Packet Pair Research Articles

Accurate estimation of capacities and cross-traffic of all links in a path using ICMP timestamps

The estimation of the link capacity and its available bandwidth in an end-to-end path is crucial for network management, admission control and flow control for adaptive applications. This paper introduces an estimation mechanism able to accurately estimate the available bandwidth of all links in an end-to-end path, through its capacity and cross-traffic estimation. The estimation procedures resort to the dispersions of packet pairs and trains using the concepts of ICMP Timestamp and Traceroute that, efficiently used together, enable the measurement of the dispersions in all links in the path. These mechanisms were evaluated through simulation experiments, where we analyzed the influence of the several network parameters on each estimation mechanism. The results show that these estimation methods are able to accurately estimate both the capacity and the cross-traffic of all links in a path with moderate length and with low overhead.

The probe gap model can underestimate the available bandwidth of multihop paths

The Probe Gap Model (PGM) was proposed as a lightweight and fast available bandwidth estimation method. Measurement tools such as Delphi and Spruce are based on PGM. Compared to estimation methods that require multiple iterations with different probing rates, PGM uses a single probing rate and it infers the available bandwidth from a direct relation between the input and output rates of measurement packet pairs. An important assumption behind the PGM model is that the measured path has a single bottleneck link that determines the available bandwidth of the end-to-end path. In this letter, we show that, even though PGM is accurate in the case of a single queue, it cannot estimate the available bandwidth of multi-hop paths, even if there is a single bottleneck in the path. Whether PGM is accurate or not depends on the routing of cross traffic relative to the measurement traffic. PGM is accurate when the cross traffic follows the same path with the measurement traffic. In the general case, however, PGM can significantly underestimate the available bandwidth of an end-to-end path.

Network loss tomography using striped unicast probes

In this paper, we explore the use of end-to-end unicast traffic as measurement probes to infer link-level loss rates. We leverage off of earlier work that produced efficient estimates for link-level loss rates based on end-to-end multicast traffic measurements. We design experiments based on the notion of transmitting stripes of packets (with no delay between transmission of successive packets within a stripe) to two or more receivers. The purpose of these stripes is to ensure that the correlation in receiver observations matches as closely as possible what would have been observed if a multicast probe followed the same path to the receivers. Measurements provide good evidence that a packet pair to distinct receivers introduces considerable correlation which can be further increased by simply considering longer stripes. Using an M/M/1/K model for a link, we theoretically confirm this benefit for stripes. We also use simulation to explore how well these stripes translate into accurate link-level loss estimates. We observe good accuracy with packet pairs, with a typical error of about 1%, which significantly decreases as stripe length is increased.

Granular model of packet pair separation in Poissonian traffic

We introduce a theoretical model of packet pair separation based on a transient solution of the Takacs integro-differential equation. We show that in addition to the parameters of the fluid approximation (physical bandwidth and the average cross traffic rate) a new parameter characterizing the granularity of the cross traffic is necessary. These three parameters determine the dynamics of the queue in the diffusive approximation and all important distributions and averages of the packet separation even in multi-hop scenarios (assuming independent cross traffic on different hops). The model describes correctly the data collected in simulations, laboratory and Internet experiments. The adjustable model parameters are the physical bandwidth, the available bandwidth and the weighted average of the packet size of the cross traffic. We show that an implementation of the theoretical results can be used to estimate such parameters in packet chirp type measurements and can be a good candidate for improved available bandwidth estimation.

Packet triplet: a novel approach to estimate path capacity

Packet Pair is an end-to-end approach to estimate the capacity of an Internet path. Many studies have shown that such estimation exhibits a lot of bias due to the dynamic cross traffic. In this paper, we propose a novel approach, called Packet Triplet, which uses three back-to-back probing packets - different from two back-to-back probing packets in Packet Pair - to estimate the capacity of a network path. A new filtering technique is also introduced into Packet Triplet to further refine the path capacity estimation approach. The experiments demonstrate that such probing approach is effective under different cross traffic.

Stochastic analysis of packet-pair probing for network bandwidth estimation

In this paper, we perform a stochastic analysis of the packet-pair technique, which is a widely used method for estimating the network bandwidth in an end-to-end manner. There has been no explicit delay model of the packet-pair technique primarily because the stochastic behavior of a packet pair has not been fully understood. Our analysis is based on a novel insight that the transient analysis of the G/D/1 system can accurately describe the behavior of a packet pair, providing an explicit stochastic model. We first investigate a single-hop case and derive an analytical relationship between the input and the output probing gaps of a packet pair. Using this single-hop model, we provide a multi-hop model under an assumption of a single tight link. Our model shows the following two important features of the packet-pair technique: (i) The difference between the proposed model and the previous fluid model becomes significant when the input probing gap is around the characteristic value. (ii) The available bandwidth of any link after the tight link is not observable. We verify our model via ns-2 simulations and empirical results. We give a discussion on recent packet-pair models in relation to the proposed model and show that most of them can be regarded as special cases of the proposed model.

A Network Layer Approach to Enable TCP over Multiple Interfaces

The mobile Internet is set to become ubiquitous with the deployment of various wireless technologies. When heterogeneous wireless networks overlap in coverage, a mobile terminal can potentially use multiple wireless interfaces simultaneously. In this paper, we motivate the advantages of simultaneous use of multiple interfaces and present a network layer architecture that supports diverse multi-access services. Our main focus is on one such service provided by the architecture: Bandwidth Aggregation (BAG), specifically for TCP applications.While aggregating bandwidth across multiple interfaces can improve raw throughput, it introduces challenges in the form of packet reordering for TCP applications. When packets are reordered, TCP misinterprets the duplicate ACKS received as indicative of packet loss and invokes congestion control. This can significantly lower TCP throughput and counter any gains that can be had through bandwidth aggregation. To improve overall performance of TCP, we take a two-pronged approach: (1) We propose a scheduling algorithm that partitions traffic onto the different paths (corresponding to each interface) such that reordering is minimized. The algorithm estimates available bandwidth and thereby minimizes reordering by sending packet pairs on the path that introduces the least amount of delay. (2) A buffer management policy is introduced at the client to hide any residual reordering from TCP. We show through simulations that our network-layer approach can achieve good bandwidth aggregation under a variety of network conditions.

Packet-Dispersion Techniques and a Capacity-Estimation Methodology

The packet-pair technique aims to estimate the capacity of a path (bottleneck bandwidth) from the dispersion of two equal-sized probing packets sent back to back. It has been also argued that the dispersion of longer packet bursts (packet trains) can estimate the available bandwidth of a path. This paper examines such packet-pair and packet-train dispersion techniques in depth. We first demonstrate that, in general, packet-pair bandwidth measurements follow a multimodal distribution and explain the causes of multiple local modes. The path capacity is a local mode, often different than the global mode of this distribution. We illustrate the effects of network load, cross-traffic packet-size variability, and probing packet size on the bandwidth distribution of packet pairs. We then switch to the dispersion of long packet trains. The mean of the packet-train dispersion distribution corresponds to a bandwidth metric that we refer to as average dispersion rate (ADR). We show that the ADR is a lower bound of the capacity and an upper bound of the available bandwidth of a path. Putting all of the pieces together, we present a capacity-estimation methodology that has been implemented in a tool called pathrate. We report on our experiences with pathrate after having measured hundreds of Internet paths over the last three years.

Comparison of routing metrics for static multi-hop wireless networks

Routing protocols for wireless ad hoc networks have traditionally focused on finding paths with minimum hop count. However, such paths can include slow or lossy links, leading to poor throughput. A routing algorithm can select better paths by explicitly taking the quality of the wireless links into account. In this paper, we conduct a detailed, empirical evaluation of the performance of three link-quality metrics---ETX, per-hop RTT, and per-hop packet pair---and compare them against minimum hop count. We study these metrics using a DSR-based routing protocol running in a wireless testbed. We find that the ETX metric has the best performance when all nodes are stationary. We also find that the per-hop RTT and per-hop packet-pair metrics perform poorly due to self-interference. Interestingly, the hop-count metric outperforms all of the link-quality metrics in a scenario where the sender is mobile.

CapProbe

We present a new capacity estimation technique, called CapProbe. CapProbe combines delay as well as dispersion measurements of packet pairs to filter out samples distorted by cross-traffic. CapProbe algorithms include convergence tests and convergence speed-up techniques by varying probing parameters. Our study of CapProbe includes a probability analysis to determine the time it takes CapProbe to converge on the average. Through simulations and measurements, we found CapProbe to be quick and accurate across a wide range of traffic scenarios. We also compared CapProbe with two previous well-known techniques, pathchar and pathrate. We found CapProbe to be much more accurate than pathchar and similar in accuracy to pathrate, while providing faster estimation than both. Another advantage of CapProbe is its lower computation cost, since no statistical post processing of probing data is required.

CapProbe

The problem of estimating the capacity of an Internet path is one of fundamental importance. Due to the multitude of potential applications, a large number of solutions have been proposed and evaluated. The proposed solutions so far have been successful in partially addressing the problem, but have suffered from being slow, obtrusive or inaccurate. In this work, we evaluate CapProbe, a low-cost and accurate end-to-end capacity estimation scheme that relies on packet dispersion techniques as well as end-to-end delays. The key observation that enabled the development of CapProbe is that both compression and expansion of packet pair dispersion are the result of queuing due to cross-traffic. By filtering out queuing effects from packet pair samples, CapProbe is able to estimate capacity accurately in most environments, with minimal processing and probing traffic overhead. In fact, the storage and processing requirements of CapProbe are orders of magnitude smaller than most of the previously proposed schemes. We tested CapProbe through simulation, Internet, Internet2 and wireless experiments. We found that CapProbe error percentage in capacity estimation was within 10% in almost all cases, and within 5% in most cases.

그리드 네트워크에서 종단간 전송 지연 향상을 위한 bottleneck 링크 대역폭 측정 알고리즘

본 논문은 그리드 네트워크의 패킷 지연시간단축을 위한 병목 링크 대역폭 측정 알고리즘을 제안하고자 한다. 병목 링크 대역폭 측정방법으로 packet pair 알고리즘과 Paced Probes가 있으나 이들은 pacer 패킷으로 cross traffic의 영향을 줄이면서 Paced probe 패킷들이 링크를 지날 때 발생하는 시간 차이를 이용하여 병목 링크 대역폭을 구한다. 그러나 경로상의 병목 링크 대역폭을 측정할 뿐이지 병목현상이 발생한 지점은 알 수 없는 문제점이 있다. 이를 hop-by-hop 측정에 기반한 병목 대역폭 측정을 제시하여 그리드 네트워크에서 전체적인 패킷 지연 개선을 이룰 수 있다. hop-by-hop의 병목 대역폭을 측정하기 위해 paced probe 패킷에 타임스탬프 옵션을 이용하며 병목현상 해소를 통하여 패킷 전송 지연의 단축을 시뮬레이션을 통해 알 수 있다. 본 논문에서 제시한 알고리즘으로 병목 링크의 대역폭과 위치를 파악하여 그리드 ETP나 실시간 QoS를 보장하는 데이터 전송에 기여할 것으로 기대된다. This paper proposes a bottleneck link bandwidth measurement algorithm for reducing packet transmission delay within the grid network. There are two methods for measuring bottleneck link bandwidth：Packet Pair algorithm and Paced Probes algorithm. They measure bottleneck link bandwidth using the difference in arrival times of two paced probe packets of the same size traveling from the same source to destination. In addition, they reduce the influences of cross traffic by pacer packet. But there are some problems on these algorithms：it's not possible to know where bottleneck link occurred because they only focus on measuring the smallest link bandwidth along the path without considering bandwidth of every link on the path. So hop-by-hop based bottleneck link bandwidth measurement algorithm can be used for reducing packet transmission delay on grid network. Timestamp option was used on the paced probe packet for the link level measurement of bottleneck bandwidth. And the reducing of packet transmission delay was simulated by the solving a bottleneck link. The algorithm suggested in this paper can contribute to data transmission ensuring FTP and realtime QoS by detecting bandwidth and the location where bottleneck link occurred.

Unicast-based inference of network link delay distributions with finite mixture models

Providers of high quality-of-service over telecommunication networks require accurate methods for remote measurement of link-level performance. Recent research in network tomography has demonstrated that it is possible to estimate internal link characteristics, e.g., link delays and packet losses, using unicast probing schemes in which probes are exchanged between several pairs of sites in the network. We present a new method for estimation of internal link delay distributions using the end-to-end packet pair delay statistics gathered by back-to-back packet-pair unicast probes. Our method is based on a variant of the penalized maximum likelihood expectation-maximization (PML-EM) algorithm applied to an additive finite mixture model for the link delay probability density functions. The mixture model incorporates a combination of discrete and continuous components, and we use a minimum message length (MML) penalty for selection of model order. We present results of Matlab and ns-2 simulations to illustrate the promise of our network tomography algorithm for light cross-traffic scenarios.

Evaluation and characterization of available bandwidth probing techniques

The packet pair mechanism has been shown to be a reliable method to measure the bottleneck link capacity on a network path, but its use for measuring available bandwidth is more challenging. In this paper, we use modeling, measurements, and simulations to better characterize the interaction between probing packets and the competing network traffic. We first construct a simple model to understand how competing traffic changes the probing packet gap for a single-hop network. The gap model shows that the initial probing gap is a critical parameter when using packet pairs to estimate available bandwidth. Based on this insight, we present two available bandwidth measurement techniques, the initial gap increasing (IGI) method and the packet transmission rate (PTR) method. We use extensive Internet measurements to show that these techniques estimate available bandwidth faster than existing techniques such as Pathload, with comparable accuracy. Finally, using both Internet measurements and ns simulations, we explore how the measurement accuracy of active probing is affected by factors such as the probing packet size, the length of probing packet train, and the competing traffic on links other than the tight link.

Network delay tomography

The substantial overhead of performing internal network monitoring motivates techniques for inferring spatially localized information about performance using only end-to-end measurements. In this paper, we present a novel methodology for inferring the queuing delay distributions across internal links in the network based solely on unicast, end-to-end measurements. The major contributions are: 1) we formulate a measurement procedure for estimation and localization of delay distribution based on end-to-end packet pairs; 2) we develop a simple way to compute maximum likelihood estimates (MLEs) using the expectation-maximization (EM) algorithm; 3) we develop a new estimation methodology based on recently proposed nonparametric, wavelet-based density estimation method; and 4) we optimize the computational complexity of the EM algorithm by developing a new fast Fourier transform implementation. Realistic network simulations are carried out using network-level simulator ns-2 to demonstrate the accuracy of the estimation procedure.

Probing molecular dynamics with attosecond resolution using correlated wave packet pairs.

Spectroscopic measurements with increasingly higher time resolution are generally thought to require increasingly shorter laser pulses, as illustrated by the recent monitoring of the decay of core-excited krypton using attosecond photon pulses. However, an alternative approach to probing ultrafast dynamic processes might be provided by entanglement, which has improved the precision of quantum optical measurements. Here we use this approach to observe the motion of a D2+ vibrational wave packet formed during the multiphoton ionization of D2 over several femtoseconds with a precision of about 200 attoseconds and 0.05 ångströms, by exploiting the correlation between the electronic and nuclear wave packets formed during the ionization event. An intense infrared laser field drives the electron wave packet, and electron recollision probes the nuclear motion. Our results show that laser pulse duration need not limit the time resolution of a spectroscopic measurement, provided the process studied involves the formation of correlated wave packets, one of which can be controlled; spatial resolution is likewise not limited to the focal spot size or laser wavelength.

Revisiting the fair queuing paradigm for end-to-end congestion control

Today, the dominant paradigm for congestion control in the Internet is based on the notion of TCP friendliness. To be TCP-friendly, a source must behave in such a way as to achieve a bandwidth that is similar to the bandwidth obtained by a TCP flow that would observe the same round-trip time (RTT) and the same loss rate. However, with the success of the Internet comes the deployment of an increasing number of applications that do not use TCP as a transport protocol. These applications can often improve their own performance by not being TCP-friendly, which severely penalizes TCP flows. To design new applications to be TCP-friendly is often a difficult task. The idea of the fair queuing (FQ) paradigm as a means to improve congestion control was first introduced by Keshav (1991). While Keshav made a fundamental step toward a new paradigm for the design of congestion control protocols, he did not formalize his results so that his findings could be extended for the design of new congestion control protocols. We make this step and formally define the FQ paradigm as a paradigm for the design of new end-to-end congestion control protocols. This paradigm relies on FQ scheduling with per-flow scheduling and longest queue drop buffer management in each router. We assume only selfish and noncollaborative end users. Our main contribution is the formal statement of the congestion control problem as a whole, which enables us to demonstrate the validity of the FQ paradigm. We also demonstrate that the FQ paradigm does not adversely impact the throughput of TCP flows and explain how to apply the FQ paradigm for the design of new congestion control protocols. As a pragmatic validation of the FQ paradigm, we discuss a new multicast congestion control protocol called packet pair receiver-driven layered multicast (PLM).

Flow Control in ServerNet ® Clusters

This paper investigates the performance implications of several end-to-end flow control protocols in clusters based on the ServerNet® system-area network. The Static Window (SW) and Packet Pair (PP) flow control protocols are studied. Based on them, the Simplified Packet Pair (SPP) and the Alternating Static Window (ASW) protocols are defined. Previously, it has been proven that PP is stable for store-and-forward networks based on Rate Allocation Servers. The applicability of PP to wormhole-routing networks is studied. Simulation results for the performance characteristics are obtained and evaluated. It is shown that if high throughput is desired, ASW is the best method for controlling the average latency. On the other hand, if low throughput is acceptable, SPP can be applied to maintain low latencies.

PLM

A major challenge in the Internet is to deliver live audio/video content with a good quality and to transfer files to large number of heterogeneous receivers. Multicast and cumulative layered transmission are two mechanisms of interest to accomplish this task efficiently. However, protocols using these mechanisms suffer from slow convergence time, lack of inter-protocol fairness or TCP-fairness, and loss induced by the join experiments. In this paper we define and investigate the properties of a new multicast congestion control protocol (called PLM) for audio/video and file transfer applications based on a cumulative layered multicast transmission. A fundamental contribution of this paper is the introduction and evaluation of a new and efficient technique based on packet pair to infer which layers to join. We evaluated PLM for a large variety of scenarios and show that it converges fast to the optimal link utilization, induces no loss to track the available bandwidth, has inter-protocol fairness and TCP-fairness, and scales with the number of receivers and the number of sessions. Moreover, all these properties hold in self similar and multifractal environment.

End-to-end Internet packet dynamics

We discuss findings from a large-scale study of Internet dynamics conducted by tracing 20000 TCP bulk transfers between 35 Internet sites. Because we traced each 100-kbyte transfer at both the sender and the receiver, the measurements allow us to distinguish between the end-to-end behavior due to the different directions of the Internet paths, which often exhibit asymmetries. We: (1) characterize the prevalence of unusual network events such as out-of-order delivery and replication; (2) discuss a robust receiver-based algorithm for estimating bottleneck bandwidth that addresses deficiencies discovered in techniques based on packet pair; (3) investigate patterns of loss, finding that loss events are not well modeled as independent and, furthermore, that the distribution of the duration of loss events exhibits infinite variance; and (4) analyze variations in transit delays as indicators of congestion periods, finding that congestion periods also span a wide range of time scales.