Download presentation
Presentation is loading. Please wait.
Published byLogan Dixon Modified over 9 years ago
1
Network Architecture Laboratory Experiments and Analysis of Voice over Mobile IP Soonuk Seol and Myungchul Kim {suseol,mckim}@icu.ac.kr
2
2 Motivation o Voice over IP – Internet telephony is one of the most promising services – low cost, efficient bandwidth utilization, integration with data traffic – Support only best effort service, more obstacles to deteriorate voice quality, e.g., delay, delay jitter, packet loss, etc. – There are two competing approaches for VoIP ITU’s H.323 [1,2], IETF’s SIP [3] o Mobility demand – VoIP needs to support most functionalities that the current PSTN does, especially mobility support. o All-IP trends – Recently, it is believed all mobility-related functionality should be handled at the IP (network) layer [10,11,12,13].
3
3 Related Work o Extensions to H.323 for mobility [8,9] : – Additional messages and functionalities to H.323 system – Requires applications to perform mobility management o Mobility support to SIP – Moh et al. [5] Address several major issues for supporting mobility on SIP – Wedlund and Schulzrinne [6] An application level approach for real-time mobile communication. Mobility support is limited to SIP-aware applications and SIP-aware correspondent hosts. Networks should support DHCP to assign IP addresses. Overhead with mobile IP – A waste of resources to keep duplicated information about the hosts current address. (both in SIP servers and Home agents) o In our experiments – Need a homogeneous mobility solution regardless of wireless interfaces and applications. – Based on Mobile IP [4] for mobility support
4
4 What we have achieved o Examine the feasibility of Voice over Mobile IP for Internet telephony – Investigate various factors that affect delay, packet loss, and load on the network – Experiment with encapsulation and decapsulation delay time and interarrival time in many aspects, comparing with normal IP. o Find the desirable number of frames per packet in Mobile IP as a function of packet transmission delay and bandwidth utilization.
5
5 Backgrounds o Mobile IP – Allows a mobile node to communicate with other nodes transparently in spite of address change due to its mobility – Triangular routing problem which increases delays – Route optimization solves the problem by using binding updates.
6
6 Backgrounds o Session Initiation Protocol (SIP) – SIP allows two or more participants to establish a session consisting of multiple media streams. – In SIP, callers and callees are identified by SIP address. – When making a SIP call, a caller first locates the appropriate server and then sends a SIP request. – SIP server can act in two different modes Proxy server – requests to the next hop or user-agent within an IP cloud Redirect server – informs their clients of the address of the requested server – allow for the client to contact that server directly – In our experiment, we make calls through peer-to-peer communications without any server.
7
7 Testbed Configuration Mobile IP: Dynamics, http://www.cs.hut.fi/Research/Dynamics/ SIP: Linphone, http://www.linphone.org, GSM codec is used o Analysis with TCPDUMP (for capturing packets) and Ping IEEE 802.11 PC Card 11 Mbit/s
8
8 RTP packet format o Length of a packet : 87 bytes – IP header : 20 bytes, IP option : 14 bytes – UDP header : 8 bytes – RTP message : 45 bytes ( RTP header : 8bytes, Voice data: 33 bytes)
9
9 Encapsulation delay o Encapsulation and decapsulation delay in Mobile IP: ~ 1ms – Measure the encapsulation and decapsulation delay by configuring the routing path between MH and CH in mobile IP to be identical to that of not using mobile IP. FA1 HA CH FA1 HA CH
10
10 Interarrival time in one-way calls over conventional IP – Sending interval : 20 ms – Interarrival time : 19.95 ~ 20.05 ms with 99% confidence – Standard deviation : 0.5 ms – Number of samples : 700 packets (14 seconds)
11
11 Interarrival time in one-way calls over Mobile IP – Sending interval : 20 ms – Interarrival time : 19.91 ~ 20.09 ms with 99% confidence – Standard deviation : 0.89 ms – Number of samples : 700 packets (14 seconds)
12
12 Interarrival time in voice conversation(1) o Bi-directional voice conversation for 60 seconds. o Average: 20ms, overall within 42ms for every case: (a) IP (b) Mobile IP without handoffs (c) Mobile IP with 5 times of handoffs
13
13 Interarrival time in voice conversation(2) o Overall packets arrive within 42 ms. (be made up with buffers) o No many differences during the handoff time. The reason is that a mobile node – can receive packets from the old foreign agent. – gets a care-of address from the FA not from the DHCP server. – Cells are overlapped enough. FA1 HA CH FA2 h->FA1 c->m h->FA2 c->m MH FA1 HA CH FA2 MH
14
14 Interarrival time under background traffic – five extra sessions for MN with different hosts, totally 6300 packets (~2min) for each call – The longest : Normal IP = 25 ms, Mobile IP = 30 ms – 98% of packets = 18 ~ 22 ms – Traditional packet loss
15
15 Bandwidth of GSM codec o SIP application, Linphone with GSM – One frame of 33 bytes in a single packet – Size of packet headers : 54 bytes – Frame duration : 20 ms Table 1. Bandwidth of GSM codec (frame size = 33 bytes, frame duration = 20ms) frames/pktpkts/secpayloadbits/secpkt sizebits/sec% optimallatency bytes ms 150.0033132008734800264 %20 225.00661320012024000182 %40 316.67991320015320400155 %60 412.501321320018618600141 %80 510.001651320021917520133 %100 68.331981320025216800127 %120 77.142311320028516286123 %140 86.252641320031815900120 %160 95.562971320035115600118 %180 105.003301320039415360116 %200 114.553631320041715164115 %220 124.173961320045015000114 %240
16
16 Total Data Size for Different frames/pkt o One-way voice data – Totally, 297 Kbytes for 180 sec (one frame : 33 bytes ) – IP & UDP headers: add 54 bytes – Encapsulation (from HA to FA): adds 20 bytes
17
17 The Desirable Number of Frames o Mobile IP Network – need to save the bandwidth (esp., wireless network) o End to end delays – Smaller than 150 ms : not perceived (our goal) – Between 150 and 400 ms : acceptable but not ideal – If f/p=3: about 60ms’ latency to aggregate three frames. The rest 90ms (150-60) are remained for packet transfer.
18
18 Conclusion and Future work o Feasibility of Mobile IP-based SIP – Mobile IP’s encapsulation and decapsulation delay is short enough for interactive audio applications. – Interarrival time does not vary much. o Desirable number of frames per packet – Sends three frames per packet to reduce loads on the campus- sized network. o Future work – Simulate SIP over Mobile IP for large scaled networks – study various kinds of codecs in the same context and in terms of the number of hops. – delay-aware and/or load-aware scheme for Internet Telephony
19
19 References [1] Gary A. Thom, “H.323: the Multimedia Communications Standard for Local Area Networks,” IEEE Communications Magazine, December 1996. [2] ITU-T Rec. H.323v2, “Packet Based Multimedia Communications Systems,” March 1997. [3] M. Handley et al., “SIP: Session Initiation Protocol,” IETF RFC 2543, March 1999. [4] C. Perkins, “IP Mobility Support,” RFC 2002, IETF, October 1996. [5] Melody Moh, Gregorie Berquin, and Yanjun Chen, “Mobile IP Telephony: Mobility Support of SIP,” Eighth International Conference on Computer Communications and Networks, 1999. [6] Elin Wedlund and Henning Schulzrinne, “Mobility Support using SIP,” Proceedings of the second ACM International Workshop on Wireless Mobile Multimedia (WoWMoM), 1999. [7] X. Zhao, C. Castelluccia, and M. Baker, “Flexible Network Support for Mobility,” in Proceedings of Mobicom, October 1998. [8] ITU-T Draft Recommendation H.MMS.1, “Mobility for H.323 Multimedia Systems,” March 2001. [9] Wanjiun Liao, “Mobile Internet Telephony: Mobile Extensions to H.323,” INFOCOM ’99. Eighteenth Annual Joint Conference of the IEEE Computer and Communications Societies. Proceedings. IEEE, June 1999. [10] Ramachandran Ramjee, Thomas F. La Porta, Luca Salagrelli, Sandra Thuel, and Kannan Varadhan, “IP-based Access Network Infrastructure for Next-Generation Wireless Data Networks,” IEEE Personal Communications, August 2000. [11] Shingo Ohmori, Yasushi Yamao, and Nobuo Nakajima, “The Future Generations of Mobile Communications Based on Broadband Access Technologies,” IEEE Communications Magazine, December 2000.
20
20 References (cont.) [12] Ramón Cáceres and Venkata N. Padmanabhan, “Fast and Scalable Wireless Handoffs in Supports of Mobile Internet Audio,” Mobile Networks and Applications 3, December 1998. [13] Mihailovic, A., Shabeer, M., and Aghvami, A.H., “Multicast for Mobility Protocol (MMP) for Emerging Internet Networks,” The 11th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), 2000. [14] H. Schulzrinne and J. Rosenberg, “A Comparison of SIP and H.323 for Internet Telephony,” http://www.cs.columbia.edu/~hgs/sip/papers.html. [15] James F. Kurose and Keith W. Ross, “Computer Networking – A Top-Down Approach Featuring the Internet”, Addison Wesley Longman, 2001. [16] Charles Perkins and David B. Johnson, “Route Optimization in Mobile IP,” draft-ietf-mobileip- optim-11.txt (Work in progress), September 2001. [17] David B. Johnson and Charles Perkins, “Mobility Support in IPv6,” draft-ietf-mobileip-ipv6-13.txt (Work in progress), July 2001. [18] Dynamics – HUT Mobile IP, available at http://www.cs.hut.fi/Research/Dynamics/index.html. [19] Linphone – a SIP application, available at http://simon.morlat.free.fr/english/linphone.html.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.