Long-lived Flows Research Articles

On AIMD Congestion Control in Multiple Bottleneck Networks

We consider a linear algebraic model of the Additive-Increase Multiplicative-Decrease congestion control algorithm and present results on average fairness and convergence for multiple bottleneck networks. Results are presented for networks of both long-lived and short-lived flows.

Differentiated TCP User Perception over Downlink Packet Data Cellular Systems

Current downlink scheduling algorithms in the (enhanced) third-generation (3G) cellular packet systems exploit instantaneous channel status of multiple users, but most of them are blind to traffic information. To improve TCP users' perception of quality-of-services (QoSs), characterized by response delay, goodput, and always-on connectivity, we propose a cross-layer hierarchical scheduler with traffic awareness and channel dependence to properly prioritize buffer and radio resource allocation among different TCP classes. The scheduler has two tiers: at the IP layer, an intrauser scheduler enhances a common practice, i.e., the DiffServ-based buffer management, by dequeuing same-user TCP packets according to per-class specified and measured responsiveness; at the MAC layer, an interuser scheduler transmits the dequeued packets by considering the opportunistic channel states, mean throughput, and class ID of all users. Both tiers consider the online measured throughput, a cross-layer metric, to achieve resource and performance fairness and TCP classification. Experiments show that, compared with (variations of) proportional fairness (PF) and other schemes, our scheduler can notably speed up time-critical interactive TCP services (HTTP and TELNET) or TCP slow-starts with minor cost to bulk file transfer (FTP) or long-lived flows. It offers scalable and low-cost TCP performance enhancement over the emerging cellular systems

Night-to-night variation in the emission lines of the nucleus spectrum of the Seyfert galaxy NGC 3227

The results of the emission line study of the optical spectrum of the Seyfert galaxy NGC 3227 nucleus are presented. 53 spectra obtained during the maximum nucleus brightness on 12–15 January 1977 with the 6 m telescope were the basis of the investigations. It was shown that the profiles of the hydrogen lines broadened during 3 days. The amounts of broadening at the 0.5 intensity level of the Hα, Hβ and Hγ line profile peaks were 12, 35 and 44%, respectively. The Hβ line profile broadening was accompanied by a decrease in its equivalent width EWβ. The increase in the eqiuvalent width EW[O III] of the [O III] line during 3 days was more than 3σ. It was assumed that a 3 day flare is observed in the galaxy nucleus, which could be caused by shock in long-lived flows from the galaxy nucleus.

Night-to-night variability of the emission lines in the spectrum of the Seyfert galaxy NGC 3227 nucleus

This is a study of optical emission lines of the nucleus of the Seyfert galaxy NGC 3227. This work is based on 53 spectrograms obtained with the 6-m telescope at the brightness maximum of this galactic nucleus on January 12–15, 1977. A set of profiles of the hydrogen Hα, and Hβ lines and the equivalent widths (EWλ) of six emission lines and blends averaged over series of observations are presented. Broadening of the hydrogen line profiles was observed over the three days. At a level of 0.50 times the peak brightness the hydrogen line profiles were broadened by 12%, 35%, and 44%, respectively. The broadening of the Hβ line was accompanied by a reduction in its equivalent width (EW ; ). It is suggested that a flare was observed in the nucleus of the galaxy on a time scale of 3 days, possibly caused to a shock wave in long-lived flows from the galaxy nucleus.

Fair service for mice in the presence of elephants

We show how randomized caches can be used in resource-poor partial-state routers to provide a fair share of bandwidth to short-lived flows that are known as mice when long-lived flows known as elephants are present.

Ancient Lacustrine Hyperpycnites: A Depositional Model from a Case Study in the Rayoso Formation (Cretaceous) of West-Central Argentina

Abstract Hyperpycnal flows originate when sediment-laden fluvial discharges enter standing, lower-density water. Because of their excess density, the flows plunge near the river mouth and continue to travel basinward as a quasi-steady and fully turbulent underflow. The related deposits are hyperpycnites, and constitute a particular type of turbidite with poorly known facies and facies tracts. Although hyperpycnal flows seem to be quite common in present times, their occurrence in fossil strata is poorly documented. This paper addresses the characteristics and depositional processes of shallow lacustrine sandy hyperpycnites, on the basis of the field analysis of well-exposed Lower Cretaceous strata (Rayoso Formation) in the Neuquen Basin of west-central Argentina. The Rayoso Formation is composed of clastic (and minor evaporitic) red beds up to 1200 m thick, deposited in a shallow perennial lake of variable salinity affected by long-lived hyperpycnal flows. Main clastic facies are composed of fine-grained sandstones with climbing ripples and plane beds. Other common sandstone facies include massive beds and low-angle cross stratification. Most sandstone facies are related to traction plus fallout processes, and often show a vertical fluctuation between sedimentary facies originated under different traction-plus-fallout conditions within single beds. These fluctuations are interpreted to be evidence of deposition from flow fluctuations in sustained hyperpycnal flows. Most beds internally show the existence of three depositional phases, acceleration (AP), erosion plus bypass (EP), and deceleration (DP), which record the complete evolution of a single long-lived hyperpycnal flow at a fixed point. Additionally, the depositional evolution of a single long-lived hyperpycnal flow with distance records initially the progressive basinward migration of the AP and EP phases, and finally an overall deposition under the DP phase both in proximal and distal areas. This evolution provides an adequate explanation for the basinward extension of channelized features, and for the occurrence of fine-grained sandstones with climbing ripples both in proximal and distal positions within the same hyperpycnal system. Consequently, facies analysis derived from application of the Bouma sequence is not valid for deposits of quasi-steady hyperpycnal flows.

Long-Term Variability Of The Optical Emission Lines In The Nuclear Spectrum Of The Seyfert Galaxy Ngc 3227

Fifty-three spectrograms in the optical region (3700–7300 A) with the spectral resolution ∼8 A have been obtained for the Seyfert nucleus of the galaxy NGC 3227 with the 6-m telescope on January 1977, while the nucleus was in the historically important epoch of its extreme maximum brightness. Width of the slit was 1″, length of the box during the spectra measurements was 1.5″. Data obtained by us and those compiled from literature showed that profiles of the Balmer lines Hα, Hβ and Hγ are different, evidencing that the gas emitting these lines is highly self-absorbed. It was shown that narrow components of the profiles revealed by Rubin and Ford kept their positions (radial velocities) over 25 years. The components showed intensity variations compared to the central one from minimum to maximum of the nucleus brightness. The same variations were observed by us earlier in the emission line profiles of the NGC 7469 nucleus spectrum. Narrow profile components can reflect long-lived flows or jets in the broad line region (BLR). Obtained facts evidenced that long-lived gas streams and flows causing narrow components of broad line profiles presented not only when BLR of accretion disc is strong but when BLR of accretion disc declined. Blue bump at radial velocity of −5000 km/s in Hγ profile was revealed in spectra of high states of the nucleus, which disappeared in low state. One of the interpretations of this event can be in the framework of a model of one-sided or two-sided gas ejection during the high state of the nucleus, positive radial velocities of which being screened out by a circumnuclear disk.

SHRiNK: a method for enabling scaleable performance prediction and efficient network simulation

As the Internet grows, it is becoming increasingly difficult to collect performance measurements, to monitor its state, and to perform simulations efficiently. This is because the size and the heterogeneity of the Internet makes it time-consuming and difficult to devise traffic models and analytic tools which would allow us to work with summary statistics. We explore a method to side step these problems by combining sampling, modeling, and simulation. Our hypothesis is this: if we take a sample of the input traffic and feed it into a suitably scaled version of the system, we can extrapolate from the performance of the scaled system to that of the original. Our main findings are as follows. When we scale an IP network which is shared by short- and long-lived TCP-like and UDP flows and which is controlled by a variety of active queue management schemes, then performance measures such as queueing delay and drop probability are left virtually unchanged. We show this in theory and in simulations. This makes it possible to capture the performance of large networks quite faithfully using smaller scale replicas.

Design principles and algorithms for effective high-speed IP flow monitoring

In this paper, we present design principles and related implementation experience for building effective and scalable IP flow monitoring systems. We focus on the monitoring of high-speed links, where the short packet inter-arrival time and the huge number of simultaneous flows impose a number of challenging requirements. First, the small inter-arrival times imply that algorithms for packet attribution to flows must be fast and efficient. To this purpose, an appropriate model for hash-based packet classification is proposed. Second, also the update of per-flow information must be fast, which suggests that fast memories are needed in order to avoid that memory access becomes the system bottleneck. But fast memories are still expensive and small, while the number of simultaneous flow in high-speed links is large. Therefore, the need arises to introduce strategies that help in keeping memory requirements low: one of these is the fast identification of timed out flows. Finally, identifying and periodically reporting information about long-lived flows receiving a lot of traffic is of paramount importance for those applications that cannot simply wait for the termination of these flows to receive the corresponding information. We provide design principles and algorithms that can be applied to all these tasks. A comparative study of some of them is carried out and performance figures are obtained using significant metrics.

Improvement of LRU cache for the detection and control of long-lived high bandwidth flows

The issue of long-lived high bandwidth flow detection and control with partial flow state is addressed in this paper. The starting point of this work is the least recently used cache proposal, which is generally effective in detecting long-term fast (LTF) flows. However, its performance can be impaired by the following two factors. The first one is the existence of a large amount of short-lived or slow (SLS) flows. Because randomly sampled SLS flows take up quite a lot cache spaces, long-lived high bandwidth flows are often expelled from the cache mistakenly, which leads to degraded performance. The second is the attack from malicious high-bandwidth on/off flows, which can avoid detection by injecting packets on and off alternatively. This paper proposes a novel cache update scheme, Landmark LRU scheme, to improve the performance of the LRU cache in the detection and control of long-lived high bandwidth flows. Firstly, the superiority of the Landmark LRU scheme over the original LRU cache is demonstrated by Markovian analysis. Then simulations driven by real Internet traces prove that the Landmark LRU scheme achieves higher accuracy in long-lived high bandwidth flow detection than the original LRU scheme. Finally, the Landmark LRU cache is incorporated into a well-known active queue management scheme, Random Early Detection, to show its effectiveness in the control of malicious on/off LTF applications. NS-2 simulation shows that the L-LRU-RED scheme outperforms the LRU-RED stably in regulating on/off flows.

Using TCP flow-aggregation to enhance data experience of cellular wireless users

All over the world Global System for Mobile Communication (GSM) cellular mobile networks have been upgraded to support the "always-on" general packet radio service (GPRS). Despite the apparent availability of levels of bandwidth not dissimilar to that provided by conventional fixed-wire telephone modems, the user experience using GPRS is still considerably poor. In this paper, we examine the performance of protocols such as transmission control protocol (TCP) over GPRS, and show how certain network characteristics interact badly with TCP to yield problems such as: link underutilization for short-lived flows, excess queueing for long-lived flows, acknowledgment bunching, poor loss recovery, and gross unfairness between competing flows. We present the design and implementation of a transparent TCP proxy that mitigates many of these problems without requiring any changes to the TCP implementations in either mobile or fixed-wire end systems. The proxy is interposed in the cellular provider's network, and splits TCP connections transparently into two halves-the wired and wireless sides. Connections destined for the same mobile host are treated as an aggregate due to their statistical dependence. We demonstrate packet scheduling and flow control algorithms that use information shared between the connections to maximize performance of the wireless link, while interworking with unmodified TCP peers. We also demonstrate how fairness between flows and response to loss is improved, and that queueing and, hence, network latency is reduced. We discuss how TCP enhancing proxies could be transparently deployed, and conclude that installing such a proxy into GPRS network would be of significant benefit to users.

FlowMate: scalable on-line flow clustering

We design and implement an efficient on-line approach, FlowMate, for clustering flows (connections) emanating from a busy server, according to shared bottlenecks. Clusters can be periodically input to load balancing, congestion coordination, aggregation, admission control, or pricing modules. FlowMate uses in-band (passive) end-to-end delay measurements to infer shared bottlenecks. Delay information is piggybacked on feedback from the receivers, or, if impossible, TCP or application round-trip time estimates are used. We simulate FlowMate and examine the effects of network load, traffic burstiness, network buffer sizes, and packet drop policies on clustering correctness, evaluated via a novel accuracy metric. We find that coordinated congestion management techniques are more fair when integrated with FlowMate. We also implement FlowMate in the Linux kernel v2.4.17 and evaluate its performance on the Emulab testbed, using both synthetic and tcplib-generated traffic. Our results demonstrate that clustering of medium to long-lived flows is accurate, even with bursty background traffic. Finally, we validate our results on the Internet Planetlab testbed.

Optimizing Web delivery over wireless links: design, implementation, and experiences

World over wide-area wireless Global System for Mobile Communication (GSM) networks have been upgraded to support the general packet radio service (GPRS). GPRS brings always-on wireless data connectivity at bandwidths comparable to that of conventional fixed-line telephone modems. Unfortunately many users have found the reality to be rather different, experiencing very disappointing performance when, for example, browsing the Web over GPRS. In This work, we show what causes the web and its underlying transport protocol TCP to underperform in a GPRS wide-area wireless environment. We examine why certain GPRS network characteristics interact badly with TCP to yield problems such as: link underutilization for short-lived flows, excess queueing for long-lived flows, ACK compression, poor loss recovery, and gross unfairness between competing flows. We also show that many Web browsers tend to be overly aggressive, and by opening too many simultaneous TCP connections can aggravate matters. We present the design and implementation of a web optimizing system called GPRSWeb that mitigates many of the GPRS link-related performance problems with a simple software update to a mobile device. The update is a link-aware middleware (a local proxy) that sits in the mobile device, and communicates with a server proxy located at the other end of the wireless link, close to the wired-wireless border. The dual-proxy architecture collectively implements a number of key enhancements-an aggressive caching scheme that employs content-based hash keying to improve hit rates for dynamic content, a preemptive push of Web page support resources to mobile clients, resource adaptation to suit client capabilities, delta encoded data transfer of modified pages, DNS lookup migration, and a UDP-based reliable transport protocol that is specifically optimized for use over GPRS. We show that these enhancements results in significant improvement in web performance over GPRS links.

Night-to-night variation in the optical emission lines in the nuclear spectrum of the Seyfert galaxy NGC 3227

Fifty-three spectrograms in the optical region (3700–7300 A)with a spectral resolution of about 8 A have been obtained for the Seyfert nucleus of the galaxy NGC 3227 with the 6 m telescope on 12–15 January 1977 while the nucleus was in the historically important epoch of its extreme maximum brightness. The width of the slit was 1″, and the length of the box during the spectral measurements was 1.5″. The data obtained by us and those compiled from literature showed that profiles of the Balmer lines Hα, Hβ and Hγ are different, demonstrating that the gas emitting these lines is highly self-absorbed. The profiles of the Balmer lines contain various components that kept their positions (radial velocities) over 10 years. The components can reflect long-lived flows or jets in the broad-line region (BLR). A blue bump at a radial velocity of −5000 km s−1 in the Hγ profile was revealed. Variations in the intensities of the revealed components and broad wings of the emission lines Hβ and Hγ profiles were detected o...

Lifetime-based TCP service differentiation

We present a lifetime-based differentiation framework for TCP flows. The separation into two classes is based on a threshold technique. We introduce a scheme, FairShare, that handles the long-lived flows and achieves global max–min fairness. The short-lived flows are bundled together and a separate family of mechanisms, DAS, dynamically allocate bandwidth to match the load of newly instantiated short flows. Thus, two different objectives are met: fairness for the long flows, as well as reduced response times and reduced response time variance for the short flows. We argue that the applications are better served this way. Namely, applications generating short transfers are predominantly interested in short response times (e.g. HTTP requests/responses) while those generating long transfers (e.g. long FTP transfers) are at least provided a guarantee they are not penalized compared to other similar connections. By way of an example, we also demonstrate that elaborate traffic control schemes that do not perform classification of flows based on their anticipated lifetimes, may fail to efficiently utilize the network links.

Closed queueing network models of interacting long-lived TCP flows

This paper presents a new analytical model for the estimation of the performance of TCP connections. The model is based on the description of the behavior of TCP in terms of a closed queueing network. The model is very accurate, deriving directly from the finite state machine description of the protocol. The assessment of the accuracy of the analytical model is based on comparisons against detailed simulation experiments developed with the ns-2 package. The protocol model interacts with an IP network model that can take into account meshed topologies with several bottlenecks. Numerical results indicate that the proposed closed queueing network model provides accurate performance estimates in all situations. A novel and interesting property of the model is the possibility of deriving ensemble distributions of relevant parameters, such as, for instance, the transmission window size or the timeout probability, which provide useful insight into the protocol behavior and properties.

SIGNS OF FLOWS AND SHOCKS IN VARIABLE BALMER LINE PROFILES OF THE SEYFERT GALAXIES NGC 3227 AND NGC 7469

There is a widely held opinion that the narrow components of the emission line profiles in the spectra of AGNs are produced by gas streams and flows. The spectral data obtained and collected for the Seyfert galaxies NGC 3227 and NGC 7469 showed that the profiles of the Balmer lines contain components that maintained their positions over 9-25 years. We suppose that the components can be caused by long-lived flows or jets in broad-line regions. Night-to-night variations in the emission line Hβ and Hγ profiles over 3 days were detected. We suppose that short-term variability of the Balmer lines can be produced in short-lived shocks in gas flows or jets.

The performance of circuit switching in the internet

Recently there has been renewed interest in circuit switching because of the ease of building very high capacity optical circuit switches. This poster studies the performance, as seen by the end user, of an Internet that uses circuit switching instead of, or in addition to, packet switching. TCP represents over 90% of current Internet traffic, which makes our usage of the network very connection oriented. Therefore, the key performance metric for the end user is the completion time for each connection, which we refer to as response time.On the face of it, simply considering circuit switching as an alternative would seem a pointless exercise; the Internet is packet switched, and was deliberately built that way to enable the efficiencies afforded by statistical multiplexing, and thus it should provide a response time that is faster than that of circuit switching. However, link utilization is low, and falling, particularly at the core of the Internet, which means that statistical multiplexing is less important than it once was.In this poster, we explore the performance of a network where a new circuit is created for each application flow, with particular emphasis on the response time experienced by users. We use simple M/G/1 and M/G/N queues to model application flows in both packet switched and circuit switched networks, as well as ns-2 simulations to validate the results. We conclude that because of high-bandwidth long-lived flows, it does not make sense to use circuit switching in shared access or local area networks. At the same time, our results suggest that in the core of the network, where access links limit the maximum flow rate, and where high capacity is needed most, there is little or no difference in performance between circuit switching and packet switching. Given that circuit switches can be built to have higher capacity than packet switches, this suggests that a circuit switched core warrants further investigation.There are other performance metrics that are also very important when deciding whether circuit switching is a good technology for the Internet, such as scalability, simplicity, cost, reliability and availability, Quality of Service (QoS), or traffic engineering. We have used such criteria in other studies, but they will not be discussed in the poster with great detail given the space constraints. The end result is that we greatly benefit from the use of circuit switching in the backbone of the Internet.Our current use of the network is very connection oriented, and, thus, it fits well with a network core that is circuit switched, while the edges use packet switching. We may just use a gateway that maps packet switched flows (e.g. TCP connections) to circuits. This gateway creates and destroys circuits as flows come and go. Then the deployment of circuit switching in the core does not require any change in existing end hosts and routers, which will continue to use IP the same way as today, as the gateways will hide the existence of circuits in the core. With this network architecture, circuit switched clouds can be inserted gradually, without requiring a flag day in the Internet.

Instability of Zonal Flows in Rotating Spherical Shells: An Application to Jupiter

Measurements from the Galileo probe suggest that the zonal winds are deep rooted. Jupiter's high rotation rate makes it likely that the whole outer molecular H/He layer is involved in these long-lived jet flows. Assuming that the primary flows are geostrophic, and that the banded surface structure stretches right through the molecular H/He layer, we examine the conditions for such flows to be stable. As a first step, the linear stability of some prescribed banded zonal flows in a rotating spherical shell is explored. Incompressibility is assumed for simplicity, and the boundary condition is stress-free. We compare solutions for two aspect ratios, appropriate for the molecular H/He layers of Jupiter and Saturn, and two Ekman numbers ( E=10 −2 and E=10 −4). Convective and shear flow instabilities compete in our system. The convective instabilities are of the well-known columnar structure. Shear flow instabilities for the larger Ekman number are similar to the Taylor–Couette instability in rotating annuli. At the lower Ekman number, shear flow instabilities adopt a geostrophic character, assuming the form of rotating columns, similar to the convective instabilities. While the convective instability always sets in outside the tangent cylinder, shear instability can become unstable inside the tangent cylinder. If even a weak zonal flow is present inside the tangent cylinder, the flow is unstable to shear instability. This offers an explanation why the jovian zonal jet structure is much weaker at the higher latitudes that correspond to locations inside the tangent cylinder.

Modelling the performance of distributed admission control for adaptive applications

We outline a methodology for the performance analysis of distributed admission control for adaptive applications, or long-lived flows, where each flow can transmit at one of a discrete set of rates, and can switch between these rates. Specific examples are certain voice over IP or conferencing applications. In the Internet, UDP based protocols such as RTP have the potential to allow sources to alter their sending rates using periodic feedback, where again the rate can take values from a limited set.