The Virtue of Gentleness: Improving Connection Response Times with SYN Priority Active Queue Management

We have analyzed network traces of TCP connections and observed that there are many more losses during the handshake than for the remainder of the data exchange. Although recently developed AQM schemes can efficiently reduce latency related to bufferbloat, only more complex solutions relying on Fair Queueing (FQ) can improve the long delays resulting from the loss of a packet during the establishment of a TCP connection. In this paper, we propose SPA (SYN Priority Active queue management), a new low-complexity queue management scheme that combines the benefits and simplicity of the most recent AQM schemes while achieving performance comparable to more complex combinations of Fair Queueing and AQM. Our evaluation shows that the SPA performance is close to FQ CoDel for only a fraction of the complexity and resource usage.

[1]  Mirja Kühlewind,et al.  Evaluation of ARED, CoDel and PIE , 2014, EUNICE.

[2]  Paul E. McKenney,et al.  Stochastic fairness queueing , 1990, Proceedings. IEEE INFOCOM '90: Ninth Annual Joint Conference of the IEEE Computer and Communications Societies@m_The Multiple Facets of Integration.

[3]  Andrzej Duda,et al.  TCP over large buffers: When adding traffic improves latency , 2014, 2014 26th International Teletraffic Congress (ITC).

[4]  Ramesh Govindan,et al.  Reducing web latency: the virtue of gentle aggression , 2013, SIGCOMM.

[5]  Marco Mellia,et al.  TCP model for short lived flows , 2002, IEEE Communications Letters.

[6]  J Gettys,et al.  Bufferbloat: Dark Buffers in the Internet , 2011, IEEE Internet Computing.

[7]  Michael Welzl,et al.  Why Is This Web Page Coming Up so Slow? Investigating the Loss of SYN Packets , 2009, Networking.

[8]  Pascal Anelli,et al.  TCP Syn Protection: An Evaluation , 2012, ICON 2012.

[9]  Marco Mellia,et al.  Two schemes to reduce latency in short lived TCP flows , 2009, IEEE Communications Letters.

[10]  Stefan Savage,et al.  Modeling TCP latency , 2000, Proceedings IEEE INFOCOM 2000. Conference on Computer Communications. Nineteenth Annual Joint Conference of the IEEE Computer and Communications Societies (Cat. No.00CH37064).

[11]  Injong Rhee,et al.  Tackling bufferbloat in 3G/4G networks , 2012, Internet Measurement Conference.

[12]  Vaibhav Bajpai,et al.  Dissecting Last-mile Latency Characteristics , 2017, CCRV.

[13]  Michael Welzl,et al.  The new AQM kids on the block: An experimental evaluation of CoDel and PIE , 2014, 2014 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS).

[14]  Deborah Estrin,et al.  A first look at traffic on smartphones , 2010, IMC '10.

[15]  Donald F. Towsley,et al.  Modeling TCP throughput: a simple model and its empirical validation , 1998, SIGCOMM '98.

[16]  Matthew Mathis,et al.  The macroscopic behavior of the TCP congestion avoidance algorithm , 1997, CCRV.

[17]  Van Jacobson,et al.  BufferBloat: What’s Wrong with the Internet? , 2011, ACM Queue.

[18]  Sally Floyd,et al.  Adaptive RED: An Algorithm for Increasing the Robustness of RED's Active Queue Management , 2001 .

[19]  Emmanuel Lochin,et al.  Revisiting old friends: is CoDel really achieving what RED cannot? , 2014, CSWS@SIGCOMM.

[20]  Andrzej Duda,et al.  Two-way TCP connections: old problem, new insight , 2011, CCRV.

[21]  Vijay Subramanian,et al.  PIE: A lightweight control scheme to address the bufferbloat problem , 2013, 2013 IEEE 14th International Conference on High Performance Switching and Routing (HPSR).

[22]  Brighten Godfrey,et al.  More is less: reducing latency via redundancy , 2012, HotNets-XI.

[23]  Mark Allman,et al.  Comments on bufferbloat , 2013, CCRV.

[24]  Van Jacobson,et al.  Controlling queue delay , 2012, Commun. ACM.