IEEE in the Large: Observations at the IETF Meeting

Slides:



Advertisements
Similar presentations
Load Balance Based Access Point Association in Wireless Mesh Networks Ma Tianze Computer Network and Protocol Testing Laboratory, Dept. of Computer Science.
Advertisements

Fast L3 Handoff in Wireless LANs Andrea G. Forte Sangho Shin Henning Schulzrinne.
College of Engineering Optimal Access Point Selection and Channel Assignment in IEEE Networks Sangtae Park Advisor: Dr. Robert Akl Department of.
Application, Network and Link Layer Measurements of Streaming Video over a Wireless Campus Network Passive & Active Measurement Workshop 05 Boston, MA,
Analysis of an IEEE Network Activity during a Small Workshop Zhe Zhou, Mark Claypool, and Robert Kinicki Computer Science, Worcester Polytechnic.
Experimental Measurement of VoIP Capacity in IEEE WLANs Sangho Shin Henning Schulzrinne Department of Computer Science Columbia University.
Cooperation Between Stations in Wireless Networks Andrea G. Forte and Henning Schulzrinne Department of Computer Science Columbia University, New York.
1 Solutions to Performance Problems in VOIP over Wireless LAN Wei Wang, Soung C. Liew Presented By Syed Zaidi.
In-Band Flow Establishment for End-to-End QoS in RDRN Saravanan Radhakrishnan.
WBest: a Bandwidth Estimation Tool for IEEE Wireless Networks Presented by Feng Li Mingzhe Li, Mark Claypool, and.
VoIP over Wireless LANs Sangho Shin Ph.D. Candidate Department of Computer Science Columbia University.
IEEE in the Large: Observations at the IETF Meeting Henning Schulzrinne, Andrea G. Forte, Sangho Shin Department of Computer Science Columbia University.
VoIP over Wireless LANs Sangho Shin Ph.D. Candidate Department of Computer Science Columbia University.
Fast Wireless Handoff in Networks Sangho Shin Andrea G. Forte Anshuman S. Rawat Henning Schulzrinne.
Handoff Delay for b Wireless LANs Masters Project defense Anshul Jain Committee: Dr. Henning Schulzrinne, Columbia University Dr. Zongming Fei, University.
Experimental Measurement of the Capacity for VoIP Traffic in IEEE WLANs Authors : Sangho Shin, Henning Schulzrinne [INFOCOM 2007] Reporter : 林緯彥.
1 Algorithms for Bandwidth Efficient Multicast Routing in Multi-channel Multi-radio Wireless Mesh Networks Hoang Lan Nguyen and Uyen Trang Nguyen Presenter:
COGNITIVE RADIO FOR NEXT-GENERATION WIRELESS NETWORKS: AN APPROACH TO OPPORTUNISTIC CHANNEL SELECTION IN IEEE BASED WIRELESS MESH Dusit Niyato,
Divert: Fine-grained Path Selection for Wireless LAN Allen Miu, Godfrey Tan, Hari Balakrishnan, John Apostolopoulos * MIT Computer Science and Artificial.
A Study on Dynamic Load Balance for IEEE b Wireless LAN Proc. 8th International Conference on Advances in Communication & Control, COMCON 8, Rethymna,
Unwanted Link Layer Traffic in Large IEEE Wireless Network By Naga V K Akkineni.
Introducing Reliability and Load Balancing in Home Link of Mobile IPv6 based Networks Jahanzeb Faizan, Mohamed Khalil, and Hesham El-Rewini Parallel, Distributed,
Dynamic Load Balancing through Association Control of Mobile Users in WiFi Networks 2013 YU-ANTL Seminal November 9, 2013 Hyun dong Hwang Advanced Networking.
Handoff in IEEE Andrea G. Forte Sangho Shin Prof. Henning Schulzrinne.
Standard for a Convergent Digital Home Network for Heterogeneous Technologies Zhimeng Du 12/5/2013.
Deployment Guidelines for Highly Congested IEEE b/g Networks Andrea G. Forte and Henning Schulzrinne Columbia University.
Computer Networks Performance Metrics. Performance Metrics Outline Generic Performance Metrics Network performance Measures Components of Hop and End-to-End.
Voice Capacity analysis over Introducing VoIP and WLans IEEE based Wireless Local Area Networks (WLANs) are becoming popular While WLANs.
1 A Dynamical Redirection Approach to Enhancing Mobile IP with Fault Tolerance in Cellular Systems Jenn-Wei Lin, Jichiang Tsai, and Chin-Yu Huang IEEE.
Call Admission Control in IEEE Wireless Networks using QP-CAT Sangho Shin Henning Schulzrinne Department of Computer Science Columbia University.
Passive DAD Henning Schulzrinne Columbia University.
MOJO: A Distributed Physical Layer Anomaly Detection System for WLANs Richard D. Gopaul CSCI 388.
DISCERN: Cooperative Whitespace Scanning in Practical Environments Tarun Bansal, Bo Chen and Prasun Sinha Ohio State Univeristy.
November 4, 2003APOC 2003 Wuhan, China 1/14 Demand Based Bandwidth Assignment MAC Protocol for Wireless LANs Presented by Ruibiao Qiu Department of Computer.
Packet Dispersion in IEEE Wireless Networks Mingzhe Li, Mark Claypool and Bob Kinicki WPI Computer Science Department Worcester, MA 01609
Muhammad Mahmudul Islam Ronald Pose Carlo Kopp School of Computer Science & Software Engineering Monash University, Australia.
Doc.: IEEE /0648r0 Submission May 2014 Chinghwa Yu et. al., MediaTekSlide 1 Performance Observation of a Dense Campus Network Date:
Architectures and Algorithms for Future Wireless Local Area Networks  1 Chapter Architectures and Algorithms for Future Wireless Local Area.
1 ECE453 - Introduction to Computer Networks Lecture 1: Introduction.
August 27, 2003 Evaluation of WiNc Manager A Wireless Network Management Software from Cirond Technologies Inc. by Kassim Olawale Radio Science Laboratory.
Cooperation between stations in wireless networks Andrea G. Forte, Henning Schulzrinne Department of Computer Science, Columbia University Presented by:
Efficient Group Key Management in Wireless LANs Celia Li and Uyen Trang Nguyen Computer Science and Engineering York University.
October 17, 2007 Cooperation Between Stations in Wireless Networks Andrea G. Forte Henning Schulzrinne Department of Computer Science Columbia University.
Performance Limitations of ADSL Users: A Case Study Matti Siekkinen, University of Oslo Denis Collange, France Télécom R&D Guillaume Urvoy-Keller, Ernst.
1 Apricot2001 Effectiveness of VLAN Chan Wai Kok Faculty of Information Technology Salim Beg Faculty of Engineering.
A Comparison of RaDiO and CoDiO over IEEE WLANs May 25 th Jeonghun Noh Deepesh Jain A Comparison of RaDiO and CoDiO over IEEE WLANs.
1 Data Overhead Impact of Multipath Routing for Multicast in Wireless Mesh Networks Yi Zheng, Uyen Trang Nguyen and Hoang Lan Nguyen Department of Computer.
Passive Duplicate Address Detection (DAD) Sangho Shin Andrea Forte Henning Schulzrinne Columbia University.
Prof. Alfred J Bird, Ph.D., NBCT Office – Science 3rd floor – S Office Hours – Monday and Thursday.
Doc.: IEEE /492r00 Submission Orange Labs Date: Collaboration between 2.4/5 and 60 GHz May 2010 Slide 1 Authors:
Prof. Alfred J Bird, Ph.D., NBCT Office – McCormick 3rd floor 607 Office Hours – Monday 3:00 to 4:00 and.
Submission doc.: IEEE /0587r0 May 2016 Peter Khoury, Ruckus WirelessSlide 1 Dwell Time In Probe Request Presentation Date: Authors:
Andrea G. Forte Sangho Shin Henning Schulzrinne
Balancing Uplink and Downlink Delay of VoIP Traffic in WLANs
Fast MAC Layer Handoff in Networks
SUBMITTED BY DINEEJ A 28 S3 EC
Management in mobile wireless networks: The case for client-assistance
Managing the performance of multiple radio Multihop ESS Mesh Networks.
The Network Beacon Announcement scanning method
BlueScan: Boosting Wi-Fi Scanning Efficiency Using Bluetooth Radio
Roaming Interval Measurements
Using Dynamic PCF to improve the capacity of VoIP traffic in IEEE 802
Proposed Metrics for TGT and Call to Action
VoIP in IEEE Networks Henning Schulzrinne
Distributed Channel Assignment in Multi-Radio Mesh Networks
University of Arkansas at Little Rock
Collaboration between 2.4/5 and 60 GHz
Muhammad Niswar Graduate School of Information Science
Cooperation Between Stations in Wireless Networks
Wireless Performance Prediction – Rationale and Goals
Presentation transcript:

IEEE 802.11 in the Large: Observations at the IETF Meeting Andrea G. Forte Sangho Shin Henning Schulzrinne Department of Computer Science Columbia University

Introduction 65th IETF meeting Data collection Goals Dallas, TX March 19th ~ 24th Hilton Anatole hotel 1200 attendee Data collection 21st ~ 23rd for three days 25GB data, 80 millions frames Goals Identify unusual behaviors due to the highly congested environment 9/22/2018

Overview Introduction Related work Wireless network setup Measurement setup User behavior Load balancing Handoff behavior Overhead of multiple APs on a channel Conclusions 9/22/2018

Related Work Jardosh et al. Rodrig et al. Balachandran et al. Target 62nd IETF (’05) SIGCOMM ’04 SIGCOMM ’01 Attendee 1100 550 200 Clients detected 500 377 195 WLAN 802.11b Channels 1, 6, 11 (dynamic) 1, 8, 11 1, 4, 7, 11 Number of APs 38 5 4 Focused on New metric for channel congestion, ‘Link reliability’ Overhead of 802.11, the data transmission rate Load balancing: “No correlation b/w num of clients and traffic load” * * The number does not mean the total number of wireless clients in the IETF meeting, but the number of clients detected from the measurements. 9/22/2018

Wireless network in IETF meeting Many hotel 802.11b APs on channel 6 91 additional APs in 802.11a/b by IETF Cisco Aironet 1200 AP Channel 1 and 11 in 802.11b One subnet with one ESSID ‘ietf65 ’ No wireless security The largest wireless network measured so far Hotel APs 802.11b channel 6 IETF APs 802.11a/b ESSID : ietf65 9/22/2018

Measurement setup Four sniffers Channel assignment IBM T42 Think Pad Proxim ORiNOCO 11 a/b/g Sniffer software: Airopeek NX Channel assignment Three sniffers on channel 1, 6 and 11, respectively Fourth sniffer on all 8 channels in 802.11a 9/22/2018

Measurement setup Measurement place Room Chantilly The biggest room - 142’ x 80’, 600 persons Two biggest IETF sessions, and the plenary session in the evening Six IETF APs + Six Hotel APs in the room Session 2 & break Session 1 (AM) Lunch Session 1 (PM) Plenary 9:00 11:30 13:00 15:00 17:00 19:30 9/22/2018

The positions of the APs in the room Room Chantilly AP5 AP4 Seating area during IETF sessions and plenary Another seating area during plenary AP1 AP2 AP3 AP6 sniffers AP on channel 1 AP on channel 11 AP on channel 11 (estimated position) 142 feet screen Gate 80 feet Lobby 9/22/2018 AP9

Overview Introduction Related work Wireless network setup Measurement setup User behavior Load balancing Handoff behavior Overhead of multiple APs on a channel Conclusions 9/22/2018

User behavior Overall traffic in 802.11b 9/22/2018

User behavior Protocols 9/22/2018

User behavior Number of clients (all 8 channels) 9/22/2018

Overview Introduction Related work Wireless network setup Measurement setup User behavior Load balancing Handoff behavior Overhead of multiple APs on a channel Conclusions 9/22/2018

Load balancing Distribution of clients No load balancing feature was used Client distribution is decided by the relative proximity from the APs Ch 6 (ceiling) > Ch 1 > Ch 11 AP5 AP4 Seating area during IETF sessions and plenary Another seating area during plenary AP1 AP2 AP3 AP6 sniffers 9/22/2018

Load balancing Distribution of clients No load balancing feature was used Client distribution is decided by the relative proximity from the APs Ch 6 (ceiling) > Ch 1 > Ch 11 AP1 > AP2 > AP3 Number of clients in channel 1 9/22/2018

Load balancing Throughput Average throughput in 802.11a/b Average throughput in channel 1 9/22/2018

Load balancing Throughput per client Average throughput per client in 802.11a/b Average throughput per client in channel 1 9/22/2018

Load balancing How? Study of Balanchandran et al. [1] Number of clients? – Simple, but not accurate Throughput ? - Accurate, but too complex Study of Balanchandran et al. [1] “No correlation between the number of clients and traffic load” “Throughput per client needs to be considered” In the large scale wireless network like IETF meeting? [1] A. Balanchandran et al. “Characterizing user behavior and network performance in a public wireless LAN”, SIGMETRICS ‘02 9/22/2018

Capacity in the channel Load balancing Number of clients vs. throughput in Ch. 6 Capacity in the channel 9/22/2018

Capacity in the channel Load balancing Number of clients vs. Throughput Capacity in the channel Clear correlation between the number of clients and throughput The number of clients can be used for load balancing with low complexity of implementation, in large scale wireless networks 9/22/2018

Effect of screen Small number of clients on AP3 Effect of a screen? Number of clients in channel 1 9/22/2018

Effect of screen An experiment 30 ~ 100 feet 9/22/2018

Overview Introduction Related work Wireless network setup Measurement setup User behavior Load balancing Handoff behavior Overhead of multiple APs on a channel Conclusions 9/22/2018

The number of handoff per hour Handoff behavior Handoff is triggered generally, by low signal strength in congested channel, by frame loss Effect of layer 2 handoff Increase of traffic Disruption of network (0.5 ~ 1.5 sec) The number of handoff per hour in each IETF session 9/22/2018

Handoff behavior Handoffs between channels Handoff to the same channels : 72% Handoff to the same AP : 55% 9/22/2018

Handoff behavior Distribution of session time: time between handoffs Too often handoff Disruption of network 0.5 ~1.5 sec per handoff Increase of traffic due to handoff related frames – probe request and response 10.4% of total 9/22/2018

(based on the number of handoffs during the whole day) Handoff behavior Handoffs per vendor Distribution of vendors Distribution of handoffs per vendor (based on the number of handoffs during the whole day) 9/22/2018

Handoff behavior Handoffs per vendor (session time) 9/22/2018 Cisco Apple 9/22/2018

Overview Introduction Related work Wireless network setup Measurement setup User behavior Load balancing Handoff behavior Overhead of multiple APs on a channel Conclusions 9/22/2018

Overhead of having multiple APs Overhead from replicated multicast and broadcast frames All broadcast and multicast frames are replicated by all APs  Increase traffic Router A channel 9/22/2018

Overhead of having multiple APs Overhead from replicated multicast and broadcast frames All broadcast and multicast frames are replicated by all APs DHCP request (broadcast) frames are replicated and sent back to each channel Router A channel 9/22/2018

Overhead of having multiple APs Overhead from replicated multicast and broadcast frames All broadcast and multicast frames are replicated by all APs. DHCP request (broadcast) frames are replicated and sent back to each channel. Multicast and broadcast frames : 10% 9/22/2018

Overhead of having multiple APs Co-channel interference 14 APs around the room Chantilly in channel 6 Co-channel interference and too many clients are responsible for the very low throughput of channel 6. 9/22/2018

Conclusions Uneven distribution of clients and throughput among channels and APs Correlation between the throughput and the number of clients The number of clients can be used for load balancing with low complexity in highly congested wireless networks Too many inefficient handoffs Disrupt network and increase traffic Need better handoff algorithms Having multiple APs on a channel Increases the traffic due to replication of multicast and broadcast frames Decrease throughput due to co-channel interference 9/22/2018

Most Recent Results 9/22/2018

Ongoing Work Apple vs. Intel (Thank you Norbert!) Apple When to scan? Who to associate to? Intel (newest chipset) Roaming aggressiveness Too many handoffs? 9/22/2018

Thank you! Questions ? 9/22/2018

Backup slides 9/22/2018

Handoff behavior Total number of handoffs Total number of handoffs Number of handoffs per hour 9/22/2018