VIEWS: 44 PAGES: 8 CATEGORY: Mobile Devices POSTED ON: 7/17/2011
B3G: Beyond Third Generation in mobile communication system, as opposed to 3G mobile communications, B3G has a higher transmission efficiency and all types of business.
Motivation • The future leads towards an ambient where wireless communications Cross Layer Roadmap to Cross-Layer Optimization will exist in every scenario of life to provide the end user the “flexibility and choice”. • The beyond 3G communication vision concerns scenarios where Department of Information and Communication Systems Engineering service user will be able to attain any service, at any time on effectively any University of the Aegean network that is optimised for the application at hand. Karlovassi, 83200, Samos, Greece – Thus creating a future global infrastructure, where several systems can coexist to support transparent end-to-end communications, in an efficient, cost-effective manner; ABCS (Always Best Connected Systems) philosophy. • An important architectural issue is that of defining a next-generation wireless system, which acts as a “network-of-wireless-networks” g y g accommodating a variety of radio technologies and mobile service requirements in a seamless manner. – Significant challenges in view of the devices heterogeneity, systems requirements. scale and network robustness requirements George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 1 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 2 Motivation Motivation • The next generation of mobile technology will probably incorporate • At an inter-system level the coexistence of communication entities innovations at many levels: raises several issues: – At the device level new technologies such as UWB (Ultra Wideband) and • First the quantification of interference levels, due to coexistence. MIMO (Multiple Input Multiple Output) have the potential to significantly – Many weak signals can accumulate and become strong enough to cause increase the physical data rates; bl i ti t k problems on coexisting networks. – At the networking level multi-access technologies are proposed to • Second of interoperability, given that currently Radio Access improve the reachability, availability and efficiency of wireless networks. Network (RAN) systems are designed and deployed to operate level, • At an intra-system level previous theoretical developments have ffi i tl in t i efficiently i a mono-system environment. t indicated strong interactions between the PHY layer signalling schemes and channel conditions with the networking issues of • The ultimate vision of a network of wireless networks where the user resource allocation. connects to the best available system, while conceptually simple, brings significant challenges: b i i ifi t h ll • Several problems traditionally considered as networking issues and – Detection: What accesses are currently available? are typically designed independently of the transmission techniques – Selection: What access to choose; which is “best”, one or multiple re examined. need to be re-examined accesses in parallel? • Cross-layer design and optimisation will be the key to adequately – Reselection: Under what conditions reselection is necessary; which exploit all the potential of next generation wireless communication network to choose? How reselection is accomplished? technologies technologies. George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 3 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 4 Cross-Layer Optimisation Cross-Layer Optimisation • Layering is the dominating design methodology of • To cope with this variability of the wireless channel, communication protocol stacks. An essential feature techniques have been developed either at the physical or at the MAC layer. of the layering principle is layer-independence • The main shortcoming of the strictly based PHY layer (modularity), which represents the classical approaches is that they do not take into account the impact on engineering approach to solve complex problems. the upper layers. • However, the strict layering approach presents • The main shortcoming of strict layering MAC layer based significant limitations when it comes to wireless schemes is that they are based on “hard” channels, i.e. they communications. i ti layer. use very limited information from the physical layer This implies that the throughput achievable at the upper layers is • The wireless networks are characterised by a only a small fraction of the capacity offered by the PHY layer. d i topology ( dynamic t l d l t (users move around, also enter • For example, consider IEEE 802.11a & g (fastest on the and leave); the link exhibits a time-varying quality market): the highest PHY rate is 54 Mbps, but above the MAC fading, shadowing due to fading shadowing, in addition to multiuser Mbps there is always less than 30 Mbps, and with RTS/CTS less than interference. 24 Mbps is achieved, even under no contention. Page 5 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 6 Cross-Layer Optimisation Cross-Layer Optimisation and Coexistence • The loss of efficiency is further aggravated if instead • It is also easy to anticipate that if the optimum of considering only the two bottom layers, one performance is not achieved at the intra-system level considers E2E performance involving the whole because of cross-layer inefficiencies, performance is stack. further hampered when considering the coexistence • For example, certain routing protocols require that of heterogeneous systems, which will be the nodes broadcast control messages to all neighbours di f paradigm for B3G scenarios. i and wait for all of them to reply. This imposes a • One of the key aspects to get the best from wireless t d t i th tremendous strain on the medium access layer. Also, di l Al t ill l systems will rely on: routing protocols tend to treat all neighbours equally – Cross-layer interactions; accessible accessible. Definition f ffi i t t d – D fi iti of efficient cross-system procedures. • It is very important that the designers of the routing protocol of a system are aware of the limitations of the MAC protocol of that system, and vice versa. Page 7 Page 8 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Cross Layer Interactions Cross Layer Interactions • A cross-layer architecture encompasses an • No universal answer to both questions. additional complexity relatively to a strictly-layered • The answer to the first question depends on the one, due to the fact that additional information specific algorithm at the given protocol layer. p g g p y besides the one that defines the basic service • The answer to the second question also depend on provided by the layer has to be exchanged. The need several algorithm dependent aspects (centralised vs. g p p ( t exchange additional cross-layer information (CLI) to h dditi l l i f ti distributed) and on the specific systems leads to two fundamental questions: architecture. – What information should be exchanged across protocol layers, and, how frequently should this exchange proceed? Important Note – What are the adequate / efficient procedures to exchange • Although, the use of cross-layering has the potential g , y g p this information? to significantly improve E2E throughput, its nature obviously implies some loss of flexibility which is inherent of the modular structure of a layered stack. George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 9 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 10 Cross Layer Interactions Definition of efficient cross-system procedures • There are therefore significant challenges • Beyond the problem of cross-layering optimisation associated with the development of a cross-layered that occurs at the intra-system level, the next protocol stack: problem that will be of paramount importance in the • Architectural challenges: “network of wireless networks” scenario envisioned – How to answer previous questions without excessive for B3G will be the definition of efficient, and secure overheads? t d ll i th cross-system procedures allowing the user to gett t – How far to go in the cross-layering principle? always the best (quality, cheap, secure, etc.) system – Wh t are th t d What ff between improved th the trade-offs b t i h t d d throughput and from the network network. loss of modularity? George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 11 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 12 Definition of efficient cross-system procedures Dynamic Resource Allocation (DRA) Issues • A cross-layer optimised radio access may lose its optimal • Research focuses on how to exploit additional cross- attributes in the presence of a heterogeneous network. system information (e.g. QoS metrics from IPv6 • Thus there is a need for further L1-to-L3 algorithmic packets, accurate routing information, knowledge of mono-system optimisation in a bid to achieve virtual mono system resource allocation strategies in neighbouring performance, and ensure transparent connectivity to multiple networks. systems, current system interference statistics, • This raises new design challenges in algorithmic design: d l i underlying PHY l d h l t layer and channel parameters) to)t – Which set of additional cross-system information is required, and allocate available radio resources to attain a trade- what signalling overhead is required to deliver this information throughput, off among maximising cell service throughput and at which rate? minimising signalling overhead, reducing complexity, p y, q y • Specifically, solutions are required for current dynamic users. and achieving the intended priority between users resource allocation, and mobility protocols to ensure both efficient, and seamless connectivity and mobility between achieved. wireless systems can be achieved George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 13 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 14 Cross-system Mobility Issues Cross-Layer Optimization Current Status • Although mobility protocols at L3 have extensively been • Physical layer engineers use specific tools where they investigated, there is a need: integrate very detailed models, and even at the PHY layer – To address a cohesive solution between L2, and L3, so that the there are differences in terms of time-scale for cross-system exchanged. relevant cross system parameters are exchanged base-band designs. simulations between the RF and base band designs – To study the type of cross-system parameters, and signalling • At the upper level, radio planning tools or network cost, which will have implications on service throughput. simulators basically rely on simple abstract models simulators, – To analyse the detailed implications of the L3 mobility protocols on L2 handover. describing the behaviour of the lower layers. – To consider a joint design for the L2 L3 mobility protocol so L2-L3 protocol, p – While for some scenarios simplified abstract models may be y seamless mobility and connectivity between systems can be sufficient and do not significantly compromise accuracy of the achieved. simulation results, they are often in place because detailed models are too difficult to implement and to run efficiently. George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 15 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 16 “Cross-Layer Design: A Survey and the Road Ahead” V. Srivastava, M. Motani Review on Cross-Layer Algorithms and Design Issues Magazine vol. 43, no.12, Dec. IEEE Communications Magazine, vol 43 no 12 Dec 2005 • the upward information flow, where specific information parameters from lower layers are used from upper layers, • downward information flow, parameters from upper layers are used for optimizing behaviour of lower layers, • back-and-forth information flow, , which is the combination of the first two cases, • merging of adjacent layers, which creates a ‘super’ layer from the cross layer Different kinds of cross-layer design proposals combination of two existing ones, • the design coupling without new Implementation: • Layer-to-Layer direct Communication via new interfaces interfaces, , • Layers share a common database • the vertical calibration, where the • Create new abstractions (no more typical layering approach) performance seen at the level of the application is a function of the t t ll th l b l it parameters at all the layers below it. George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 17 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 18 “A Framework for Cross-layer Design of Energy-efficient Communication with QoS Provisioning “Cross-Layer-Based Modeling for Quality of Service Guarantees in Mobile Wireless Networks” in Multi-hop Wireless Networks” X. Jhang, J. Tang, H. Chen, S. Chi, M. Guizani, C. Kozat, I. Koutsopoulos, L. Tassiulas Magazine vol. 44, no.1, Jan. IEEE Communications Magazine, vol 44 no 1 Jan 2006 IEEE INFOCOM-2004 , Hong Kong, 2004 work In this work, cross – layer design is used between This cross-layer approach investigates the impact of the physical and MAC layers. Power control and physical layer on the data-link layer in mobile wireless Scheduling are jointly used in order to achieve networks. minimizing the total transmit power, and at the same At the physical layer MIMO diversity schemes as well as AMC time maintain the proper QoS levels, in terms of account. (Advanced Modulation Coding) are taken into account rate. bandwidth and bit error rate At the data-link layer, the influence of the physical Input: Link gains, QoS parameters infrastructure on real time multimedia performance is p studied. O t t T i i P f h d Output: Transmission Power for each node Input: Physical layer diversity schemes (number of transmit antennas) QoS reliability requirements (packet error rate - PER) Output: Effective capacity George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 19 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 20 “Topology-Aided Cross-Layer Fast Handoff Designs for IEEE 802.11/Mobile IP Environments” “Cross-Layer Feedback Architecture for Mobile Device Protocol Stacks” C. Tseng, L. Yen, H. Chang, K. Hsu V. Raisinghani, S. Lyer IEEE Communications Magazine, vol. 43, no.12, Dec. 2005 vol. 44, no.1, Jan. IEEE Communications Magazine vol 44 no 1 Jan 2006 p gy q Topology-aided fast handoff technique for 802.11 p p g g y This paper highlights the need for cross-layer feedback networks. architecture and identifies key design goals for such Layer 3 handoff-related procedures (e.g., agent architectures. discovery address configuration and registration) can be discovery, Cross layer information exchange can be realised by: Cross-layer performed in parallel with or prior to those of layer 2 • additional blocks of code into layers, handoff (i.e. probe, authentication and reassociation). • the usage of ICMP messages, g g , cross-layer information Using cross layer topology information, such as association between APs and Mobile IP mobility agents, • injection of cross-layer information into packet headers handoff related layer-2 triggers may reduce the delay and between layer-2 handoff completion and the associated b t l 2 h d ff l ti d th i t d i lli h i • signalling mechanisms. layer-3 handoff activation. This way, the handoff latency In ECLAIR, a tuning layer (TL) for each layer, provides an time (e.g., AP probe delay, authentication delay, move interface to read and update these protocol data-structures. p p detection d l t di d l i t ti d l d t ti delay, agent discovery delay, registration delay, TLs are then used by protocol optimizers (POs), which etc.) can be reduced. contain cross-layer feedback algorithms, allowing even user p p p gy Input: Link performance, topology information (e g feedback (e.g., application prioritization) for performance Output: Handoff triggers optimization. George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 21 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 22 “Cross-layer Design of Ad Hoc Networks for Real-Time Video Streaming” “Cross-Layering in Mobile Ad Hoc Network Design” E. Setton, T. Yoo, X. Zhu, A. Goldsmith, B. Girod Marco Conti, Gaia Maselli, Giovanni Turi, Silvia Giordano Magazine IEEE Wireless Communications Magazine, August 2005 Computer IEEE Computer, February 2004 This paper explores a cross-layer design framework p p p y g Advocates that the level of cross – for real-time video streaming, which maintains a layering that is necessary is not clear. general layered structure and identifies the key parameters to be exchanged between adjacent It proposes a new architecture for layers. MANETs that introduces a new core • Link layer information adjusts packet size, symbol rate, and constellation size according to channel component, Network Status, conditions. responsible for the cross – layering •TTransport layer i f tl ti information enables smart bl t d t sharing. data h i scheduling protecting video stream, and • MAC and Network layer information adjusts the Through this new architecture, the supported traffic rate. pp y g cross layering that is vital for functions Layers/systems referred to: Application, Transport, like energy management is present, all Network, MAC and Link Layer layers are capable of knowing Input: Symbol rate, constellations size, channel conditions capacity, video rate distortion and conditions, capacity layers, information from all other layers and latency requirements Cross-layer design for real-video streaming in ad-hoc networks at the same time, keep overhead to Output: Packet size, traffic rate, packet scheduling, acceptable levels. MobileMan architecture. appropriate source rate George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 23 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 24 “Design and Implementation of a Simulator Based on a Cross-Layer Protocol between MAC and PHY Layers in a WiBro Compatible IEEE 802.16e OFDMA System” “Multivariate Analysis of the Cross-Layer Interaction in Wireless Networks Simulations” T. Kwon, H. Lee, S. Choi, J. Kim, D. Cho, S. Cho, S. Yun, W. Park, K. Kim Jean-Michel Dricot, Philippe De Doncker, Esteban Zimànyi IEEE Communications Magazine, December 2005 Ad-hoc International Workshop on Wireless Ad hoc Networks 2005 In this paper cross-layer design frameworks for 802.16e cross layer p p y This paper uses the mathematical tool of ANalysis Of VAriance OFDMA systems that are compatible with WiBro are (ANOVA) in multivariate data, in order to study the interactions between the layers of a wireless system. proposed The analysis p y y g y proves not only that a single layer can affect the In the proposed model, the MAC layer contains a user whole network operation, but also reveals the existence of grouper, scheduler, and resource controller. Each interesting interactions between the layers of the system, as functional entity exploits PHY information to increase system shown between four variables: throughput. • routing algorithm, • propagation conditions, In addition, a simulation framework for cross-layer analysis density, • nodes density and between the MAC and PHY layers in 802.16e systems is • mobility scheme. provided indicating improvement for the average cell The study also reveals a correlation of the routing protocol with throughput throughput. the type of environment (indoor / outdoor Output: User and packet scheduling George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 25 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 26 “A Cautionary Perspective on Cross-Layer Design” Vikas Kawadia, P. R. Kumar Projects for Cross-layer Optimization Magazine IEEE Wireless Communications Magazine, February 2005 The trend to use cross – layer interaction in order to improve y p • The cross-layer optimisation projects are currently wireless networks’ performance is significant in our days. being targeted to attain an optimised solution, using This paper is on the opposite side, defending the traditional architecture. a cross profile support entity to gather and maintain In order to prove its point it conducts some simulation comparing user-specific and system information from all layers, some cross – layer algorithms with others not using this approach. e.g., MAGNET, 4MORE, WINNER. Agree that cross – layering is the only way to optimize certain parameters, and suggest points that protocol designers must consider: modularity, • loss of modularity • the fact that every time a protocol is replaced, there will probably be a need for total redesign of the algorithm, • the dependencies between parameters shared between layers, • guarantee that the interactions between protocols are only those intended and the absence of unintended interactions. George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 27 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 28 Projects for Cross-layer Optimization Project for Cross-layer/Cross-system Optimization • The WHYNET project (Scalable Testbed for Next Generation • UNITE builds through open APIs a VDT of Mobile Wireless Networking Technologies) [WHYNET] clusters/systems giving also the chance for cross- proposes to design and develop a Wireless HYbrid NETwork testbed to facilitate a detailed study of cross-layer cross layer system and cross-layer operational scenarios interactions, that extend from the application layer down to the running horizontally among a number of physical devices, and their impact on application level heterogeneous clusters. f i h t i l performance in heterogeneous wireless systems. t • The hybrid testbed is a networked federation of geographically distributed heterogeneous wireless physical testbeds with distributed, multiple protocol stacks. • The testbed itself will be accessible by the research community via a web-based mechanism that will allow remote uploading of models, implementations, and configurations. George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 29 George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 30 THANK YOU George Kormentzas, “Roadmap to Cross-Layer Optimization for B3G” Page 31
"Roadmap to Cross-Layer Optimization Roadmap to Cross Layer "