Packet Loss Burstiness: Measurements and Implications for Distributed Applications

Many modern massively distributed systems deploy thousands of nodes to cooperate on a computation task. Network congestions occur in these systems. Most applications rely on congestion control protocols such as TCP to protect the systems from congestion collapse. Most TCP congestion control algorithms use packet loss as signal to detect congestion. In this paper, we study the packet loss process in sub-round-trip-time (sub-RTT) timescale and its impact on the loss-based congestion control algorithms. Our study suggests that the packet loss in sub-RTT timescale is very bursty. This burstiness leads to two effects. First, the sub-RTT burstiness in packet loss process leads to complicated interactions between different loss-based algorithms. Second, the sub-RTT burstiness in packet loss process makes the latency of data transfers under TCP hard to predict. Our results suggest that the design of a distributed system has to seriously consider the nature of packet loss process and carefully select the congestion control algorithms best suited for the distributed computation environments.

[1]  Mark Handley,et al.  Equation-based congestion control for unicast applications , 2000, SIGCOMM 2000.

[2]  Sally Floyd,et al.  The NewReno Modification to TCP's Fast Recovery Algorithm , 2004, RFC.

[3]  Injong Rhee,et al.  Limitations of Equation-Based Congestion Control , 2005, IEEE/ACM Transactions on Networking.

[4]  Sally Floyd,et al.  Quick-Start for TCP and IP , 2007, RFC.

[5]  Luigi Rizzo,et al.  Dummynet: a simple approach to the evaluation of network protocols , 1997, CCRV.

[6]  Ao Tang,et al.  Heterogeneous Congestion Control: Efficiency, Fairness and Design , 2006, Proceedings of the 2006 IEEE International Conference on Network Protocols.

[7]  Michael S. Borella,et al.  Internet packet loss: measurement and implications for end-to-end QoS , 1998, Proceedings of the 1998 ICPP Workshop on Architectural and OS Support for Multimedia Applications Flexible Communication Systems. Wireless Networks and Mobile Computing (Cat. No.98EX206).

[8]  Nick McKeown,et al.  Why flow-completion time is the right metric for congestion control , 2006, CCRV.

[9]  Amit Aggarwal,et al.  Understanding the performance of TCP pacing , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[10]  Craig Partridge,et al.  A Simulation Study of Paced TCP , 2000 .

[11]  Steven H. Low,et al.  Fairness Convergence of Loss-based TCP , 2006 .

[12]  Cheng Jin,et al.  FAST TCP: Motivation, Architecture, Algorithms, Performance , 2006, IEEE/ACM Transactions on Networking.

[13]  Mark Handley,et al.  Equation-based congestion control for unicast applications , 2000, SIGCOMM.

[14]  QUTdN QeO,et al.  Random early detection gateways for congestion avoidance , 1993, TNET.

[15]  Hao Jiang,et al.  Why is the internet traffic bursty in short time scales? , 2005, SIGMETRICS '05.

[16]  Sanjay Ghemawat,et al.  MapReduce: Simplified Data Processing on Large Clusters , 2004, OSDI.

[17]  Vern Paxson,et al.  End-to-end Internet packet dynamics , 1997, SIGCOMM '97.

[18]  GhemawatSanjay,et al.  The Google file system , 2003 .

[19]  Vern Paxson.  End-to-end internet packet dynamics , 1999, TNET.