(Cross-layer design and network planning for B3G systems) Aspects of Coexistence Between WiFi and HSDPA Orlando Cabral Valdemar Monteiro 2005, it - instituto de telecomunicações. Todos os direitos reservados.
Outline Global view of Cross-layer design and network planning for B3G systems (CROSSNET) Cross-layer design Cross-system and convergence Cross-layer studies for IEEE 802.11e Differentiation mechanisms Simulator validation Cross-layer design proposal and results RAT selection for HSDPA and IEEE 802.11e Interoperability scenario Algorithm description Results and scientific impact Conclusions and future work 2 Lisboa, 19 de Fevereiro de 2008
Global view of CROSSNET - Objectives and work Organization The project was divided into three main areas: Traffic modelling and Quality of Service (QoS) Medium Access Control (MAC) definition and cross-layer design Network planning and optimisation, synthesis and integration In a scenario of heterogeneous networks, the interoperability among available radio access technologies should be seen by mobile users as the ability to successively attach to the access points and bases stations in a transparent way 3 Lisboa, 19 de Fevereiro de 2008
Coexistence 4 Lisboa, 19 de Fevereiro de 2008
Global view of CROSSNET - Cross-layer Traffic Model NodeB DRA and Implementation Queues and validation Radio of the High-Speed Channel Models Resources and Interference Downlink Packet Access (HSDPA) IEEE 802.11e Power H- MAC protocol L L I ARQ H- text Scheduler Link ARQ Adaptation Cell Deployment and HSDPA R5 system level simulator facilitates the design, Mobile Dynamics CQI evaluation and optimization of scheduling and linkadaptation Queue Manager policies Code Slot UE ARQ msg (Ack/Nack) SIR 5 Lisboa, 19 de Fevereiro de 2008
Global view of CROSSNET - Cross-system MOTION (HSDPA) and WE-MOVE (WiFi) simulation tools were integrated under a common simulation platform to perform Common Radio Resource Management studies RAT selection for multimedia services Also UMTS (HSDPA) WiMAX interoperability is being considered, however at this stage the WiMAX simulation platform is being validated for IEEE 802.16e Dynamic Resource Allocation strategies 6 Lisboa, 19 de Fevereiro de 2008
IEEE 802.11e IFSs relationships Priorities SIFS (Short Inter Frame Spacing, IFS) highest priority, for ACK, CTS, polling response PIFS (PCF IFS) medium priority, for time-bounded service using PCF AIFS (EDCA, arbitrary IFS, used by the QoS facility) Priority depends of the access category. 7 Lisboa, 19 de Fevereiro de 2008
IEEE 802.11e Block ACK 2.5E+07 The Block Ack mechanism VO is 2.0E+07 initialized by an exchange VI-BLOCK of ADDBA Request/Response BK-BLOCK 1.5E+07 frames. After 1.0E+07 initialization, blocks of QoS data frames can be 5.0E+06 transmitted from the originator to the recipient. 0.0E+00 Goodput [b/s] BK VI 0 10 20 30 40 50 Stations 8 Lisboa, 19 de Fevereiro de 2008
IEEE 802.11 - State transition diagram 9 Lisboa, 19 de Fevereiro de 2008
Validation PHY layer without fragmentation 40 Goodput [Mb/s] 35 30 25 20 15 10 5 0 Mod 8 Mod 7 Mod 6 Mod 5 Mod 4 Mod 3 Mod 2 Mod 1 0 5 10 15 20 25 30 SINR [db] 10 Lisboa, 19 de Fevereiro de 2008
Cross-layer concept used OSI Model BK BE VI VO Backoff AIFS [0] BC [0] Backoff AIFS [1] BC [1] Backoff AIFS [2] BC [2] Backoff AIFS [3] BC [3] Virtual Collision Handler Transmission Attempt 11 Lisboa, 19 de Fevereiro de 2008
Cross-Layer for WiFi EDCA mode Scenario Description WiFi system where a cell has a set N+1 IEEE 802.11e stations communicating through the same wireless channel Station 0 is the Access Point or QoS Access Points (QAP), and the other N wireless terminals are QoS stations (QSTA) Novelties of Algorithm Dynamically changing the Arbitration Inter Frame Spacing (AIFS) and the contention window (CW) size, using as inputs delay and channel state information, can lead to enhanced channel utilization and system QoS By sharing network state information among users, optimal radio resource allocation parameters can be attained 12 Lisboa, 19 de Fevereiro de 2008
Cross-Layer for WiFi EDCA mode Algorithm Description The algorithm is based on dynamically changing the CW and the AIFS to assign a priority Fairness is provided by using as an input the metric PhyDataRat ej M j =. AverageThroughput j ( p) QoS is ensured by using as an input DelayWeight CurrentDel ay K AC AC( CurrentDel ay) = C AC. Delay per packet [m s] 500 400 300 200 100 0 with scheduler-ap without scheduler-ap 0 10 20 30 40 50 Stations Results The proposed cross-layer algorithm showed enhanced performance for voice traffic can support at least 35% extra stations with QoS Future work will target optimisation for the remaining service classes (background, streaming, and best effort) 13 Lisboa, 19 de Fevereiro de 2008 Goodput [b/s] 3.E+06 2.E+06 1.E+06 0.E+00 without scheduler-ap with scheduler-ap 0 10 20 30 40 50 Stations
Conclusions - 802.11e Cross-layer Studies The IEEE 802.11e event driven simulator allows the evaluation and optimization of MAC protocols that includes acknowledgment policies and radio resource allocation strategies Utilizing a small CW for the voice access category leads to reduced goodput when the number of stations exceeds 65% Due to increased collisions causing the packet losses to exceed the retransmission threshold The proposed cross-layer algorithm showed enhanced performance for voice traffic Can support at least 35% extra stations with the required QoS 14 Lisboa, 19 de Fevereiro de 2008
Future work - 802.11e Cross-layer Studies This simulator will be extended: to address the evaluation of interoperability protocols for WiFi and HSDPA (in the context of the IT-MOTION tool) to investigate and validate Common Radio Resource Management (CRRM) techniques for HSDPA, WiMAX and WiFi Future work will target: The joint application of block acknowledgment and mixed traffic. Handover policies between APs for multi-cell scenarios 15 Lisboa, 19 de Fevereiro de 2008
RAT selection for HSDPA and IEEE 802.11e 16
RAT selection for HSDPA and IEEE 802.11e - Scenario Description Common coverage scenario with UMTS cell and with a WiFi hotspot IP-based core network assumed that acts as the bridge between WiFi, and HSDPA NRTV services that can be streamed either over HSDPA, or WiFi system 17 Lisboa, 19 de Fevereiro de 2008
RAT selection for HSDPA and IEEE 802.11e - Algorithm description Suitability calculation for RAT selection among cells when new call is requested Targeted to flexible traffic Flexibility imposes on the system, meaning that the service can be held by each RAT δ S (netw ork-centric suitability) 1 2 S( L( cell i, j)) = 1 L( cell i, j) 1 LThj if L( cell ) LTh i, j if L( cell ) > LTh i, j j j 1 0 0 1 η LTh j LThj is the threshold load for RAT j 18 Lisboa, 19 de Fevereiro de 2008
RAT selection for HSDPA and IEEE 802.11e -Novelties of Algorithm RAT selection based on load suitability where the systems are HSDPA and WiFi, under delay constraints services Concept of suitability is used in terms of preferred access system to accommodate the service, where Suitability can change as load increase The goal is to optimize the load in each RAT, measuring the diversity gain in quality of service (QoS) provisioning 19 Lisboa, 19 de Fevereiro de 2008
RAT selection for HSDPA and IEEE 802.11e - Results and Scientific Impact Results Gain of 60% on the supported QoS load achieved with the CRRM (intelligently adapted algorithm) over HSDPA alone Scientific impact QoS/ service throughput (Kbps) 3000 2500 2000 1500 1000 500 0 CRMM vs HSDPA QoS Throughput CRRM Goodput NRTV offered load HSDPA QoS 8 14 20 26 32 38 44 Number of user Exploring Common Management of the whole Radio Resources such that the mobile assignment to different RATs is flexible and applicable to heterogeneous radio systems 20 Lisboa, 19 de Fevereiro de 2008
Conclusions and future work - HSDPA / WiFi CRRM studies CRRM was employed exploring the common management of the whole radio resources such that the assignment of mobiles to different radio technologies is flexible and considers the specific features of each system The goal of the proposed CRRM is to select the most appropriate radio network in co-located systems without deteriorating the quality of service Thanks to the CRRM algorithm, a 60% gain on system throughput (with QoS) can be achieved in a co-located HSDPA- WiFi system over HSDPA alone 21 Lisboa, 19 de Fevereiro de 2008
Future work - HSDPA / WiFi CRRM studies CRRM future work includes Evaluation of the proposed algorithm for the HSDPA and WiMAX 802.16e Improvement on the algorithm to include load balancing between HSDPA and 802.16e for established connections 22 Lisboa, 19 de Fevereiro de 2008 Thank you! http://www.e-projects.ubi.pt/crossnet