Practical, distributed network coordinates

Cox, Russ; Dabek, Frank; Kaashoek, Frans; Li, Jinyang; Morris, Robert

doi:10.1145/972374.972394

Cited by 70 publications

(65 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…1 shows the error, figure 2 the stability of the original Vivaldi Neighbor Error algorithm with continuous adjustment (CA) and different values for c c and c e using the Azureus-to-PlanetLab dataset. In accordance to Cox et al [1] improved stability is only relevant in the initial phase: After 500 seconds, the stability of the embedding is similar for all parameter combinations. Moreover, we find through the course of our simulations that c e has only little influence on the embedding error and the stability.…”

Section: Resultssupporting

confidence: 81%

“…In order to evaluate the different Vivaldi variants, we built a simulator that takes either the static RTT matrices or the dynamic RTT traces as input. The method of deriving a Vivaldi simulation from a static matrix of pairwise RTTs was introduced by Cox et al [1]. We extended their idea with a dynamic trace-based simulation.…”

Section: Data Sets and Simulatormentioning

confidence: 99%

“…They assign positions in an euclidean space to the peers, so that their metric distance reflects the respective latencies. There are several such systems, but Vivaldi [1] has obtained the most attention. Its simple spring model is easy to understand and led to a practical implementation in the BitTorrent [2] client Azureus (now called Vuze).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Tuning Vivaldi: Achieving Increased Accuracy and Stability

Elser

Forschler²,

Fuhrmann

2009

Self-Organizing Systems

View full text Add to dashboard Cite

Abstract. Network Coordinates are a basic building block for most peer-to-peer applications nowadays. They optimize the peer selection process by allowing the nodes to preferably attach to peers to whom they then experience a low round trip time. Albeit there has been substantial research effort in this topic over the last years, the optimization of the various network coordinate algorithms has not been pursued systematically yet. Analyzing the well-known Vivaldi algorithm and its proposed optimizations with several sets of extensive Internet traffic traces, we found that in face of current Internet data most of the parameters that have been recommended in the original papers are a magnitude too high. Based on this insight, we recommend modified parameters that improve the algorithms' performance significantly.

show abstract

Section: Resultssupporting

confidence: 81%

Section: Data Sets and Simulatormentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Tuning Vivaldi: Achieving Increased Accuracy and Stability

Elser

Forschler²,

Fuhrmann

2009

Self-Organizing Systems

View full text Add to dashboard Cite

show abstract

“…RTT is important for applications such as anycast services [3] and content delivery overlays [4]. Vivaldi [5] is a well known system that estimates RTT by embedding end hosts in a 2+1D coordinate system in order to produce estimates of delays between arbitrary host pairs. iPlane and iPlane nano [6], [7], create an "atlas" of the Internet based on approximate routing paths known as BGP atoms [8].…”

Section: A Backgroundmentioning

confidence: 99%

On the relationship between fundamental measurements in TCP flows

Clegg

Araújo

Landa

et al. 2013

2013 IEEE International Conference on Communications (ICC)

View full text Add to dashboard Cite

Abstract-This paper considers fundamental measurements which drive TCP flows: throughput, RTT and loss. It is clear that throughput is, in some sense, a function of both RTT and loss. In their seminal paper Padyhe et al [1] begin with a mathematical model of the TCP sliding window evolution process and come up with an equation showing that TCP throughput is (roughly) proportional to 1/RT T √ p where p is the probability of packet loss. Their equation is shown to be consistent with data gathered on several links. This paper takes the opposite approach and analyses a large number of packet traces from well-known sources in order to create a data-driven estimate of the functions which relate TCP, loss and RTT. Regression analysis is used to fit models to connect the quantities. The fitted models show different behaviour from that expected in [1].

show abstract

“…Cox, Dabek et. al. proposed the Vivaldi algorithm [16], [7] based on an analogy to a network of physical springs. In this approach, the problem of minimizing the sum of errors is related to the problem of minimizing the potential energy of a spring system.…”

Section: A Previous Workmentioning

confidence: 99%

IDES: An Internet Distance Estimation Service for Large Networks

Mao

Saul

Smith

2006

IEEE J. Select. Areas Commun.

View full text Add to dashboard Cite

Abstract-The responsiveness of networked applications is limited by communications delays, making network distance an important parameter in optimizing the choice of communications peers. Since accurate global snapshots are difficult and expensive to gather and maintain, it is desirable to use sampling techniques in the Internet to predict unknown network distances from a set of partially observed measurements. This paper makes three contributions. First, we present a model for representing and predicting distances in large-scale networks by matrix factorization which can model sub-optimal and asymmetric routing policies, an improvement on previous approaches. Second, we describe two algorithms -singular value decomposition (SVD) and nonnegative matrix factorization (NMF)-for representing a matrix of network distances as the product of two smaller matrices.Third, based on our model and algorithms, we have designed and implemented a scalable system-Internet Distance Estimation Service (IDES)-that predicts large numbers of network distances from limited samples of Internet measurements. Extensive simulations on real-world data sets show that IDES leads to more accurate, efficient and robust predictions of latencies in largescale networks than existing approaches.

show abstract

Practical, distributed network coordinates

Abstract: -

Cited by 70 publications

References 17 publications

Tuning Vivaldi: Achieving Increased Accuracy and Stability

Tuning Vivaldi: Achieving Increased Accuracy and Stability

On the relationship between fundamental measurements in TCP flows

IDES: An Internet Distance Estimation Service for Large Networks

Contact Info

Product

Resources

About