On the self-similar nature of Ethernet traffic (extended version)
IEEE/ACM Transactions on Networking (TON)
High performance TCP in ANSNET
ACM SIGCOMM Computer Communication Review
SIGCOMM '95 Proceedings of the conference on Applications, technologies, architectures, and protocols for computer communication
The macroscopic behavior of the TCP congestion avoidance algorithm
ACM SIGCOMM Computer Communication Review
Dimensioning bandwidth for elastic traffic in high-speed data networks
IEEE/ACM Transactions on Networking (TON)
Loss probability calculations and asymptotic analysis for finite buffer multiplexers
IEEE/ACM Transactions on Networking (TON)
ICNP '97 Proceedings of the 1997 International Conference on Network Protocols (ICNP '97)
NIST Net: a Linux-based network emulation tool
ACM SIGCOMM Computer Communication Review
Proceedings of the 2004 conference on Applications, technologies, architectures, and protocols for computer communications
Self-configuring network traffic generation
Proceedings of the 4th ACM SIGCOMM conference on Internet measurement
Queueing Networks and Markov Chains
Queueing Networks and Markov Chains
Why is the internet traffic bursty in short time scales?
SIGMETRICS '05 Proceedings of the 2005 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
Part I: buffer sizes for core routers
ACM SIGCOMM Computer Communication Review
Part II: control theory for buffer sizing
ACM SIGCOMM Computer Communication Review
Part III: routers with very small buffers
ACM SIGCOMM Computer Communication Review
Open issues in router buffer sizing
ACM SIGCOMM Computer Communication Review
Update on buffer sizing in internet routers
ACM SIGCOMM Computer Communication Review
A critique of recently proposed buffer-sizing strategies
ACM SIGCOMM Computer Communication Review
Beyond the Model of Persistent TCP Flows: Open-Loop vs Closed-Loop Arrivals of Non-persistent Flows
ANSS-41 '08 Proceedings of the 41st Annual Simulation Symposium (anss-41 2008)
Measuring the congestion responsiveness of internet traffic
PAM'07 Proceedings of the 8th international conference on Passive and active network measurement
IEEE Journal on Selected Areas in Communications
Buffer sizing for 802.11-based networks
IEEE/ACM Transactions on Networking (TON)
Future internet video multicasting with essentially perfect resource utilization and QoS guarantees
Proceedings of the Nineteenth International Workshop on Quality of Service
FIFO Service with Differentiated Queueing
Proceedings of the 2011 ACM/IEEE Seventh Symposium on Architectures for Networking and Communications Systems
Hybrid Simulation of Packet-Level Networks and Functional-Level Routers
PADS '12 Proceedings of the 2012 ACM/IEEE/SCS 26th Workshop on Principles of Advanced and Distributed Simulation
SABRE: a client based technique for mitigating the buffer bloat effect of adaptive video flows
Proceedings of the 4th ACM Multimedia Systems Conference
Hi-index | 0.00 |
The issue of router buffer sizing is still open and significant. Previous work either considers open-loop traffic or only analyzes persistent TCP flows. This paper differs in two ways. First, it considers the more realistic case of nonpersistent TCP flows with heavy-tailed size distribution. Second, instead of only looking at link metrics, it focuses on the impact of buffer sizing on TCP performance. Specifically, our goal is to find the buffer size that maximizes the average per-flow TCP throughput. Through a combination of testbed experiments, simulation, and analysis, we reach the following conclusions. The output/input capacity ratio at a network link largely determines the required buffer size. If that ratio is larger than 1, the loss rate drops exponentially with the buffer size and the optimal buffer size is close to 0. Otherwise, if the output/input capacity ratio is lower than 1, the loss rate follows a power-law reduction with the buffer size and significant buffering is needed, especially with TCP flows that are in congestion avoidance. Smaller transfers, which are mostly in slow-start, require significantly smaller buffers. We conclude by revisiting the ongoing debate on "small versus large" buffers from a new perspective.