ENSC 835: COMMUNICATION NETWORKS CMPT 885: SPECIAL TOPICS: COMMUNICATION NETWORKS FINAL PROJECT PRESENTATION Spring 2008

Deployment of Mobile and Fixed Video Conferencing over an Existing IP Infrastructure

Victor Gusev http://www.sfu.ca/~vga9/ensc835.htm vga9@sfu.ca
1

Roadmap
Introduction Related Work Problem Description Models Simulation Study Results and Discussions Conclusions and Future Improvements References

Motivation and Overview

Use already existing IP infrastructure Minimal deployment cost Video conferencing deployment not yet analysed for WiFi in OPNET Commercial interest (simulate before investing time and money into hardware/software setup) Research interest (optimization, capacity)

Introduction (continues)
Typical scenario: plant/warehouse floor (wireless), and office floor (desktops)
Plant maintenance (chemical, power)
Personnel with PDAs reporting to office

Production environment
Warehouse, equipment maintenance, report to office

Office with 2 floors

Mix of wireless clients and desktops

Existing standards: 802.11a/b/g, 802.11e, 802.11n 802.11a/b/g, 802.11e, 802.11n This project concentrates on the most popular deployed hardware: 802.11b and 802.11g, installed on one floor 802.11b 802.11g, and wired workstations on the second floor

Related Work
X. Cao, G. Bai, and C. Williamson, Media streaming performance in a portable wireless classroom network
   

Ad-hoc 802.11b testbed 400 kbps video 128 kbps audio Demonstrates that up to 8 clients can be supported

[3] X. Cao, G. Bai, and C. Williamson, Media streaming performance in a portable wireless classroom network, in Proceedings of the IASTED European Workshop on Internet Multimedia Systems and Applications (EuroIMSA), Feb. 2005, pp. 246252.
5

Related Work
Salah, K., and Alkhoraidly, A., An Opnet-based OpnetSimulation Approach for Deploying VoIP
VoIP parameters Chooses G.711

[2] K. Salah and A. Alkhoraidly, An OPNET-based Simulation Approach for Deploying VoIP, International Journal of Network Management, vol. 16, no. 3, May 2006, pp. 149-183.
6

Related Work
[1] K. Salah, Analytic approach for deploying desktop videoconferencing, IEE Proceedings Communications, Communications, vol. 153, no. 3, June 2006, pp. 434 153, 2006, 434 444. 444.

Problem Description: Technical Details

Want to extend existing desktop conferencing simulation [1] to include wireless links Previously simulated capacity using 100BaseT must be larger than our WLAN's capacity WLAN becomes a bottleneck Want to examine throughput, drop rate, delay, number of supported clients Compare 802.11b, 802.11g in a typical environment with a mix of desktop and mobile clients

Model: Typical Office Environment

Model: Typical Office with Wireless Clients

Floor 1: Plant/warehouse floor with equipment Floor 2: Office, dispatch Examples:
UPS, CanadaPost Manufacturing, chemical, power and other plants

Enable interactive voice or video call to report to the main office or between personnel

10

Choice of parameters: H.323

Widely implemented by voice and videoconferencing equipment manufacturers Used within various Internet real-time applications (i.e. realGnuGK, NetMeeting and X-Meeting) XDeployed worldwide by service providers and enterprises for voice and video services over IP networks. H.323s strength: H.323s strength: multimedia communication functionality designed specifically for IP networks.

11

Choice of parameters: G.711 G.711

Popular audio encoder scheme Payload of VoIP: 160 OPNET uses: 32 voice samples, 8 bits each Set Voice frames per packet: 5

12

Scenarios
Scenario 1: 802.11b (OPNET), increasing # calls model 802.11b Scenario 2: 802.11g (OPNET), increasing # calls model 802.11g Scenario 3: 802.11b verification (IT Guru) 802.11b All 3 scenarios have 2 floors (wireless and wired)

13

Scenario 1: 802.11b Global pps

IP background traffic starts at 40s, VConf at 70s Tx/Rx pps mismatch at 1m 38s

14

Scenario 1: 802.11b
Wireless LANs Dropped data (bits/s), Delay (s) and Media access delay (s)

15

Scenario 1: 802.11b 802.11b

Wireless LANs Queue Size (packets)

16

Scenario1 Scenario1: Discussion

# Video conferencing calls supported: Traffic Rx/Tx Mismatch at 1m 38s, good at 1m 36s. 38s, 36s. Started with 2 video packets at 70s 70s Added 2 new calls every 2 seconds Hence: 2calls + 2calls*((1m * 60s/m + 36s 70s)/2s) = 28 calls*((1 60s/m 36s 70s)/2 videoconferencing calls

17

Scenario 1: 802.11b
Wireless LANs Load, Throughput and Data Dropped (bits/s)

18

Scenario1: Discussion
802.11b 802.11b reaches the typical throughput of 6 Mbps (max theoretical rate is 11Mbps) 11Mbps) Due to the CSMA/CA protocol overhead, the actual throughput is 4.3-5.9 Mbps (TCP) and 7.1 (UDP) Mbps.

19

Scenario 2: 802.11g
Videoconferencing traffic sent/received (packets/sec)

20

Scenario 2: 802.11g 802.11g

Videoconferencing traffic sent/received (bytes/sec)

21

Scenario 2: 802.11g
Wireless LANs Data Dropped (bits/s), Queue size (packets)

22

Scenario 2: Discussion
# Video conferencing calls supported: Traffic Rx/Tx Mismatch at 2m 48s, good at 2m 46s. 48s, 46s. Started with 2 video packets at 70s 70s Added 2 new calls every 2 seconds Hence: 2calls + 2calls*((2m * 60s/m + 46s 70s)/2s) = 98 calls*((2 60s/m 46s 70s)/2 videoconferencing calls

23

Scenario 2: 802.11g
Wireless LANs Load, Throughput, Data Dropped (bits/s)

24

Scenario 2: Discussion
The modulation scheme is OFDM with data rates of up to 54 Mbit/s Higher speeds than 802.11b, higher link capacity 802.11b, 802.11g 802.11g reaches its typical 23 Mbps (note: 54 Mbps is a theoretical limit

25

Scenario 3: Further Analysis

# Videoconferencing calls:
First floor 802.11b: 28 First floor 802.11g: 98

Are these the actual number of calls 802.11 can support? Careful! Each of these calls can be made between the clients of the same floor! Verify

26

Scenario 3: 802.11b 802.11b Throughput

Verify the validity of Scenario 1 Used IT Guru for a quicker set of simulations for verification purposes One server supports all services Only one desktop - fixed Manually add wireless clients until the packets are dropped (in contrast to automatic # calls increase) Compare results

27

Scenario 3: 802.11b revisited: 802.11b IT Guru

28

Scenario 3: 9 wireless, 1 desktop

Note the mismatch between sent and received VC traffic

29

Scenario 3: 8 wireless, 1 desktop

Videoconferencing traffic sent = received

30

Results
Limit for 802.11b is 9 video calls (all wireless)! 802.11b Agrees with X. Caos results (Media streaming performance in a portable wireless classroom network paper) Similar check can be done for 802.11g 802.11g

31

Conclusions
This study analyzed a deployment of videoconferencing clients over an existing IP infrastructure Both, inter and intra floor communication was taking into account 802. 802.11 reached its bandwidth limits before the IP network as expected Longer simulations would result in IP packet drop rate >0 # Videoconferencing (inter and intra floor) calls:
First floor 802.11b: 28 802.11b: First floor 802.11g: 98 802.11g:
32

Future Work
Compare with 802.11e, 802.11n 802.11e, 802.11n Add wireless to all floors Add background wireless traffic Add weighted call destinations to limit calls within the same floor Multiple access points spaced far apart Security

33

References
[1] K. Salah, Analytic approach for deploying desktop videoconferencing, IEE Proceedings Communications, vol. 153, no. 3, June 2006, pp. 434444. Communications, 153, 2006, 434444. [2] K. Salah and A. Alkhoraidly, An Opnet-based Simulation Approach for OpnetDeploying VoIP, International Journal of Network Management, vol. 16, no. 3, May Management, 16, 2006, 2006, pp. 149183. 149183. [3] X. Cao, G. Bai, and C. Williamson, Media streaming performance in a portable wireless classroom network, in Proceedings of the IASTED European Workshop on Internet Multimedia Systems and Applications (EuroIMSA), Feb. 2005, pp. 246 (EuroIMSA), 2005, 246 252. 252. [4] Y. E. Liu, J. Wang, M. Kwok, J. Diamond and M. Toulouse, Capability of IEEE Toulouse, 802.11g 802.11g Networks in Supporting Multi-player Online Games, Consumer MultiCommunications and Networking Conference, Jan. 2006, pp. 11931198. Conference, 2006, 11931198. [5] A. Wijesinha, Y. Song, M. Krishnan, V. Mathur, J. Ahn, and V. Shyamasundar, Throughput measurement for UDP traffic in an IEEE 802.11g WLAN, in 802.11g Proceedings of the 6th International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing, 2005 and First ACIS International Workshop on Self-Assembling Wireless Networks, May 2005, pp. SelfNetworks, 2005, 220225. 220225.
34