Flow-Level Packet Loss Analysis of a Markovian Bottleneck Buffer

Buffer overflow in intermediate network routers is the prime cause of packet loss in wired communication networks. Packet loss is usually quantified by the packet loss ratio , the fraction of packets that are lost in a buffer. While this measure captures part of the loss performance of the buffer, we show that it is insufficient to quantify the effect of loss on user-perceived quality of service for multimedia streaming applications. In this contribution, we refine the quantification of loss in two ways. First, we focus on loss of a single flow, rather than loss in a buffer. Second, we focus on the different moments of the time and number of accepted packets between losses, rather than just the mean number of accepted packets between losses (which directly relates to the packet loss ratio). The network node is modelled as a Markov-modulated M/M/1/N-type queueing system which is sufficiently versatile to capture the arrival correlation while keeping the analysis tractable. We illustrate our approach by some numerical examples.

[1]  Pascal Frossard.  FEC performance in multimedia streaming , 2001, IEEE Communications Letters.

[2]  Israel Cidon,et al.  The ballot theorem strikes again: Packet loss process distribution , 2000, IEEE Trans. Inf. Theory.

[3]  Andrzej Chydzinski,et al.  On the loss process in a batch arrival queue , 2009 .

[4]  H. Michiel,et al.  Teletraffic engineering in a broad-band era , 1997, Proc. IEEE.

[5]  Dieter Fiems,et al.  On higher-order packet loss characteristics , 2005 .

[6]  Predrag R. Jelenkovic,et al.  Subexponential loss rates in a GI/GI/1 queue with applications , 1999, Queueing Syst. Theory Appl..

[7]  Tatsuya Suda,et al.  Cell loss and output process analyses of a finite-buffer discrete-time ATM queueing system with correlated arrivals , 1995, IEEE Trans. Commun..

[8]  Matti Latva-aho,et al.  Throughput Optimization in Wireless Networks Under Stability and Packet Loss Constraints , 2014, IEEE Transactions on Mobile Computing.

[9]  Gunnar Karlsson,et al.  On the effects of the packet size distribution on the packet loss process , 2006, Telecommun. Syst..

[10]  Tetsuya Takine,et al.  Analysis of packet loss in transport layer over ATM networks , 1997 .

[11]  Alain Jean-Marie,et al.  Loss probabilities for messages with redundant packets feeding a finite buffer , 1998, IEEE J. Sel. Areas Commun..

[12]  Israel Cidon,et al.  Analysis of packet loss processes in high-speed networks , 1993, IEEE Trans. Inf. Theory.

[13]  Fumio Ishizaki,et al.  Loss Probability for a Finite Buffer Multiplexer with the M/G/∞ Input Process , 2005, Telecommunications Systems.

[14]  Donald F. Towsley,et al.  Inferring network characteristics via moment-based estimators , 2001, Proceedings IEEE INFOCOM 2001. Conference on Computer Communications. Twentieth Annual Joint Conference of the IEEE Computer and Communications Society (Cat. No.01CH37213).

[15]  Dieter Fiems,et al.  An analytic study of a scalable video buffer , 2009, Telecommun. Syst..

[16]  Dieter Fiems,et al.  Packet loss characteristics for M/G/1/N queueing systems , 2009, Ann. Oper. Res..

[17]  Che Soong Kim,et al.  The MAP/PH/1/N queue with flows of customers as a model for traffic control in telecommunication networks , 2009, Perform. Evaluation.

[18]  Vaidyanathan Ramaswami,et al.  Introduction to Matrix Analytic Methods in Stochastic Modeling , 1999, ASA-SIAM Series on Statistics and Applied Mathematics.

[19]  San-qi Li,et al.  Spectral analysis of packet loss rate at a statistical multiplexer for multimedia services , 1994, TNET.

[20]  Ness B. Shroff,et al.  Loss probability calculations and asymptotic analysis for finite buffer multiplexers , 2001, TNET.

[21]  Eitan Altman,et al.  On loss probabilities in presence of redundant packets with random drop , 2003, Perform. Evaluation.

[22]  Herwig Bruneel,et al.  An efficient solution technique for discrete-time queues fed by heterogeneous traffic. , 1997 .

[23]  Gunnar Karlsson,et al.  On the effects of the packet size distribution on FEC performance , 2006, Comput. Networks.

[24]  Gunnar Karlsson,et al.  User-oriented QoS in packet video delivery , 1998, IEEE Netw..

[25]  Tetsuya Takine,et al.  Analysis of the loss probability in the M/G/1+G queue , 2015, Queueing Syst. Theory Appl..

[26]  Henning Schulzrinne,et al.  Loss correlation for queues with bursty input streams , 1992, [Conference Record] SUPERCOMM/ICC '92 Discovering a New World of Communications.

[27]  Eitan Altman,et al.  On Loss Probabilities in Presence of Redundant Packets and Several Traffic Sources , 1999, Perform. Evaluation.

[28]  Ofer Hadar,et al.  Quality measurements for compressed video transmitted over a lossy packet network , 2004 .

[29]  Annie Gravey,et al.  Simultaneity in Discrete-Time Single Server Queues with Bernoulli Inputs , 1992, Perform. Evaluation.

[30]  Tetsuya Takine,et al.  The M/D/1+D queue has the minimum loss probability among M/G/1+G queues , 2015, Oper. Res. Lett..

[31]  M. Pihlsgård.  Loss Rate Asymptotics in a GI/G/1 Queue with Finite Buffer , 2005 .