Range Mindful Burden Adjusting for WLANs.


41 views
Uploaded on:
Category: Sports / Games
Description
Versatile Channel Width (ACW). Versatile Channel Width is a key empowering innovation for Cognitive Radio NetworkingWhy? . Versatile Channel Width (ACW). Versatile Channel Width is a key empowering innovation for Cognitive Radio NetworkingWhy? . Decent Properties (reach, power, throughput). Application: Music sharing, specially appointed correspondence,
Transcripts
Slide 1

Range Aware Load Balancing for WLANs Victor Bahl Ranveer Chandra Thomas Moscibroda Yunnan Wu

Slide 2

Adaptive Channel Width (ACW) Adaptive Channel Width is a key empowering innovation for Cognitive Radio Networking Why?

Slide 3

Adaptive Channel Width (ACW) Adaptive Channel Width is a key empowering innovation for Cognitive Radio Networking Why? Decent Properties (go, control, throughput) Application: Music sharing, specially appointed correspondence, …

Slide 4

Adaptive Channel Width (ACW) Adaptive Channel Width is a key empowering innovation for Cognitive Radio Networking Why? Adapt to Fragmented Spectrum (Primary clients) Application: TV-Bands, White-spaces, …

Slide 5

Adaptive Channel Width (ACW) Adaptive Channel Width is a key empowering innovation for Cognitive Radio Networking Why? (Another handle for) Optimizing Spectrum Utilization Application: Infrastructure-based systems ! This discussion!

Slide 6

Outline Adaptive Channel Width is a key empowering innovation for Cognitive Radio Networking Nice Properties (run, control, throughput) Cope with Fragmented Spectrum This discussion Optimizing Spectrum Utilization Cognitive Networking MATH… ? Models Algorithms Theory This discussion MATH

Slide 7

Infrastructure-Based Networks (e.g. Wi-Fi) Each customer partners with AP that offers best SINR Hotspots can show up ��  Client throughput endures! Thought: Load-Balancing

Slide 8

Previous Approaches - 1 Change relationship amongst customers and get to focuses (APs) e.g. [ Bejerano , Mobicom\'04] , [ Mishra , Infocom\'06]

Slide 9

Previous Approaches - 1 Change relationship amongst customers and get to focuses (APs) e.g. [ Bejerano , Mobicom\'04] , [ Mishra , Infocom\'06] Problem: Clients interface with far APs Lower SINR ��  Lower datarate/throughput

Slide 10

Previous Approaches – 1I Cell-breating : Use transmission powers for load adjusting e.g. [Bahl et al. 2006]

Slide 11

Previous Approaches – 1I Cell-breating : Use transmission powers for load adjusting e.g. [Bahl et al. 2006] Problem: Not generally conceivable to accomplish great arrangement Clients still associated with far APs TPC - Difficult practically speaking

Slide 12

Previous Approaches – III Coloring: Assign best (slightest congested) channel to most-stacked APs e.g. [ Mishra et al. 2005] Channel 1 Channel 1 Channel 2 Channel 2 Channel 1 Channel 3 Channel 3 Channel 2 Channel 3 Channel 1 Channel 2 Channel 3

Slide 13

Previous Approaches – III Coloring: Assign best (minimum congested) channel to most-stacked Aps e.g. [ Mishra et al. 2005] Channel 1 Channel 1 Channel 2 Channel 2 Problem: Good thought – yet constrained potential. ��  Still just a single channel for every AP ! Channel 1 Channel 3 Channel 3 Channel 2 Channel 3 Channel 1 Channel 2 Channel 3

Slide 14

Load-Aware Spectrum Allocation Our thought: Assign range where range is required! (Versatile Channel Width) ��  ACW as a key handle of improving range usage

Slide 15

Load-Aware Spectrum Allocation Our thought: Assign range where range is required! (Versatile Channel Width) ��  ACW as a key handle of advancing range use Advantages: Assign Spectrum where range is required Clients can remain related to ideal AP Better per-customer decency conceivable Channel cover can be stayed away from ��  Conceptually, it appears the regular method for taking care of the issue

Slide 16

Load-Aware Spectrum Allocation Problem definition: Assign (non-meddling) range groups to APs to such an extent that, Overall range usage is expanded Spectrum is doled out reasonably to customers Trade-off Assignment with ideal range use : ��  All range to leafs! Stack: 2 Load: 2 Load: 2 Load: 2 Load: 2

Slide 17

Load-Aware Spectrum Allocation Problem definition: Assign (non-meddling) range groups to APs to such an extent that, Overall range use is amplified Spectrum is alloted reasonably to customers Trade-off Assignment with ideal range usage : ��  All range to leafs! Task with ideal per-stack reasonableness: ��  Every AP gets half the range Load: 2 Load: 2 Load: 2 Load: 2 Load: 2

Slide 18

Our Results [Moscibroda et al. , submitted] Different range designation calculations 1) Computationally costly ideal calculation Computationally less costly guess calculation ��  Provably productive even in most pessimistic scenario situations Computationally cheap heuristics Significant increment in range usage!

Slide 19

Why is this issue fascinating? Customary channel task/recurrence task issues guide to chart shading issues (or variations thereof!) 2 6 2 5 2 1. Spatial reuse (like shading issue) 1 Self-instigated discontinuity 2. Stay away from self-initiated fracture (no proportionate in shading issue) MATH ��  Fundamentally new issue space ��  More troublesome than shading!

Slide 20

MATH Cognitive Networks: Challenges Models: New remote correspondence standards (network coding, versatile channel width, … .) ��  How to show these frameworks? ��  How to outline calculations for these new models… ? ��  Changes in models can have colossal effect! (Case: Physical model versus Convention demonstrate!) ��  Understand relationship between models

Slide 21

Example: Graph-based versus SINR-based Model Hotnets\'06 IPSN\'07 A needs to sent to D, B needs to send to C (single recurrence!) B A C D 4m 2m 1m SINR-based models (Physical models) ��  Possible Graph-based models (Protocol models) ��  Impossible Models impact convention/calculation outline! ��  Better conventions conceivable when thinking in new models Thomas Moscibroda, Microsoft Research

Slide 22

Example: Improved "Channel Capacity" Consider a channel comprising of remote sensor hubs What throughput-limit of this channel...? time Channel limit is 1/3 Thomas Moscibroda, Microsoft Research

Slide 23

Example: Improved "Channel Capacity" No such (diagram based) technique can accomplish limit 1/2! For certain remote settings, the accompanying technique is better! time Channel limit is 1/2 Thomas Moscibroda, Microsoft Research

Slide 24

Cognitive Networks: Challenges MATH Algorithms/Theory: Cognitive Networks will conceivably be gigantic Cognitive calculations are neighborhood, dispersed calculations! Hypothesis of nearby processability ! [PODC\'04, PODC\'05, ICDCS\'06, SODA\'06, SPAA\'07 ] 1) Certain undertakings are inalienably worldwide MST (Global) Leader race Count number of hubs 2) Other errands are inconsequentially nearby Count number of neighbors and so on 3) Many issues are "in the center" Clustering, neighborhood coordination Coloring, Scheduling Synchronization Spectrum Assignment, Spectrum Leasing Task Assignment Thomas Moscibroda, Microsoft Research

Slide 25

Summary Load-adjusting in framework based systems Assign range where range is required! Enormous potential for better decency and range use Building frameworks and applications essential! In any case, likewise a lot of in a general sense new hypothetical issues ��  new models ��  new algorithmic standards (calculations for new models) ��  new hypothetical underpinnings MATH

Recommended
View more...