A. Acharya and J. Saltz, "A study of internet round-trip delay," U. Maryland, Tech. Rep. CS-TR-3736, Jan. 1996.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....determine its cost. We then normalize this so that the maximum (one way) latency along an overlay link is 100ms. This forms the base latency on an overlay link. We then vary the latency between this base value and twice the base value. The variation is based on results from a wide area RTT study [19]. The details are not important here since our results do not depend on this; the interested reader may refer to [8] Setup for the study of the LIAC metric: For our study of the behavior of the LIAC metric, we use an emulation setup with a 40 node overlay network, with 119 overlay links. The ....

A. Acharya and J. Saltz, "A Study of Internet Round-Trip Delay, " Tech. Rep. CS-TR 3736, UMIACS-TR 96-97, University of Maryland, College Park, 1996-97.

....loss and the round trip delay experienced by data packets and their ACKs. Thus, in order to model TCP, round trip time and packet loss must be modeled. A. A Diffusion Model of round trip Time There has been extensive work focused on understanding the distribution of latency, e.g. 14] 15] [16]. However, there has been less work on developing a dynamical model of latency. Work that has investigated the dynamics of the roundtrip time includes [17] and [18] In [18] diffusion models of round trip delay are developed. These models are validated with extensive real network measurements. ....

A. Acharya and J. Saltz, "A study of internet round-trip delay," Department of Computer Science, University of Maryland, Tech. Rep. CS-TR3736, 1996.

....to differentiate between delay changes due to path changes and those due to congestion. We summarize our minimum RTT measurement process in Section V. Finally, we conclude in Section VI. II. RELATED WORK A large number of studies on packet delay have been presented in the literature, e.g. [11], 1] 12] 13] 14] To the best of our knowledge, however, no previous studies of packet delay and loss on the Internet have reported the appropriate number of probes and the required frequency to capture the minimum RTT of a path. Mukherjee in [13] reported on a study in which he collected ....

....packets over a 10minute period with an inter probe interval ranging from 8 ms to 500 ms. He also introduced the phase plot technique used in this paper as a method to analyze packet delays. Acharya and Saltz conducted a large scale study to characterize the roundtrip delay on 90 different paths [11]. Over a 48 hour period, they sent 64 byte ICMP packets to different hosts with an inter probe interval of 1 second. Their main observations are: the probability of capturing minimum RTT of a given path, within a short period of time is high (a similar observation was also reported in [1] 15] ....

[Article contains additional citation context not shown here]

A. Acharya and J. Saltz, "A Study of Internet Round-Trip Delay," University of Maryland, Department of Computer Science Technical Report CS-TR-3736, Dec. 1996.

....determine its cost. We then normalize this so that the maximum (one way) latency along an overlay link is 100ms. This forms the base latency on an overlay link. We then vary the latency between this base value and twice the base value. The variation is based on results from a wide area RTT study [19]. The details are not important here since our results do not depend on this; the interested reader may refer to [8] Intuitively, load balancing depends on the feedback loop, and not the finer variations in latency) Setup for the study of the LIAC metric: For our study of the behavior of the ....

A. Acharya and J. Saltz, "A Study of Internet Round-Trip Delay," University of Maryland, College Park, Tech. Rep. CS-TR 3736, UMIACSTR 96-97, 1996-97.

....we achieve reasonable levels of utilization in practice. As in TCP Vegas, maintaining running measures of minRTT and maxRTT have their limitations for example, if the network is in a state of persistent congestion, a bad estimate of minRTT is likely to be obtained. However, past studies [1, 44] have indicated that a good estimate of the minimum round trip delay can typically be obtained in a short time; our experience supports this claim. The use of minimum and maximum values makes the prototype sensitive to outliers. Using the first and ninety ninth percentile values could improve the ....

Anurag Acharya and Joel Saltz. A study of internet round-trip delay. Technical Report CS-TR-3736, University of Maryland, 1996.

....instance, when communicating with a server in Vanuatu from the UK, at peak times. In general, the value of latency depends on the network distance i.e. number of hops between the sender and the receiver, the bandwidth of the intermediate links and the current network tra#c conditions. Studies [29, 1, 27] show that, unlike previously prevailing opinion [3] there appears to be a strong link between geographic distance and communication delay, as longer physical distance means, in most cases, more network hops, and therefore longer round trip delay times. XenoServers will encourage the placement ....

Anurag Acharya and Joel Saltz. A study of internet round-trip delay. Technical Report CS-TR-3736, UMIACS and Department of Computer Science, University of Maryland, 1996.

....we achieve reasonable levels of utilization in practice. As in TCP Vegas, maintaining running measures of minRTT and maxRTT have their limitations for example, if the network is in a state of persistent congestion, a bad estimate of minRTT is likely to be obtained. However, past studies [1, 44] have indicated that a good estimate of the minimum round trip delay can typically be obtained in a short time; our experience supports this claim. The use of minimum and maximum values makes the prototype sensitive to outliers. Using the first and ninety ninth percentile values could improve the ....

Anurag Acharya and Joel Saltz. A study of internet round-trip delay. Technical Report CS-TR-3736, University of Maryland, 1996.

....will be able to submit code for execution on our platform, and the sponsor of the code will be billed for all the resources used or reserved during the course of execution. The XenoServers architecture is essential to help addressing some significant problems and needs of today. First, as studies [10, 1, 9] show, unlike previously prevailing opinion [2] there appears to be a strong link between geographic distance and communication latency. The XenoServers platform supports the placement of servers physically close to the clients, and therefore helps reducing latency and network load. Also, ....

Anurag Acharya and Joel Saltz. A study of internet round-trip delay. Technical Report CS-TR-3736, UMIACS and Department of Computer Science, University of Maryland, 1996.

....ployed 11 sites in 7 countries; whereas the majority of other researchers monitored transit TCP traffic at a single backbone router [2] 16] or inside several campus networks [7] for the duration ranging from several hours to several days. In addition, several other studies employed ICMP probes [1], 17] 18] UDP echo packets [5] 6] and multicast backbone (MBone) audio packets [25] With the exception of the last study, neither the setup, nor the type of probe traffic of prior work represented realistic real time streaming scenarios. Among the studies that specifically sent ....

A. Acharya and J. Saltz, "A Study of Internet Round-trip Delay," Technical Report CS-TR-3736, University of Maryland, December 1996.

....retransmission schemes may require a different retransmission timeout (RTO) estimator and leads us to believe that research in this area should be extended. We should further mention that the Internet has been extensively studied by various researchers using ICMP ping and traceroute packets [1], 9] 16] 17] 20] UDP echo packets [6] 7] and multicast backbone (MBone) audio packets [24] With the exception of the last one, similar observations apply to these studies neither the setup, nor the type of probe traffic represented realistic real time streaming scenarios. In ....

A. Acharya, and J. Saltz, "A Study of Internet Round-trip Delay, " Technical Report CS-TR-3736, University of Maryland, December 1996.

....A. Methodology We performed network measurements on the Internet by using the tool traceroute, which allows to trace the routing path between two Internet hosts [18] The resulting data allows to derive the hop count and a very rough estimation of the round trip time (rtt) As Acharya and Saltz [19] have shown, rtt has a large temporal variation and a value that well characterizes a rtt distribution requires a larger sample of single rtt measurements. However, we decided to take these rough estimations for evaluating the behaviour of our algorithms for different distance metrics. A second ....

....and their closest mServer is rather high (compared to the general average of hop count distances) To overcome this deficiency, a larger set of mServers would be needed. Secondly, our method for measuring round trip times is too simple for achieving characteristic rtt values. As Acharya and Saltz [19] have shown, the mode value of a larger sample of rtt measurements is a good characterisation of rtt distribution and remains valid for about 45 minutes. Despite these deficiencies, distance maps have been already successfully deployed for the coordination of mobile agents [9] The distance ....

A. Acharya, J. Saltz, "A Study of Internet Round-trip Delay," Technical report CS-TR-3736, Department of Computer Science, University of Maryland, USA, December 1996

....load when compared with the other document transfer times for document sizes of 1KB, 10KB and 100KB. While there is stability in the internet in terms of throughput, there is large temporal and spatial variation in round trip times (RTTs) that were found out by pinging a particular site repeatedly [1]. A pinging experiment was also introduced that was done whenever these documents were transferred. The site is sent a series of ping packets, five in total, one packet every second. The response times are noted and the average of these response times will give us an idea about the prevailing ....

Anurag Acharya and Joel Saltz, "A Study of Internet Round-trip Delay", Technical Report CS-TR-3736, University of Maryland, College Park, 1994.

....Inter cluster work exchange usually follows different schemes. Wide area networks (WAN) like the Internet, may offer enormous power, by combining hundreds of systems. On the other hand, they introduce many additional factors to the problem of interconnections, such as unpredictable latency [AS] or fault intolerance. There are a few works that endeavour to overcome these weaknesses. Some papers present distributed systems for arbitrary networks [H95] and some others offer solutions for the Internet [LSCI] S96] Yet, almost all of them merely cope with the tasks of processing or simple ....

A.Acharya,J.Saltz, "A Study of Internet Round-trip Delay" < http://www.cs.umd.edu/~acha/papers/latency-tr.html > mod. Jan 97

No context found.

A. Acharya and J. Saltz. A Study of Internet Round-Trip Delay. Technical Report CS-TR-3736, University of Maryland, December 1996.

....the ping period or five times the current round trip time estimate) the packet is retransmitted. Bandwidth values are reported without filtering whereas the latency values are filtered to eliminate shortterm variations. This filter was based on the results of an extensive study of Internet latency [1]. This study revealed that: 1) there is a lot of short term jitter in the latency measures but in most cases, the jitter is small; 2) there are occasional jumps in latency that appear only for a single observation; and (3) In most cases, a short window of values around the mode contains a large ....

A. Acharya and J. Saltz. A Study of Internet Round-Trip Delay. Technical Report CS-TR-3736, University of Maryland, December 1996.

....If an echo is not received within an expected interval, the maximum of the ping period or five times the current round trip time estimate) the packet is retransmitted. To eliminate short term variation in latency measures, we developed a filter based of an extensive study of Internet latency [1]. This study revealed that: 1) there is a lot of short term jitter in the latency measures but in most cases, the jitter is small; 2) there are occasional jumps in latency that appear only for a single observation; and (3) In most cases, a short window of values around the mode contains a large ....

A. Acharya and J. Saltz. A Study of Internet Round-Trip Delay. Technical Report CS-TR-3736, University of Maryland, December 1996.

No context found.

A. Acharya and J. Saltz, "A study of internet round-trip delay," U. Maryland, Tech. Rep. CS-TR-3736, Jan. 1996.

No context found.

ACHARYA, A., AND SALTZ, J. A Study of Internet Round-trip Delays. Tech. Rep. CS-TR-3736, University of Meryland Technical Report, 1996.

No context found.

A. Acharya and J. Saltz. A Study of Internet Round-Trip Delay. Technical Report CS-TR 3736, UMIACS-TR 96-97, University of Maryland, College Park, 1996-97.

No context found.

A. Acharya and J. Saltz. A Study of Internet Round-Trip Delay. Technical Report CS-TR 3736, UMIACS-TR 96-97, University of Maryland, College Park, 1996-97.

No context found.

Anurag Acharya and Joel Saltz. A study of Internet round-trip delay. Technical Report CS-TR 3736 and UMIACS-TR-96-97, University of Maryland, Department of Computer Science and UMIACS, December 1996.

No context found.

Anurag Acharya and Joel Saltz. A study of Internet round-trip delay. Technical Report CS-TR 3736 and UMIACS-TR-96-97, University of Maryland, Department of Computer Science and UMIACS, December 1996.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC