Presentation on theme: "Concevoir et maîtriser les systèmes complexes 1 Urbanisme des Radio-Communications SP2, D2.1.2 : Specifications of the Metrology Related to Radio Resources."— Presentation transcript:
Concevoir et maîtriser les systèmes complexes 1 Urbanisme des Radio-Communications SP2, D2.1.2 : Specifications of the Metrology Related to Radio Resources and Spectrum Management 6 June 2007 Editor: PRiSM Lab
Concevoir et maîtriser les systèmes complexes 2 Content of D2.1.2 Centralised/Decentralised Metrology Concepts: Pending Definition of DSA Metrics: Pending Need of Cooperation between Different RATs belonging to the same Operator: In progress Withdrawn Tasks: Cooperation between Heterogeneous RATs and Different Operators Spectrum Sharing and Utilisation of Unlicensed Bands
Concevoir et maîtriser les systèmes complexes 3 Centralised/Decentralised Metrology Concepts Resp: ENST Status: Pending Metrology Classification: Centralised: At which network level/node(s)? Distributed: How to manage information exchange and decisions? Metrology and DSA Metrology Related to Radio Measurements in Île-de-France
Concevoir et maîtriser les systèmes complexes 4 Definition of DSA Metrics Resp: ENST; Other: PRiSM, INRETS, INT Status: Pending Possibilities of improving spectrum usage in Île-de- France. Connection to the radio measurements campaign?? Input Measurements for DSA Operations: to be completed Offered load of the RAN. At which level measurements are taken? Relating measurements to traffic and mobility models (parameters estimation). Grade of Service, GoS related to traffic and services Relation Application/Service RAT
Concevoir et maîtriser les systèmes complexes 5 Definition of DSA Metrics Blocking probability: –reject of new calls –Handover failures History of Offered load based on service nature –Streaming: Video –Telephony, visiophony –Interactive data –Elastic Issue: How to relate a service demand to the needed bandwidth? Interferences: connect interferences to service QoS degradation User inputs: –Radio metrics: C/I, eventually absolute interference level. =? Frame error rate.
Concevoir et maîtriser les systèmes complexes 6 Definition of DSA Metrics Metrics for DSA Performance Evaluation and Validation Testing (Output Metrics): Pending Spectrum allocation: fairness issues Comparison between FSA and DSA schemes QoS perceived by final users Performance and cost metrics to be used by operator Relation between Radio Measurements and DSA Metrics: Pending Standards specifications of radio measurements and how to translate them into DSA metrics. Take advantage of the measurements campaign Traffic measurements.
Concevoir et maîtriser les systèmes complexes 7 Need of Cooperation between Different RATs belonging to the same Operator Resp: PRiSM; Other: ENST, INRETS, INT Status: In progress Space-Time Variation of RATs Spectrum Demands: In progress Cooperation and Spectrum Exchange between RATs: In progress Control and Management Information Exchange between Heterogeneous Networks: to be completed Measurements and Metrics Implicated: Network oriented and User oriented RRM functions and RRM cooperation protocols Control Network for information exchange
Concevoir et maîtriser les systèmes complexes 8 Time Variation of Spectrum Demand Time-varying traffic Each service has its own temporal traffic patterns shape related to the human activity: –Ex: telephone activity culminates at the busy hours. Video demand increases in the evening. –The shape of a given service may depend on the geographical location. Periodic/predictable variations (daily or seasonal basis,...) Exceptional variations (events, network problems,...). Current spectrum status: FSA. Dimensionning and spectrum allocation is done based on busy hours.
Concevoir et maîtriser les systèmes complexes 9 Time Variable DSA (1) Temporal DSA Take advantage of the temporal variations. Dynamically adjust allocated bandwidth to the load. General Conditions for DSA on 2 RATs Temporal peaks located at different times A negative correlation is favorable Drive data Lisbon urban area: GSM
Concevoir et maîtriser les systèmes complexes 10 Time Variable DSA (2) Ideal DSA : Continuous measurements of offered load. Continuous spectrum values. In fact, DSA is discrete in time and spectrum allocation: What is the best period of DSA operation given a traffic shape? The minimum spectrum exchange unity is a RAT carrier. General constraints and issues: RATs may have different carrier sizes. GSM would have to free 40 carriers to activate a new DVB-T carrier. RAN and users harware must operate at different frequencies. (software radio)
Concevoir et maîtriser les systèmes complexes 11 Time Variable DSA (3) The time varying nature of the traffic implies the following DSA functions: Operational steps: Find the best temporal granularity that assures a good tracking of the demand variations. DSA interval. Load measurements and prediction: –Load histroy. –Prediction for the next DSA interval based on the history (e.g. regression methods). Spectrum allocation (contention and fairness issues).
Concevoir et maîtriser les systèmes complexes 12 Time varying traffic related to other tasks Relation to the measurement campaign: Get graphical shapes of daily load variations for different technologies in Île-de-France. Possible measurements at Base Stations? Relation to SP3: The simulator uses individual session/call models for the demand and traffic. The session/call models and arrival rates must be fitted so that their aggregation gives similar temporal shapes.
Concevoir et maîtriser les systèmes complexes 13 Space Variation of Spectrum Demand Spatial Spectrum demand depends on users traffic activity on each RAT and in each area Traffic modelling in space and measurements to be used Macroscopic granularity: call scale, i.e., minutes
Concevoir et maîtriser les systèmes complexes 14 Spatial Spectrum Demand Distribution Services/Applications used are space dependent and are strongly connected to underlying RATs Recreational Areas: DVB-T, HSDPA, WiMAX Business Areas: unicast, telephony -> GSM, UTRAN. Regional traffic models combine: Geographical distribution of terminal and traffic density Distribution of Applications/activities: Which application activity distribution (itself connected to RAT) for each area
Concevoir et maîtriser les systèmes complexes 15 Spatial DSA Constraints DSA areas geometry and neighbourhood constraints (coordination) Spatial gradient: different DSA schemes between adjacent areas Guard bands should be adapted when needed Reducing cell sizes to decrease a high spatial gradient
Concevoir et maîtriser les systèmes complexes 16 Spatial DSA Constraints (2) Integrate traffic correlation between areas Integrate traffic correlation between RATs Integrate total traffic demand parameter T(total) = Sum(T(RATi)) variability Global mean GoS should be maximized in the whole network Each area should enhance or at least equal FSA GoS Optimisation problem: maximising GoS gain under all listed constraints (interference, cost, guard band, …)
Concevoir et maîtriser les systèmes complexes 17 Space-Time correlation and Mobility Geographical traffic distribution is time-dependent Combine time and regional DSA. Rush hour:
Concevoir et maîtriser les systèmes complexes 18 Control and Management Metrics Measurements and Metrics Implicated Network oriented: at AN devices –Load measurements –Global performance: blocking probabilities User oriented: DSA may be involved when a global degradation of users QoS occurs => statistics collected from users
Concevoir et maîtriser les systèmes complexes 19 Control and Management RRM RRM functions and RRM cooperation protocols depend on the DSA nature: Contiguous DSA: Control Network for information exchange RRM inside a RAT, carriers redistribution. Readjustement of carriers (condensing) Interference issues.
Concevoir et maîtriser les systèmes complexes 20 Cooperation between Heterogeneous RATs and Different Operators Defining a network architecture for control and spectrum exchange Defining appropriate protocol for the control plane Centralised, distributed architecture? Coordination between different RAT and Areas TD: Traffic Distribution IU: Interface Unit RS: RAN Selection SM: Service Manager EO: Efficiency Optimisation Unit TM: Traffic Measurer TP: Traffic Predictor
Concevoir et maîtriser les systèmes complexes 21 Cooperation between heterogeneous RATs and Different Operators FT, TDF Common Architecture Definition for Operators Cooperation (broker) Cooperation Constraints and Mechanisms between different operators and heterogeneous Systems Spectrum Sharing Mechanisms
Concevoir et maîtriser les systèmes complexes 22 Spectrum Sharing and Utilisation of Unlicensed Bands FT, ENST, INRETS, INT Unlicensed Bands Identification and Characterisation Analysis and Study of DSA Achievability in Unlicensed Bands