Observations on the dynamics of a congestion control algorithm: the effects of two-way traffic
SIGCOMM '91 Proceedings of the conference on Communications architecture & protocols
SOSP '03 Proceedings of the nineteenth ACM symposium on Operating systems principles
Source-level IP packet bursts: causes and effects
Proceedings of the 3rd ACM SIGCOMM conference on Internet measurement
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
Notes on burst mitigation for transport protocols
ACM SIGCOMM Computer Communication Review
MapReduce: simplified data processing on large clusters
OSDI'04 Proceedings of the 6th conference on Symposium on Opearting Systems Design & Implementation - Volume 6
Buffer sizing theory for bursty TCP flows
IZS '06 Proceedings of the 2006 International Zurich Seminar on Communications
Performance optimization of TCP/IP over 10 gigabit ethernet by precise instrumentation
Proceedings of the 2008 ACM/IEEE conference on Supercomputing
Helios: a hybrid electrical/optical switch architecture for modular data centers
Proceedings of the ACM SIGCOMM 2010 conference
Network traffic characteristics of data centers in the wild
IMC '10 Proceedings of the 10th ACM SIGCOMM conference on Internet measurement
TritonSort: a balanced large-scale sorting system
Proceedings of the 8th USENIX conference on Networked systems design and implementation
Network characteristics of video streaming traffic
Proceedings of the Seventh COnference on emerging Networking EXperiments and Technologies
On the impact of bursting on TCP performance
PAM'05 Proceedings of the 6th international conference on Passive and Active Network Measurement
Practical TDMA for datacenter ethernet
Proceedings of the 7th ACM european conference on Computer Systems
Less is more: trading a little bandwidth for ultra-low latency in the data center
NSDI'12 Proceedings of the 9th USENIX conference on Networked Systems Design and Implementation
Design and implementation of a consolidated middlebox architecture
NSDI'12 Proceedings of the 9th USENIX conference on Networked Systems Design and Implementation
Packet Trains--Measurements and a New Model for Computer Network Traffic
IEEE Journal on Selected Areas in Communications
Trickle: rate limiting YouTube video streaming
USENIX ATC'12 Proceedings of the 2012 USENIX conference on Annual Technical Conference
Scalable rule management for data centers
nsdi'13 Proceedings of the 10th USENIX conference on Networked Systems Design and Implementation
Integrating microsecond circuit switching into the data center
Proceedings of the ACM SIGCOMM 2013 conference on SIGCOMM
FasTrak: enabling express lanes in multi-tenant data centers
Proceedings of the ninth ACM conference on Emerging networking experiments and technologies
Circuit switching under the radar with REACToR
NSDI'14 Proceedings of the 11th USENIX Conference on Networked Systems Design and Implementation
Hi-index | 0.00 |
While numerous studies have examined the macro-level behavior of traffic in data center networks---overall flow sizes, destination variability, and TCP burstiness---little information is available on the behavior of data center traffic at packet-level timescales. Whereas one might assume that flows from different applications fairly share available link bandwidth, and that packets within a single flow are uniformly paced, the reality is more complex. To meet increasingly high link rates of 10 Gbps and beyond, batching is typically introduced across the network stack---at the application, middleware, OS, transport, and NIC layers. This batching results in short-term packet bursts, which have implications for the design and performance requirements of packet processing devices along the path, including middleboxes, SDN-enabled switches, and virtual machine hypervisors. In this paper, we study the burst behavior of traffic emanating from a 10-Gbps end host across a variety of data center applications. We find that at 10--100 microsecond timescales, the traffic exhibits large bursts (i.e., 10s of packets in length). We further find that this level of burstiness is largely outside of application control, and independent of the high-level behavior of applications.