Quality of service is a prerequisite for real-time applications like voice over IP. The residential gateway plays as the first gate keeper at the residential endpoints to make sure the priority datagram can be given precedence in transmission and receiving over the other non-priority datagram. Usually the limited upstream and downstream bandwidth are given by most of the present internet service providers and the common incremental requirement of the network bandwidth for the residential users, it’s getting more and more critical to ensure the quality of service for the priority datagram like voice over IP.
Three algorithms (E-Meter, RPDTB and I-Police) were proposed and implemented to address three key portions by measuring the egress available bandwidth, shaping the egress traffic and lowering the ingress TCP traffic. The objective of the implementation of SQPV is to prove and show the output result of E-Meter, RPDTB and I-Police in the test setup with the real network deployment emulated. The results demonstrated the egress bandwidth measurement reached a good level of accuracy, the outgoing non-priority traffic was well shaped and the incoming TCP traffic was also congestion avoided. The good service of quality for 5 SIP/RTP sessions were able to be assured concurrently with the other non-priority TCP and UDP sessions.
The proposed algorithms for E-Meter, RPDTB and I-Police can be applied and reused for the enhanced quality of service provision system for various kinds of real-time streaming applications. SIP signaling protocol along with the 64 Kbps G.711 CODEC were chosen and implemented in SQPV to simply demonstrate the result of E-Meter, RPDTB and I-Police. As long as some configuration on the bandwidth requirement of the additional real-time streaming applications and modification on the Traffic Parser to support and detect the additional required signaling protocol of the real-time streaming applications, the different quality of service provision system can be provided.
Besides SIP, Skype [21] is the most amazing example of this new phenomenon. It recently reached over 170 millions of users and accounts for more 4.4% of total VoIP traffic [18]. It’s worth making a proposal to support Skype here as an example.
Skype relies on a P2P infrastructure to exchange signaling information in a distributed fashion which can be making the system highly scalable and robust [17]. Except for the user’s authentication which is performed under a classical client and server architecture by means of public key mechanisms, all further signaling is performed in P2P network.
So the Skype user’s information are entirely decentralized and distributed among nodes.
This allows the service to scale very easily to large sizes and avoid a costly centralized infrastructure. However, Skype uses a proprietary solution which is difficult to reverse engineer due to extensive use of both cryptography and obfuscation techniques [19]. It makes the traffic parser difficult to interpret the Skype signaling protocols and figure out how to add or remove the Skype session to make the run-time bandwidth allocation and admission control possible. Skype is able to select different CODEC according to the unknown algorithm. According to the nominal characteristics of Skype CODECS showed in Figure 58 [20], the bit rate of the different CODECS is at least 8 Kbps (G.729) and 80 Kbps (iPCM-wb) at most.
Figure 58 : Nominal Characteristics of Skype CODECS
The QoS provisioning mechanisms (E-Meter, RPDTB and I-Police) are applicable to Skype calls as well since the egress bandwidth measurement, egress traffic shaping and ingress congestion avoidance are all required regardless of different kinds of real-time streaming applications. Limitation is the admission control and bandwidth allocations at running time are not possible due to lack of transparency of Skype signaling protocols.
The alternative approach is to have the preconfigured amount of bandwidth for priority traffic (like Skype calls) and modify the congestion detection scheme of I-Police accordingly.
6 Bibliography
[1] L. Peterson, B. Davie, Computer Networks, A Systems Approach, Edition 3, Morgan Kaufmann Publishers, San Francisco, 2003
[2] S. Blake, D. Black, M. Carlson, E. Davies, Z. Wang, W. Weiss, An Architecture for Differentiated Services, RFC 2475, December 1998
[3] M. Handley, V. Jacobson, C. Perkins, SDP: Session Description Protocol, RFC 4566, July 2006
[4] A. Johnston, SIP: Understanding the Session Initiation Protocol, 2nd edition, Artech House, 2004
[5] H. Schulzrinne, S. Casner, R. Frederick, V. Jacobson, RFC 3550, RTP: A Transport Protocol for Real-Time Applications, July 2003
[6] Official PESQ website, http://www.pesq.org/, 2001, 2002, 2003, 2004, 2005, 2006, 2007 OPTICOM GmbH, Erlangen, Germany
[7] A. Chadda, Quality Of Service Testing Methodology, Dec, 2004
[8] H. Schulzrinne, G. Camarillo, A. Johnston, J. Peterson, R. Sparks, M. Handley, E.
Schooler, RFC 3261 - SIP: Session Initiation Protocol, June 2002
[9] National Institute of Standards and Technology Group, NIST Net Home page, http://snad.ncsl.nist.gov/nistnet/, Jan, 1998
[10] SIPp, http://sipp.sourceforge.net/index.html, 2004 [11] The Board of Trustees, University of Illinois,
http://www.dast.nlanr.net/projects/Iperf/, 1999-2006 [12] P. Brodsky, TeleGeography Research,
http://www.apritel.org/fotos/editor2/Paul_Brodsky.pdf
[13] W.S. Hwang and P.C. Tseng, A QoS-aware Residential Gateway with Bandwidth Management, IEEE Transactions on Consumer Electronics, Vol. 51, No. 3, page 840-848, August, 2005,
http://credit.csie.ncku.edu.tw/opensource/93report/CMMI_RR/38.pdf
[14] S. Floyd, Lawrence Berkeley Laboratory, Note on CBQ and guaranteed service, July, 1995
[15] D. Bansal, J. Bao, and W. Lee, Motorola Labs, QoS-Enabled Residential Gateway Architecture, IEEE Communications Magazine, Vol. 41, issue 4, page 83-89, April 2003
[16] R. Yavatkar, Intel, D. Hoffman, Teledesic, Y. Bernet, Microsoft, F. Baker, Cisco, M.
Speer, Sun Microsystems, RFC 2814 - SBM (Subnet Bandwidth Manager): A Protocol for RSVP-based Admission Control over IEEE 802-style networks, May 2000
[17] D. Rossi, M. Mellia, M. Meo, ENST Telecom Paris, France, [email protected],
Politecnico di Torino, Italy, Following Skype Signaling Footsteps, page 248-253, IT-NEWS, 2008 IEEE
[18] International carriers’ traffic grows despite Skype popularity, Tele-Geography Report and Database, available on line http://www.telegeography.com/, Dec. 2006 [19] P. Biondi, F. Desclaux, Silver Needle in the Skype, Black Hat Europe 06,
Amsterdam, the Netherlands, Mar 2006
[20] D. Bonfiglio, M. Mellia, M. Meo, Nicol`o Ritacca, Politecnico di Torino – Dipartimento di Elettronica, D. Rossi, ENST ParisTech – INFRES Department, Tracking Down Skype Traffic, page 843-851, 2008 IEEE
[21] Skype web site, http://www.skype.com