which consists of data servers. This information is sent to the train sub system by means of a network subsystem and the information is displayed on the onboarded (on train) equipment. The real time feedback from the train subsystem is provided to the station subsystem by the means of the network subsystem. This includes the backbone network and other onboard networks[3]. The highly mobile environment makes the data more susceptible to errors since the video is already sent in a compressed manner in addition to it being rapidly handed off from each changing WLANs at stages wherever necessary[3].
B. Deployed standard and fountain codes
A modified standard called IEEE802.11p is made use of in the metro PIS systems. This is a tweaked version of the original IEEE 802.11 but meant specifically for a highly mobile vehicular environment[30]. It is called as Wireless access in vehicular environment[3]. It works with the data link and physical layers of the OSI networking model3]. The MAC protocol used in this standard is called Enhanced channel distributed access (ECDA). ECDA has a slightly varying OFDM standard than the IEEE802.11a. The channel is 10-MHz instead of the conventional 20-MHz which was used in the IEEE802.11a[3].The IEEE802.11p allows a better and more enhanced capability of exchange of data in vehicular networks which need to communicate often with the road side networks[3]. This standard imposes minimal required parameters that the system needs to execute the handoff procedure.
Fountain codes are the highlight of the metro PIS system under consideration in [3]. The essence of these codes are to generate the required number of parity check packets dynamically. The original chunk of data is fragmented into k number of packets and these are termed as input symbols[3]. The packets transmitted are called as output symbols. The ideal nature of fountain codes is that any given fountain encoder can generate n number of output symbols and similarly the decoder will have the capability to reconstruct these without any trouble. This is different from the way FEC codes work, that they generate all encoded symbols prior to transmission for a channel rate that is fixed[3]. Unlike FEC, the fountain codes can recover the packets that are lost due to the handoff procedures, without the need to retransmit.
Fig.12 Real time SNR for two access points[3]
C. Train-ground video communication network[3].
The internal network structure of all the layers involved in the metro PIS system can be seen in fig.13. The earlier mentioned fountain codes are dealt with in the application layer.
Fig.13 Internal network of metro PIS[3]
The FEC codes the data packets which are encapsulated with Real time transport protocol (RTP)[3] are further sent along the intended network. The network layer has the IP packets which are encapsulated by the IEEE802.11p standard of the MAC layer[44]. This network only proves to be efficient if the system knows under what conditions handoff must take place and the corresponding policies in the application layers that need to be modified accordingly[3]. The performance of the metro PIS takes a hit under situations where the above two conditions are not properly considered to design the system.
D. Channel modelling and optimization of video transmission in metro PIS
The metro PIS has a train that traverses along a railway line which results in the multiple handoffs that are required to maintain the video transmission to the train subsystem. Fig.3 depicts the SNR from two consecutive Access points that were measured in [3] to understand the behavior of the channel before modelling it. There is a rapid change in the SNR which indicates the effects of path loss, fading and other distortion causing phenomena such as shadowing, doppler spreads and the radiation pattern of the antenna[3]. The concept called Finite state Markov channel (FSMC) is used to characterize the fading that occurs in the wireless channel[3]. In this technique, the SNR received can be fragmented into multiple discrete levels. Each of these levels represent one state in the Markov chain[3].There are L states; i represents the channel state and represents the SNR corresponding to the channel state[3]. Th transition from one state to another can be expressed as transition probability which can be deduced from the SNR measurements made real time.
The video transmission can be optimized by using a technique called as Markov decision process (MDP)[3][31]. A less generalized SMDP is mentioned to be used in the literature of [3]. The underlying concept here is that the decision maker can choose the relevant action when there is a change in the system state. The time spent in any state tends to follow a random probability distribution[3]. The previously mentioned conditions, with respect to the knowledge of when to perform handoff and what higher layer policies need to be considered, are calculated based on the value iteration algorithm[3]. The five key elements of the SMDP model used are[3]:
a) Decision epochs- These are the instants at which the controller on the train has to make decisions of handoff after a particular amount of time has lapsed.[3]
b) Action and state- At every decision epochs the controller on the train needs to determine if it needs to connect to the current chosen access point or proceed to the next one[3]. While doing so it is important to consider the application level parameters as well[3].
c) Rewards
d) Transition probabilities- These are the probabilistic rates at which the states could change and can be determined based on the real time SNRs obtained by physical measurements.
Thus, based on effective cross layer techniques such as use of fountain codes, IEEE802.11p and SMDP, the video transmission capabilities of a metro PIS could be increased based on the analysis detailed in [3]. A brief outline of the same has been provided in section IV-C-2 of this document.
V. CONCLUSION
The existing internet model based on TCP/IP model has five layers which have complete modularity and abstraction which may hinder the performance of the network on various levels, affecting the QoS mostly. A cross layer designed network brings about means for each of these layers to communicate with each other such that each of the layers is well aware of the information contained in the other which might facilitate better network operation. The benefits of cross layer design under two separate circumstances were outlined in section IV-C 1 and 2. These examples are just a speck in this vast area capable of mitigating various other problems. The field of cross layer designed networks opens up a plethora of opportunities which can indeed prove to be beneficial considering the network in its entirety.
References
[1] B. Fu, Y. Xiao, H. J. Deng and H. Zeng, "A Survey of Cross-Layer Designs in Wireless Networks," in IEEE Communications Surveys & Tutorials, vol. 16, no. 1, pp. 110-126, First Quarter 2014.
[2] J. Zuo, C. Dong, S. X. Ng, L. Yang and L. Hanzo, "Cross-Layer Aided Energy-Efficient Routing Design for Ad Hoc Networks," in IEEE Communications Surveys & Tutorials, vol. 17, no. 3, pp. 1214-1238, thirdquarter 2015.
[3] L. Zhu, F. R. Yu, B. Ning and T. Tang, "Cross-Layer Design for Video Transmissions in Metro Passenger Information Systems," in IEEE Transactions on Vehicular Technology, vol. 60, no. 3, pp. 1171-1181, March 2011.
[4] A. Goldsmith, and B. Girod, “Cross-Layer Design of Adhoc
Wireless Networks for Real-Time Media.” Available:
http://www.stanford.edu/ zhuxq/adhoc project/
[5] Y. Xiao, Y. Zhang, M Nolen, H Deng, and J Zhang, “A Cross-layer
Approach for Prioritized Frame Transmissions of MPEG-4 over the
IEEE 802.11 and IEEE 802.11e Wireless Local Area Networks,” IEEE
Systems J., Vol.5, No.4, pp.474-485, Dec. 2011.
[6] V. Srivastava, “Cross-Layer Design: A Survey and the Road Ahead,”
IEEE Communications, December 2005.
[7] F. Foukalas, V. Gazis, and N. Alonistioti, “Cross-Layer Design Proposals for Wireless Mobile Networks: A Survey and Taxonomy,” IEEE Commun. Surveys & Tutorials, pp. 70-85, 2008.
[8] G. Carneiro, J. Ruela, and M. Ricardo, “Cross-Layer Design in 4G
Wireless Terminals,” IEEE Wireless Commun., Vol. 11, No. 2, pp. 7-
13, April 2004.
[9] H. Zeng, K. Kwak, J. Deng, B. Fu, Y. Xiao, J. Jeski, “Proactive
and Adaptive Reconfiguration for Reliable Communication in Tactical
networks,” Proc. SPIE 8405, Defense Transformation and Net-Centric
Systems 2012, May 2012.
[10] R. Ferrs et al., “Cross-Layer Scheduling Strategy for UMTS Downlink
Enhancement,” IEEE Commun. Mag., Vol. 43, No. 6, pp. 24-28, June
2005.
[11] X. Lin, N. B. Shroff, and R. Srikant, “A Tutorial on Cross-Layer
Optimization in Wireless Networks,” IEEE J. Sel. Areas Commun.,
Special Issue on Non-Linear Optimization of Communication Systems,
2006.
[12] E. Setton, T. Yoo, X. Zhu, A. Goldsmith, and B. Girod, “Cross-Layer
Design of Ad Hoc Networks for Real-Time Video Streaming,” IEEE
Wireless Communications, August 2005.
[13] S. Khan et al., “Application-Driven Cross-Layer Optimization for
Video Streaming Over Wireless Networks,” IEEE Commun. Mag., Vol.
44, No. 1, pp. 122-130, 2006.
[14] C. Huang, M. Matthews, M. Ginley, X. Zheng, C. Chen, M. Chang,
“Efficient and Secure Multicast in WirelessMAN: A Cross-Layer
Design,” Journal of Communications Software and Systems 3(3), pp.
199-206, 2007.
[15] I. Chuang, C. Hsieh, and Y. Kuo, “An adaptive cross-layer design
approach for network security management,” 13th International Conference on Advanced Communication Technology (ICACT), pp. 1085-1089, 2011.
[16] A. Goldsmith, Wireless Communications. New York, NY, USA:
Cambridge Univ. Press, 2005.
[17] F.Fu and M. V. D. Schaar, “A new systematic framework for autonomous cross-layer optimization,” IEEE Trans. Veh. Technol., vol. 58, no. 4, pp. 1887–1903, May 2009.
[18] V. Srivastava and M. Motani, “Cross-layer design: A survey and the road ahead,” IEEE Commun. Mag., vol. 43, no. 12, pp. 112–119, Dec. 2005.
[19] V. Kawadia and P. R. Kumar, “A cautionary perspective on cross-layer
design,” IEEE Wireless Commun., vol. 12, no. 1, pp. 3–11, Feb. 2005.
[20] J. G. Li, D. Cordes, and J. Y. Zhang, “Power-aware routing protocols
in Ad Hoc wireless networks,” IEEE Wireless Commun., vol. 12, no. 6,
pp. 69–81, Dec. 2005.
[21] M. C. Vuran and I. F. Akyildiz, “Error control in wireless sensor networks: A cross layer analysis,” IEEE/ACM Trans. Netw., vol. 17, no. 4, pp. 1186–1199, Aug. 2009.
[22]Information Technology-Telecommunications and Information Exchange Between Systems-Local and Metropolitan Area Networks-Specific Requirements, IEEE Std. 802.11, 2007.
[23] Dynamic MANET On-Demand (DYMO) Routing Routing. [Online].
Available: http://tools.ietf.org/html/draft-ietf-manet-dymo-19.
[24] User Datagram Protocol. [Online]. Available: http://tools:ietf:org/html/
rfc768
[25] H. Cai, B. Zeng, G. Shen, Z. Xiong, and S. Li, “Error-resilient unequal
error protection of fine granularity scalable video bitstreams,” EURASIP J. Appl. Signal, vol. 2006, pp. 1–11, Apr. 2006.
[26] G. Conklin, G. Greenbaum, K. Lillevold, A. Lippman, and Y. Reznik,
“Video coding for streaming media delivery on the Internet,” IEEE Trans. Circuits Syst. Video Technol., vol. 11, no. 3, pp. 269–281, Mar. 2001.
[27] M. Srinivasan and R. Chellappa, “Adaptive source–channel subband videocoding for wireless channels,” IEEE J. Sel. Areas Commun., vol. 16, no. 9,pp. 1830–1839, Dec. 1998.
[28] Z. He, J. Cai, and C. W. Chen, “Joint source channel rate–distortion
analysis for adaptive mode selection and rate control in wireless video
coding,” IEEE Trans. Circuits Syst. Video Technol., vol. 12, no. 6, pp. 511–523, Jun. 2002.
[29] A. S. Tan, A. Aksay, G. B. Akar, and E. Arikan, “Rate–Distortion
optimization for stereoscopic video streaming with unequal error
protection,” EURASIP J. Appl. Signal Process., vol. 2009, pp. 1–14,
Jan. 2009.
[30] Draft Amendment for Wireless Access in Vehicular Environments (WAVE), IEEE Std. 802.11p/D3.0, Jul. 2007.
[31] C. Luo, F. R. Yu, H. Ji and V. C. M. Leung, "Cross-Layer Design for TCP Performance Improvement in Cognitive Radio Networks," in IEEE Transactions on Vehicular Technology, vol. 59, no. 5, pp. 2485-2495, Jun 2010.
[32] Dhivya Gomathy G., P. P. Edwin Winston and S. Kuzhaloli, "A new adaptive method to optimize SCTP during vertical handover in UMTS/WLAN integrated networks," 2014 International Conference on Communication and Signal Processing, Melmaruvathur, 2014, pp. 1890-1894.
[33] Fei Yu and V. C. M. Leung, "Mobility-based predictive call admission control and bandwidth reservation in wireless cellular networks," Proceedings IEEE INFOCOM 2001. Conference on Computer Communications. Twentieth Annual Joint Conference of the IEEE Computer and Communications Society (Cat. No.01CH37213), Anchorage, AK, USA, 2001, pp. 518-526 vol.1.
[34] R. Xie, F. R. Yu, H. Ji and Y. Li, "Energy-Efficient Resource Allocation for Heterogeneous Cognitive Radio Networks with Femtocells," in IEEE Transactions on Wireless Communications, vol. 11, no. 11, pp. 3910-3920, November 2012.
[35] Q. Guan, F. R. Yu, S. Jiang and G. Wei, "Prediction-Based Topology Control and Routing in Cognitive Radio Mobile Ad Hoc Networks," in IEEE Transactions on Vehicular Technology, vol. 59, no. 9, pp. 4443-4452, Nov. 2010.
[36] R. Xie, F. R. Yu and H. Ji, "Dynamic Resource Allocation for Heterogeneous Services in Cognitive Radio Networks With Imperfect Channel Sensing," in IEEE Transactions on Vehicular Technology, vol. 61, no. 2, pp. 770-780, Feb. 2012.
[37] F. Yu and V. Krishnamurthy, "Optimal Joint Session Admission Control in Integrated WLAN and CDMA Cellular Networks with Vertical Handoff," in IEEE Transactions on Mobile Computing, vol. 6, no. 1, pp. 126-139, Jan. 2007.
[38] K. S. Babulal, and R. R. Tewari, “Cross Layer Design for Cooperative
Transmission in Wireless Sensor Networks,” Wireless Sensor Network,
Vol. 3, pp. 209-214, 2011.
[39] C. Luo, F. R. Yu, H. Ji, and V. C. M. Leung, “Cross-Layer Design for
TCP Performance Improvement in Cognitive Radio Networks,” IEEE
Trans. Veh.Technol., Vol. 59, No. 5, pp. 2485-2495, 2010.
[40] U. Korger, C. Hartmann, K. Kusume, and J. Widmer, “Quality of service implications of power control and multiuser detection-based cross layer design,” EURASIP J. Wireless Communications and Networking,June 2011.
[41] R. W. Thomas, D. H. Friend, L. A. DaSilva, and A. B. MacKenzie,
“Cognitive Networks: Adaptation and Learning to Achieve End-to-End
Performance Objectives,” IEEE Commun. Mag., December 2006.
[42] A. Tanenbaum, Computer Networks, Prentice Hall, pp. 550, 2003.
[43] K. G. Zografos, K. N. Androutsopoulos, and V. Spitadakis, “Design and assessment of an online passenger information system for integrated multimodaltrip planning,” IEEE Trans. Intell. Transp. Syst., vol. 10, no. 2,pp. 311–323, Jun. 2009.
[44] A. Mishra, M. Shin, and W. Arbaugh, “An empirical analysis of the IEEE802.11 MAC layer handoff process,” SIGCOMM Comput. Commun. Rev.vol. 33, no. 2, pp. 93–102, Apr. 2003.
[45] B. Girod, K.W. Stuhlmueller,M. Link, and U. Horn, “Packet loss resilient Internet video streaming,” in Proc. PIE Visual Commun. Image, 1999, pp. 833–844.