TCP Data Research Articles

ImTCP: TCP with an inline measurement mechanism for available bandwidth

We introduce a novel mechanism for actively measuring available bandwidth along a network path. Instead of adding probe traffic to the network, the new mechanism exploits data packets transmitted in a TCP connection (inline measurement). We first introduce a new bandwidth measurement algorithm that can perform measurement estimates quickly and continuously and is suitable for inline measurement because of the smaller number of probe packets required and the negligible effect on other network traffic. We then show how the algorithm is applied in RenoTCP through a modification to the TCP sender only. We call the modified version of RenoTCP that incorporates the proposed mechanism ImTCP (Inline measurement TCP). The ImTCP sender adjusts the transmission intervals of data packets, then estimates available bandwidth of the network path between sender and receiver utilizing the arrival intervals of ACK packets. Simulations show that the new measurement mechanism does not degrade TCP data transmission performance, has no effect on surrounding traffic and yields acceptable measurement results in intervals as short as some RTTs (round-trip times). We also give examples in which measurement results help improving TCP performance.

Integration of streaming services and TCP data transmission in the Internet

We study in this paper the integration of elastic and streaming traffic on a same link in an IP network. We are specifically interested in the computation of the mean bit rate obtained by a data transfer. For this purpose, we consider that the bit rate offered by streaming traffic is low, of the order of magnitude of a small parameter ε ≪ 1 and related to an auxiliary stationary Markovian process ( X ( t ) ) . Under the assumption that data transfers are exponentially distributed, arrive according to a Poisson process, and share the available bandwidth according to the ideal processor sharing discipline, we derive the mean bit rate of a data transfer as a power series expansion in ε . Since the system can be described by means of an M / M / 1 queue with a time-varying server rate, which depends upon the parameter ε and process ( X ( t ) ) , the key issue is to compute an expansion of the area swept under the occupation process of this queue in a busy period. We obtain closed formulas for the power series expansion in ε of the mean bit rate, which allow us to verify the validity of the so-called reduced service rate at the first order. The second order term yields more insight into the negative impact of the variability of streaming flows.

Energy-Efficient TCP Operation in Wireless Sensor Networks

Many applications of wireless sensor networks require connectivity to external networks to let monitoring and controlling entities communicate with the sensors. By using the TCP/IP protocols inside the sensor network, external connectivity can be achieved anywhere in the sensor network. In such IP-based sensor networks, TCP can be used for remote management and reprogramming of sensor nodes. However, the high bit error rates in multi-hop sensor networks lead to energy-inefficiencies that reduce the lifetime of the sensor network. This paper introduces and compares two approaches to support energy-efficient operation of TCP in sensor networks: Distributed TCP Caching (DTC) and TCP Support for Sensor networks (TSS). Both concepts allow intermediate sensor nodes to cache TCP segments and to perform local retransmissions in case of errors. This allows reducing the total number of packet transmissions in the sensor network when transferring data to or from a sensor node. DTC caches and immediately forwards TCP data segments, whereas TSS does not forward a cached segment until it knows that the previous segment has been successfully received by the next hop node. We show by simulation that both approaches significantly reduce the number of TCP segment and acknowledgement transmissions. Their performance differs slightly depending on the error rate. Both approaches have also slightly different needs in buffer requirements and TCP options to be supported.

Architecture and Performance Evaluation for Redundant Multicast Transmission Supporting Adaptive QoS

In this paper we describe the architecture of an application that was developed for the transmission of multimedia data, using the multicast mechanism, over the Internet. There are two major issues that have to be considered when designing and implementing such a service, the fairness and the adaptation schemes. The fairness problem results from the fact that Clients with different capabilities have to be served. In our application we use a mechanism that categorizes the Clients into a number of groups according to each Client's capabilities and (the mechanism) serves each group of Clients with a different multicast stream. With the term “capabilities” we do not only mean the processing power of the Client, but also the capacity and the condition of the network path towards that Client. Because of today's Internet heterogeneity and the lack of Quality of Service (QoS) support, the Server cannot assume that the Clients will permanently be able to handle a specific bit rate. We have therefore implemented an additional mechanism for the intra-stream bit rate adaptation. The proposed mechanism uses a “friendly” to the network users congestion control policy to control the transmission of the data. We evaluate the adaptive multicast transmission mechanism through a number of experiments and a number of simulations in order to examine its behaviour to a heterogeneous group of Clients and its behaviour against TCP and UDP data streams.

Mathematical analysis of Westwood+TCP congestion control

TCP congestion control is based on an additive-increase/multiplicative-decrease (AIMD) probing paradigm aimed at adapting the sending rate of TCP data sources to match the Internet time-varying available bandwidth. Westwood+ TCP has been recently proposed to improve the tracking of available bandwidth of classic TCP. It is based on an end-to-end estimate of the available bandwidth, which is obtained by properly counting and filtering the stream of acknowledgement packets. The estimate is used to adaptively decrease the congestion window and slow start threshold after congestion so that it can be said that Westwood+ TCP substitutes the classic multiplicative decrease with an adaptive decrease paradigm. The authors propose a mathematical analysis of the additive-increase/adaptive-decrease (AIADD) paradigm to analyse the steady-state throughput provided by Westwood+ TCP and investigate the intra-protocol fairness of the AIADD paradigm and the inter-protocol friendliness between AIADD and AIMD algorithms. It is shown that (i) both classic and Westwood+ TCP provide a throughput that is proportional to 1/√p, where p is the segment drop probability, that is they are friendly to each other; and (ii) the throughput of Westwood+ TCP is proportional to 1/√RTT, where RTT is the round trip time, whereas the throughput of Reno TCP is proportional to 1/RTT, i.e. Westwood+ TCP improves the intra-protocol fairness. Finally, Ns-2 simulations are reported in order to validate the mathematical model in the presence of a wide range of network loads, loss probabilities and round trip times.

Efficient resource allocation for China's 3G/4G wireless networks

The all-IP DiffServ model is expected to be the most promising architecture for QoS provisioning in China's next-generation wireless networks, due to its scalability, convenience for mobility support, and capability of interworking heterogeneous radio access networks. This article focuses on efficient resource allocation in a wireless DiffServ architecture. Resource utilization efficiency is particularly important for China's wireless networks as the mobile user density in China is and will continue to be much higher than that in other countries. More specifically, we propose a novel buffer sharing scheme to provide assured service for real-time layer-coded multimedia traffic, which can guarantee the specific packet loss requirement of each layer with UDP as the transport layer protocol. An adaptive optimal buffer configuration can be applied to achieve maximum resource utilization over the time-varying channel. Assured service is also provided to TCP data traffic for guaranteed throughput, where the cross-layer coupling between the TCP layer and link layer is exploited to efficiently utilize the wireless resources.

TCP와 UDP 플로우 간의 공정성 개선을 위한 새로운 큐 관리 알고리즘

인터넷의 혼잡상황을 해결하기 위하여 제안된 RED(Random Early Detection)와 같은 능동적 큐 관리(Active Queue Management) 알고리즘들은 TCP 데이터에 대하여 우수한 혼잡제어 효과를 나타낸다. 그러나 TCP와 UDP가 병목 링크를 공유하는 경우 불공정성 문제와 큐에서의 지연시간이 길어지는 문제점을 가지고 있다. 본 논문에서는 공정성을 개선함과 동시에 큐 지연시간을 감소할 수 있는 새로운 큐 관리 알고리즘인 PSRED(Protocol Sensitive RED) 알고리즘을 제안하였다. PSRED 알고리즘은 트래픽의 프로토콜 필드를 이용하여 플로우의 종류를 구분하고 각기 다른 패킷폐기함수를 적용함으로써 공정성을 개선하고 평균 큐 길이를 줄일 수 있다. AQM (Active Queue Management) techniques such as RED (Random Early Detection) which be proposed to solve the congestion of internet perform congestion control effectively for TCP data. However, in the situation where TCP and UDP share the bottleneck link, they can not solve the problems of the unfairness and long queueing delay. In this paper, we proposed an simple queue management algorithm, called PSRED (Protocol Sensitive RED), that improves fairness and decreases queueing delay. PSRED algorithm improves fairness and decreases average queue length by distinguishes each type of flow in using protocol field of packets and applies different drop functions to them respectively.

Architecture for a hardware-based, TCP/IP content-processing system

The transmission control protocol is the workhorse protocol of the Internet. Most of the data passing through the Internet transits the network using TCP layered atop the Internet protocol (IP). Monitoring, capturing, filtering, and blocking traffic on high-speed Internet links requires the ability to directly process TCP packets in hardware. High-speed network intrusion detection and prevention systems guard against several types of threats. As the gap between network bandwidth and computing power widens, improved microelectronic architectures are needed to monitor and filter network traffic without limiting throughput. To address these issues, we've designed a hardware-based TCP/IP content-processing system that supports content scanning and flow blocking for millions of flows at gigabit line rates. The TCP splitter2 technology was previously developed to monitor TCP data streams, sending a consistent byte stream of data to a client application for every TCP data flow passing through the circuit. The content-scanning engine can scan the payload of packets for a set of regular expressions. The new TCP-based content-scanning engine integrates and extends the capabilities of the TCP splitter and the old content-scanning engine. IP packets travel to the TCP processing engine from the lower-layer-protocol wrappers. Hash tables are used to index memory that stores each flow's state.

Seamless handover in terrestrial radio access networks: a case study

Next-generation networks will overlay heterogeneous networking technologies. In order to enable seamless communications in such an environment, providing support for efficient handover between the various access technologies will play a crucial role. The present article provides a quantitative analysis of it Mobile IPv4-based WLAN-GPRS handover prototype, identifies a number of side effects related to the link layer and routing mechanisms, and presents the impact of handover on UDP and TCP data traffic as well as on Mobile IP signaling itself.

IGrid2002 demonstration: bandwidth from the low lands

We report on a demonstration of several complementary high-performance end-to-end active network throughput measurement tools. These include: the PingWorld Java applet that displays the Round Trip Time (RTT) and losses to sites around the world from the user’s host; the multi-path analysis tool that visualizes common paths from traceroutes to selected hosts; the IEPM high-performance BandWidth monitoring toolkit which gives achievable throughput for several types of TCP data transfer applications; and the Available Bandwidth Estimation (ABWE) tool that reports in real-time the available bandwidth to several paths within the range from Mbits/s to Gbits/s. We also demonstrated sending high-speed data from 4 hosts at iGrid2002 to over 30 hosts in 10 countries to simulate a high energy physics experiment distributing data to collaborators. The demonstration utilized the high-speed, long latency, trans-Atlantic network set up for iGrid2002 in Amsterdam during September 2002.

Remote access and display of neutron data

With the proliferation of high-intensity neutron sources and involvement of scientists from diverse areas, it is important to provide methods for remote access and display of data. We have developed an integrated spectral analysis workbench (ISAW) that can read, merge or combine, operate on, and visualize large arrays of data. ISAW is written in Java to allow it to run on common user workstations and to facilitate network communications. Recently, we added the ability to remotely access live data made available through a data server running on the control computer. This data server receives UDP data packets from a data sender running as part of the data-acquisition system and sends TCP data packets requested by ISAW clients running on user desktop systems.

TCP/IP data transfer over the DECT air interface with multibearer capability and support of asymmetric flows

This works aims at evaluating achievable performance for TCP/IP data transfer over DECT radio interface, by verifying the ability of DECT lower layers protocols, namely DLC and MAC, to mask the radio interface unreliability and to provide flexible transfer means for TCP data. Major points are: (i) the interplay of error recovery mechanisms at the various layers; (ii) the effectiveness of providing multibearer capabilities (i.e. allowing slot bundling in the radio access); (iii) the possible gains brought about by MAC asymmetric connections in view of essentially asymmetric TCP data flows.

The NGI ONRAMP test bed: reconfigurable WDM technology for next generation regional access networks

Next generation internet optical network for regional access using multi-wavelength protocols (NGI ONRAMP) is a pre-competitive consortium sponsored by DARPA. Its mission is to develop architectures, protocols, and algorithms for wavelength division multiplexing (WDM)-based regional access networks that will effectively support the NGI. A reconfigurable WDM test bed is being built to demonstrate some of the key thrusts of the consortium, including dynamic service provisioning and optical flow switching, service protection in the optical domain, medium access control protocols, and network control and management geared for the efficient transport of Internet traffic over WDM networks. The ONRAMP test bed will consist of a feeder network connecting via access nodes to distribution networks on which the end users reside. ONRAMP network reconfiguration is enabled by access nodes that contain both optical and electronic switching components, allowing data traffic to be routed all-optically through the network or to be switched and aggregated by electronic Internet protocol (IP) routers. This paper describes the goals and basic architecture of the ONRAMP test bed, as well as the design, construction, and characterization of the network access nodes. To illustrate test bed operation, we demonstrate optical flow switching over the test bed that achieves Gb/s throughput of TCP data between end user workstations.

Streaming Video Using Dynamic Rate Shaping and TCP Congestion Control

We present a new technique for streaming real time video on today's Internet, based ondynamic rate shapingandTCP congestion control. Dynamic rate shaping is a signal processing technique that adapts the rate of compressed video (MPEG-1, MPEG-2, H.26x) to dynamically varying bandwidth constraints. This provides an interface (or filter) between the source and the network, with which the encoder's output (either live or stored) can be perfectly matched to the network's available bandwidth. We couple this adaptation capability with the use of a new semi-reliable protocol that uses the TCP congestion window to pace the delivery of data into the network, but without using other TCP algorithms that are poorly suited to real time media. Use of TCP congestion control ensures that the protocol competes fairly with all other TCP data and that it optimally shares the available bandwidth. It also avoids the latency problems commonly associated with TCP. In addition, we describe a real application that uses this approach to stream MPEG video on the Internet. We present several experiments, performed in both a controlled environment and the wide area Internet, that were used to evaluate the effectiveness and fairness of the scheme. The results show that the proposed solution achieves superior video quality while at the same time providing fairness by sharing bandwidth equally with other non-real-time connections.

Predicting the radiation control probability of heterogeneous tumour ensembles: data analysis and parameter estimation using a closed-form expression

A closed-form formula describing the tumour control probability (tcp) of a heterogeneous collection of tumours has been obtained by analytically averaging the homogeneous double-exponential tcp formula over inter-tumour distributions of clonogen radiosensitivity, density and repopulation rate, tumour volume and dose. The formula can be straightforwardly and relatively quickly fitted to clinical data, yielding radiobiological parameter values for use in tcp modelling. The formula was fitted to published tcp data which catalogued tumour control records grouped by dose and tumour volume, and treatment duration. Fitted parameter values, confidence intervals and goodness-of-fit statistics were determined. The sets of parameter values obtained are unique only to within a scaling factor. The formula provides non-rejectable fits to data which grouped tcp by dose and volume when radiosensitivity parameters take values close to laboratory estimates, the fitted volume dependence parameter, however, taking rather high values. Good fits are obtainable with the intuitively reasonable volume parameter value of one, but with radiosensitivity values around one-third of their laboratory estimates. Non-rejectable fits to data which grouped tcp by dose and treatment duration may be obtained with radiosensitivity and repopulation rate parameters lying close to laboratory estimates.

Improving TCP throughput over two-way asymmetric links

The sharing of a common buffer by TCP data segments and acknowledgments in a network or internet has been known to produce the effect of ack compression , often causing dramatic reductions in throughput. We study several schemes for improving the performance of two-way TCP traffic over asymmetric links where the bandwidths in the two directions may differ substantially, possibly by many orders of magnitude. These approaches reduce the effect of ack compression by carefully controlling the flow of data packets and acknowledgments. We first examine a scheme where acknowledgments are transmitted at a higher priority than data. By analysis and simulation, we show that prioritizing acks can lead to starvation of the low-bandwidth connection. Next, we introduce and analyze a connection-level backpressure mechanism designed to limit the maximum amount of data buffered in the outgoing IP queue of the source of the low-bandwidth connection. We show that this approach, while minimizing the queueing delay for acks, results in unfair bandwidth allocation on the slow link. Finally, our preferred solution separates the acks from data packets in the outgoing queue, and makes use of a connection-level bandwidth allocation mechanism to control their bandwidth shares. We show that this scheme overcomes the limitations of the previous approaches, provides isolation, and enables precise control of the connection throughputs. We present analytical models of the dynamic behavior of each of these approaches, derive closed-form expressions for the expected connection efficiencies in each case, and validate them with simulation results.

Convergence between public switching and the Internet

The Internet has developed into a global data network that is highly accepted as a multimedia information platform which has the potential to develop into an alternative carrier network in the future. Telecom operators increasingly act as Internet service providers to maximize network utilization, to attract or retain customers, and to generate additional revenue. To leverage their installed base in the PSTN/ISDN the optimal strategy for telco ISPs is the integration of their point of presence into the central office packetizing and grooming of IP traffic in the local office relieves load on the PSTN/ISDN trunk network, resolves existing bottlenecks due to Internet traffic upstream of the CO, and creates new opportunities for revenue-generating features for both telephony and Internet subscribers. This article shows that current telecommunication network elements can be upgraded with innovate cutting-edge technology to build a solid basis for a seamless multimedia network of tomorrow, thereby enabling telecom operators' and service providers' tremendous investment in existing network infrastructure to be fully utilized. An integrated Internet services platform is presented which turns the CO switch into an optimized link between the PSTN/ISDN and the Internet. Technically, it is proposed that the central office be expanded with an integrated Internet PoP (IPOP) configured from the following IP components: an internal high-speed data backbone (ATM or Ethernet); modem pools to terminate dial-in calls from analog modems using the PPP protocol; protocol handlers for UDP, TCP, IP, and lower-layer data protocols (X.25, frame relay, SMDS, etc.); access to data networks; IP router, RADIUS server, and name server database; a contents server (optionally), enabling telco ISPs to become content providers. This effectively turns the CO into an Internet access point that integrates smoothly into the existing telco OA&M/TMN. The IP functions integrated in the IPOP can interwork closely with PSTN/ISDN call processing.

Network performance effects of HTTP/1.1, CSS1, and PNG

We describe our investigation of the effect of persistent connections, pipelining and link level document compression on our client and server HTTP implementations. A simple test setup is used to verify HTTP/1.1's design and understand HTTP/1.1 implementation strategies. We present TCP and real time performance data between the libwww robot [27] and both the W3C's Jigsaw [28] and Apache [29] HTTP servers using HTTP/1.0, HTTP/1.1 with persistent connections, HTTP/1.1 with pipelined requests, and HTTP/1.1 with pipelined requests and deflate data compression [22]. We also investigate whether the TCP Nagle algorithm has an effect on HTTP/1.1 performance. While somewhat artificial and possibly overstating the benefits of HTTP/1.1, we believe the tests and results approximate some common behavior seen in browsers. The results confirm that HTTP/1.1 is meeting its major design goals. Our experience has been that implementation details are very important to achieve all of the benefits of HTTP/1.1.For all our tests, a pipelined HTTP/1.1 implementation outperformed HTTP/1.0, even when the HTTP/1.0 implementation used multiple connections in parallel, under all network environments tested. The savings were at least a factor of two, and sometimes as much as a factor of ten, in terms of packets transmitted. Elapsed time improvement is less dramatic, and strongly depends on your network connection.Some data is presented showing further savings possible by changes in Web content, specifically by the use of CSS style sheets [10], and the more compact PNG [20] image representation, both recent recommendations of W3C. Time did not allow full end to end data collection on these cases. The results show that HTTP/1.1 and changes in Web content will have dramatic results in Internet and Web performance as HTTP/1.1 and related technologies deploy over the near future. Universal use of style sheets, even without deployment of HTTP/1.1, would cause a very significant reduction in network traffic.This paper does not investigate further performance and network savings enabled by the improved caching facilities provided by the HTTP/1.1 protocol, or by sophisticated use of range requests.

Hardware-software architecture of the SWAN Wireless ATM network

The SWAN (Seamless Wireless ATM Network) system provides end-to-end ATM connectivity to mobile end-points equipped with RF transceivers for wireless access. Users carrying laptops and multimedia terminals can seamlessly access multimedia data over a backbone wired network while roaming among room-sized cells that are equipped with basestations. The research focus on how to make ATM mobile and wireless distinguishes SWAN from present day mobile-IP based wireless LANs. This paper describes the design and implementation of the ATM-based wireless last-hop, the primary components of which are the air-interface control, the medium access control, and the low-level ATM transport and signalling. The design is made interesting by its interplay with ATM; in particular, by the need to meaningfully extend over the wireless last-hop the service quality guarantees made by the higher level ATM layers. The implementation, on the other hand, is an example of hardware-software co-design and partitioning. A key component of the wireless hop implementation is a custom designed reconfigurable wireless adapter card called FAWN (Flexible Adapter for Wireless Networking) which is used at the mobiles as well as at the basestations. The functionality is partitioned three-way amongst dedicated reconfigurable hardware on FAWN, embedded firmware on FAWN, and device driver software on a host processor. Using an off-the-shelf 625 Kbps per channel radio, several of which can be supported by a single FAWN adapter to provide multiple channels, per-channel unidirectional TCP data throughput of 227 Kbps (or, 454 Kbps bidirectional) and per-channel unidirectional native ATM data throughput of 210 Kbps (or, 420 Kbps bidirectional) have been obtained.

Buffer and bandwidth allocation for TCP data traffic: experimental results

This paper discusses the dimensioning of buffers and the bandwidth allocation for data traffic in the ATM network. Data traffic is notoriously complex and bursty, making such dimensioning a difficult task. However, the COMBINE project, when dimensioning their InterWorking Units (IWUs), adopted a Poissonian packet arrival model, based upon the argument that burstiness at timescales higher than that of a packet arrival are a problem to be tackled by flow control at higher layers. This paper presents experimental results from the COMBINE testbed that show that this hypothesis was justified and that good TCP goodput was obtained based upon this dimensioning approach, due to TCP's ability to adapt to network congestion. However, it is also shown that it was the TCP algorithm that was ultimately responsible for controlling the packet loss ratio in the network and not the bandwidth allocation or buffer size. The results highlight the importance of taking into account the mutual influence between the ATM layer and the transport layer congestion control algorithms.