Search

The last few years have witnessed an explosion of content-rich services over the Internet, and thereby content distribution itself grows into one of the most important Internet applications. Content delivery networks or content distribution networks (CDNs) first emerged to address the efficiency of content distribution over the Web for the end users, since content delivery has become important for improvement of Web performance [53]. In order to improve accessibility, decrease access delay, maximize bandwidth utilization, and maintain correctness for the users, CDN distributes content to a group of geographically dispersed cache servers. Note that these servers are located as close as possible to the users.

Part 6: Load Balancing, Analysis, and Management; International audience; The growing demand for quality and performance has become a discriminating factor in the field of software applications. Specifically in the area of web applications, performance has become a key factor for success creating the need for new types of performance evaluation models and methods capable of representing the dynamic characteristics of web environments.In this paper we will recall seminal work in this area and present AWPS, a tool for automatic web performance simulation and prediction. AWPS is capable of (a) automatically creating a web performance simulation and (b) conducting trend analysis of the system under test. The operation and usage of this tool is demonstrated in a case study of a two-tier architecture system.

An attempt was made to evaluate Web performance by making a 3D forecast of Web site performance while resource downloading using a geostatistic simulation method Turning Bands for the first time. Data for the research were obtained in an active experiment conducted by a multi-agent measurement system MWING performing monitoring of a common group of Web sites from different agent locations. In this study we use a measurement database collected by an agent from Wroclaw. A preliminary analysis of measurement data was conducted. Next a structural analysis of data was conducted and a spatial forecast of the total time of downloading data from web servers with a one-week time advance was calculated. Forecast results were analyzed in detail and then directions for further research were suggested.The article introduces the first innovative use of the Turning Bands method in the prediction of Internet performance. First results permit to find that the mentioned predictive method can be effective.

This chapter introduces Webometrics as an emerging discipline focused on the understanding and assessment of the flow of Web based academic information. It describes the principal web-based techniques and tools used to evaluate the performance of higher education websites and to explain how these information networks are created and modelled. The chapter begins with an introduction to Webometrics: its origin and evolution, its theoretical framework and its relationship with other web disciplines. The principal indicators and measures used to quantify the development of several web units (web domains, sites and pages) are described. Emphasis is placed on the properties of social-network measures in order to describe the visibility of a web site and to characterise the structure of a web space. Major developments, such as the Ranking of World universities on the Web and visualisations of web regions, are considered. Finally, there is a discussion about the implications of this discipline on the improvement of web performance and visibility of the university institutions on the Web and its impact on the development of the higher education web-based policies according to open access and e-learning initiatives.

The work presents a proposal of the application of the geostatistical simulation - the Turning Bands method, to acquire knowledge about network throughput. The data base was created on the basis of Multiagent Internet Measurement System MWING. In the research the connections between an agent in Wroclaw and European servers were considered. The preliminary structural analysis of the data which are necessary to use the Turning Bands method was conducted. Next a spatial forecast of the total time of downloading data from Web servers with a two-week time advance was calculated. The analysis of server activity on a particular week day was conducted for a period of a few weeks in selected time intervals.

While many programming languages excel in their ability to execute commands quickly, others embody a greater focus on programmer productivity and clear syntax. In ESRI's GIS software package ArcGIS, Python is now the choice language for many GIS Analysts as an alternative to the more complex ArcObjects library. ArcObjects is written in C#, Visual Basic, Java, or C++, all more difficult languages to learn than Python, but also much faster. In modern web mapping, ArcGIS Python scripts are now making their way onto the server, sometimes at the expense of application performance and stability.I have explored the idea of code performance vs. programmer productivity in the context of ArcGIS Server by writing several web-based geoprocessing services in both Python and C# ArcObjects. The goal was to identify the classes of tools which are best developed using one technology or the other, either based on performance or ease of development. From the outset, I made the assumption that under equal circumstances, it is easier to develop a service in Python, but that C# will always execute faster.The different geoprocessing services were divided into three categories: raster-based, vector-based, and server utilities. The services had different inputs and outputs ranging from text to polygons to zip files. Multi-Mechanize web performance and load testing framework was used to automate requests and make testing repeatable. Multi-Mechanize is an open source testing framework written in python which assisted in replaying requests, logging responses, and compiling statistics. Using this framework, I was able to make an assessment of the exact types of geoprocessing services which should be built using python, and which should be avoided.

Nowadays the e-Commerce and e-Business paradigm have prevailed all over the world, definitely changing the way of running businesses and shaping the new business environment. The Web-based business solutions have produced a novel type of a consumer, the e-Consumer, and a specific way of doing business on the Net, based upon the human – computer – Web interaction. The online behavior of the e-Consumer has proven to be a key not only to a successfulness of a given e-Commerce site, but also a key to obtain a significant knowledge about customer’s habits, needs, expectations, etc. Nonetheless, the e-Consumer online behavior can also be used as a basis for obtaining the basic Web performance metrics, necessary for assuring a relevant level of Quality of Service (QoS) through capacity planning. The paper aims to highlight some of the most important aspects of modeling e-Consumer online behavior regarding Web performance. We propose a novel approach to modeling online behavior, based on usage of Deterministic and Stochastic Petri Nets (DSPNs).

The growing demand for quality and performance has become a discriminating factor in the field of web applications. Advances in software technology and project management have increased the amount of web applications which are developed under these aspects. At the same time the research in web performance simulation and prediction is focusing on complex modeling methods and techniques rather than providing easy to use tools to test and ensure the performance requirements of today’s and tomorrow’s web applications. As a tool for these tasks, the automatic web performance simulation and prediction system (AWPS) will be described. The AWPS is capable of (a) automatically creating an online web performance simulation and (b) conducting trend analysis of the system under test.

Web log analysis can be helpful in gaining information about the usability of the web site, web performance, for marketing purposes, or for development of business intelligence tools in e-commerce systems. User segmentation is one of the problems solved in marketing and e-commerce sphere. Various software was developed to support web analysis. However, most of them provide only information through the tools based on statistics. User behavior and interaction with the web site is usually presented by measurement of click through rates, or by identification and sometimes visualization of popular paths only. User segmentation for further analysis (e.g. campaign analysis in marketing, web recommendation, web usage optimization) is usually allowed with the manual selection (often with variable setting). In this paper is presented the automatic user segmentation (clustering) based on the similar user's behavior on the web site. The user's behavior and behavioral patterns are extracted using process mining techniques; further user segmentation is provided by finding communities with similar behavior through two-step hierarchical clustering.

Data mining (DM) is the key process in knowledge discovery. Many theoretical and practical DM applications can be found in science and engineering. However there are still such areas where data mining techniques are still at early stage of development and application. In particular, an unsatisfactory progress is observed in DM applications in the analysis of Internet and Web performance issues. This chapter gives the background of network performance measurement and presents our approaches, namely Internet Performance Mining and Web Performance Mining as the ways of DM application to Internet and Web performance issues. The authors present real-life examples of the analysis where explored data sets were collected with the aid of two network measurement systems WING and MWING developed at our laboratory.

This paper presents analysis of Web performance data collected by MWING active Internet measurement system. MWING collects traces of HTTP transactions' performance and allows for estimation of important quality of Web user's experience metrics such as resource download time or goodput (application level throughput). We compare data collected by distributed measurement agents and analyze the responsiveness of common Web servers for different clients. We examine two time series prediction techniques: ARIMA models and pattern matching predictor.

From the user point of view, security has shown to be ambivalent; on one hand it is felt necessary, but on the other hand it is felt disturbing. Authentication solutions, for instance, are designed to keep undesired and unauthorized users out; however, allowed users need to spend some effort and waiting time when logging into a system. The question remains to which extent this effort is perceived as positive as security is increased, or negative through the waiting time spent in the process. Excessive waiting times imply the risk of user churn. As there is a lack of such studies, this paper investigates user perception (Quality of Experience, QoE) of the response times (Quality of Service, QoS) of a web authentication procedure, in particular a login to a community web page. Comparing the results to well-known user perception of web performance, we show that the users perceive logins in a similar way as standard web pages, which means that similar limits on user patience apply. The derived QoE-QoS relationship, an exponential function, serves then as the basis for assessing the performance of authentication algorithms in the domain of user acceptability. Pappret visar att fördröjningar vid webb-inloggning upplevs på nästan samma sätt än fördröjningar av "vanliga" webbsidor -- det är samma exponentiella sammanhang mellan upplevd tjänstekvalité (QoE) och svarstiden (QoS) som kan användas. upplevd tjänstekvalité, QoE, svarstid, QoS, inloggning, web

The performance prediction is a key part of the modern network traffic engineering. In this paper we present the application of nonlinear autoregressive modeling to the prediction of goodput level in web transactions. We propose the two-stage approach, with clustering step on historical data, prior to classification, to determine the most appropriate traffic intensity levels. Our study is based on the data collected by the MWING system, an ensemble of web performance measurement agents, and cover over a year of continuous observations of a group of HTTP servers.

To improve the quality and reliability of Web applications, the paper presents a quick testing model of Web performance based on testing flow. It includes successful requests rate, the user behavior modeling, the quick testing flow and treatment for errors in experimental data. It also contributes a testing method of the testing model combining with LoadRunner testing tool, this provides an effective solution to Web performance testing quickly. Finally, the experimental results between current test model and the improved test model are analyzed in detail, which proves that the improved model is more practical and effective.

In this paper we present the application of the Transform Regression (TR) algorithm to predict file transfer time. We consider the files that are transferred from Web servers using HTTP protocol. We identify parameters having impact on file transfer time and use them as inputs to transform regression. We compare the results of TR-based prediction with the moving averages methods.

With the growing use of dynamic content by mobile content distribution systems, how to manage dynamically changing files has become an important issue, since the cached replicas on different mobile sites must be updated if the originals have been changed. Therefore, this paper proposes a priority based selection method to enhance the efficient utilization of network resource and support the client mobility for Mobile Contents Delivery Network (M-CDN). On one hand, a consistency priority is calculated by analyzing the characteristics of mobile surrogates. If a given content which has been changed on its original node, only the replicas with the high consistency priority instead of all replicas are updated. On the other hand, an update priority is also proposed. If a replica has been selected for update, the latest version will be sent from the site decided by the update consistency. Simulation results show that the proposed new approach outperforms other conventional methods.

Companies and web sites store information about the user behaviour which is not shared with other people. A tool for public and free sharing of this information has been developed. This tool is installed as a Firefox Add-on and it stores, in a central system (Web Mining Service), the gathered information in the browser of the client.

In this paper, based on the research of the existing traditional software testing, according to the characteristics of Web service itself, a Web system testing model is given and achieved. And the experimental results show that the model is effective and feasible, and this model is simple and easy to achieve, which has a certain practical significance.

Web Mining is the application of data mining to discover useful knowledge from the Web. Web mining focuses now on four main research directions related to the categories of Web data: Web content mining, Web usage mining, Web structure mining, and Web user profile mining. Web content mining discovers what Web pages are about and reveals new knowledge from them. Web usage mining concerns the identification of patterns in user navigation through Web pages and is performed for the reasons of service personalization, system improvement, and usage characterization. Web structure mining investigates how the Web documents are structured, and discovers the model underlying the link structures of WWW. Web user profile mining discovers user’s profiles based on users’ behavior on the Web. We present the application of data mining in Web performance analysis. We call our approach Web performance mining (WPM). It has been defined to characterize the performance from the perspective of Web clients in the sense of the data transfer throughput in Web transactions. WPM adds a new dimension in Web mining research that focuses on using data mining techniques to analyze Web performance measurements to find interesting patterns in order to support decision-making in the use of Web, for example, to predict future state of good or poor performance in the access to particular Web servers. WPM is based on the measurements which are planned and performed using specific measurement tools and platforms. We developed the multi-agent distributed system MWING to support required active measurements.

In this paper, we proposed web performance enhancement of the e-business system using the SSD (Solid State Drive). In the e-business system, transactions must be strictly ordered, should occur at a fixed point time, and not be lost. We approached to solve e-business problem by using software and hardware techniques such as SSD. X-internet has advantage of C/S and Web applications. Since X-internet technique enables to reduce the loading time of web pages, X-internet is useful to enhance performance of the e-business system. SSD is a storage device that uses DRAM or NAND Flash as primary storage media. Since the SSD stores and accesses data directly to memory chips, which results in storage speeds far greater than conventional magnetic storage devices (HDD). Therefore x-internet techniques can be used to solve confused browsing and poor UI, and SSD can be used to solve I/O bottleneck of the e-business system.

Prefetching is an interesting technique for improving web performance by reducing the user-perceived latency when surfing the web. Nevertheless, due to its speculative nature, prefetching can increase the network traffic and the server load. This could negatively affect the overall system performance and decrease the quality of service. To minimize and maintain under control these adverse effects, in this paper we propose an intelligent prefetching mechanism that dynamically adjusts the aggressiveness of the prefetching algorithm at the server side. To this end, we also propose a traffic estimation model that permits to accurately calculate, in the server side, the extra load and traffic generated by the prefetching. The performance evaluation study shows that our proposal effectively regulates the adverse effects of web prefetching without affecting its benefits.

Today the performance of web systems is getting ever more important, as the number of users and competitors is still increasing. Therefore performance analysis tools gain importance too. There are currently several tools on the market that ensure and test for adequate performance. There are a number of simulation tools and monitoring tools, but only few that automatise and combine both approaches.This paper outlines a system that is capable of (a) automatically creating a web performance simulation and (b) conducting trend analysis of the system under test (SUT). The system requires input information like monitoring points and static-information about the SUT. Based on this information a simulation model of the system is generated. Then the simulation model is refined stepwise e.g. by adding or removing connections between the model components or adjusting the parameters until the aimed accuracy is achieved. By using this simulation model the prediction module creates an analysis of the SUT, and thereby provides as much information as possible about the current state of the system and potential trends. This predictive information can be used for pro-active server tuning or other performance optimisations.The special focus of this work is on the adjustment and prediction parts of the system described here. For all the other parts existing tools and techniques will be used wherever possible.

In this work the authors show how the asymmetry characteristics of wireless network strongly affects the Web response time of user. It is found that uplink delay degrades Web performance because additional bytes overhead and time overhead are introduced. On the other hand, the study also shows that reduction of Web response time is possible through protocol optimization on different layer of protocol stack.

Accurate performance metric models are the key to Web capacity planning related problems. Due to the complexity of Web systems, analytical modeling without integrating the performance testing process is not enough to get accurate metric models. To integrate performance testing and analytical modeling in a systematic way, a Web performance modeling process is presented based on the methodology of learning from data. The process divides the modeling activity into several phases: constructing models and hypothetical conditions, deriving test cases, estimating parameters and validating models, etc. The scalability of a real Web community system (www.igroot.com) is studied by using the proposed process. The error of estimated saturation point is within 1 percent, the error of estimated lower bound of buckle point is within 5 percent. At last, a HTTP processing bottleneck at the architecture level is identified by correlating the model with the threads data of the Web server.

Performance is a key feature in many systems nowadays. There are several tools on the market that ensure and test for adequate performance. They, can be divided into simulation tools and monitoring tools. But only a few automatise and combine both approaches. This paper describes a system capable of automatically creating a web performance simulation and conducting trend analysis of the system under test (SUT). To achieve this the system requires input information, like Monitoring Points and Static-Information about the SUT. The system monitors and analyses the SUT and based on this information generates a simulation model of the system. The simulation model is refinded stepwise e.g. by adding or removing connections between the model components or adjusting the parameters until the aimed accuracy is achieved. With the help of the simulation model a prediction module creates an analysis of the SUT, and thereby can give as much information about the current state of the system and potential trends as possible. This predictive information can be used for pro-active server tuning or other performance optimisations. The focus of my PhD thesis is on the adjustment and prediction part of the system described here. For all other parts, already existing tools and techniques will be used where possible. This initial paper outlines the complete system.

With the development whereby mobile content distribution systems can manage dynamically changing files, an important issue to be resolved is consistency management, which means the cached replicas on different mobile sites must be updated if the originals change. This paper is to design an integrated consistency-control algorithm for mobile contents delivery network (M-CDN) to enhance the efficient utilization of network resource and support the client mobility. Firstly, by carrying out an analysis of mobile surrogates' characteristics, for a given content which has been changed on its original node, only a limited number of its replicas instead of all replicas are updated. Secondly, if a replica has been selected for update, the latest version will be sent from an algorithm-decided site instead of from its original server. Simulation results show that the proposal outperforms other conventional methods.

This paper presents a study of the application of data mining algorithms to the prediction of TCP throughput in HTTP transactions. We are using data mining models built on the basis of historic measurements of network performance gathered using WING system. These measurements reflect Web performance as experienced by the end-users located in Wroclaw, Poland. Data mining models are created using the algorithms available in Microsoft SQL Server 2005and IBM Intelligent Minertools. Our results show that our data mining based TCP throughput prediction returns accurate results. The application of our method in building of so-called "best performance hit" operation mode of the search engines is proposed.

The enormous potential of locality based strategies like caching and prefetching to improve Web performance motivates us to propose a novel global framework for performance evaluation in scenarios where different parts of the Web architecture interact. Unlike existing proposals our approach is a fast and flexible tool that allows to represent faithfully the behaviour of each element of the architecture in order to study, reproduce, evaluate and design Web strategies to decrease the userpsilas perceived latency when surfing the Web.

This paper presents a load balancing solution for a cluster of Web servers that considers QoS in order to differentiate the service provided to the clients when the Web system is receiving a high volume of traffic. Adaptability is a needed feature to suit the balancing parameters to the fluctuations of the Web traffic. We monitor the incoming workload to the system and tune the parameters more frequently when the bursty arrivals trend increases. Throughput prediction enables to estimate the maximum utilization allowed in the servers for each service, and then to set a maximum number of accepted requests to set a limit in the utilization of the servers and to avoid a possible congestion situation, while guaranteeing the QoS. Classical dispatching policies are introduced and compared to analyse the behaviour of the algorithm. HTTP static and dynamic requests are considered in the workload.

Reduction of World Wide Web user perceived latency has assumed importance in the wake of the fast development of Internet services and a huge amount of network traffic and hence adaptation of Web pages to the needs of a specific user is today's trend of Web technologies. Although Web performance can be improved by caching, the benefit of using it is rather limited owing to filling the cache with documents without any prior knowledge. Web prefetching becomes an attractive solution wherein forthcoming page accesses of a client are predicted, based on access log information. This paper proposes a Zipfpsilas Law based novel approach for the determination of next page likely to be accessed by specific client.

Age-related differences in web navigation were examined in two experiments, one (N=44) examining search performance and memory for site structure and a second (N=61) investigating search and memory for site content. Participants searched a large website under one of two experimental conditions; a control condition with only basic navigation aids or a condition with more enhanced navigation aids available. In Experiment 1, the enhanced navigation aids were breadcrumbs and a dynamic side-tree (BC-DST), whereas in Experiment 2 only the dynamic side-tree (DST) was available. In Experiment 1, search was more efficient for both age groups in the BC-DST condition. Age differences in search time and recall of the website’s structure could be eliminated by controlling for working memory and processing speed. In Experiment 2, navigation speed improved with practice and there were larger improvements observed for older adults. Navigation tools did not improve performance in either age group. Younger adults had better memory for content than their older counterparts. Contrary to the hypothesis that older adults are more frequently ‘lost’ in their search, age was not associated with the number of pages or repeat pages visited. Future research should focus on alleviating cognitive demands of search in more representative groups of older users.

Peer-to-peer (P2P) applications consume most of Internet bandwidth and cause access network congestion. They severely deteriorate the performance of the traditional Internet applications. Previous researches have studied the mechanisms to improve the aggregate throughput of P2P traffic, but there doesn't exist any research to predict and quantify its impact. In this work, we develop a performance modeling system of P2P file sharing traffic and traditional Internet traffic (Web traffic) to quantify the impact of P2P file sharing traffic on Web traffic in the congested access network. We answer the following questions from a user's point of view: How many P2P concurrent connections and what proportion between P2P traffic and Web traffic will guarantee certain Web performance. The simulation results demonstrate that our model is accurate and efficient.

We present a Web system architecture using ontologies to improve the behavior of the system from the performance viewpoint. Since Web system performance indexes depend on state and parameter values on runtime period, the proposed system configuration will change during this period. In order to perform this change, the Web system is monitorized and gathered information stored into a knowledge base. We also model the performance of the different Web system elements intervening in the configuration using the knowledge base expressed by means of ontologies. An example of the use of this ontology in cache tier is also presented. We propose the use of performance reasoners to change the configuration during runtime period based on the information supplied from the knowledge base.

Accurate web application performance testing relies on the use of loading tests based on a realistic client behaviour load model. Unfortunately developing such load models and associated test plans and scripts is tedious and error-prone with most existing web performance testing tools providing limited client load modelling capabilities. We describe a new approach and toolset that we have developed, MaramaMTE+, which improves the ability to model realistic web client load behaviour, automatically generates complex web application testing plans and scripts, and integrates load behaviour modelling with a generic performance engineering tool. MaramaMTE+ uses a stochastic form chart as its client loading model. A 3rd party web crawler application extracts structural information from a target web site, aggregating the collected data into a crawler database that is then used for form chart model generation. The performance engineer then augments this synthesized form chart with client loading probabilities. Realistic web loading tests for a 3rd party web load testing tool are then automatically generated from this resultant stochastic form chart client load model. We describe the development of our MaramaMTE+ environment, example usage of the tool, and compare and contrast the results obtained from our generated performance load tests against hand-built 3rd party tool load tests

Unexpected surges in request traffic (e.g., flash crowds) can exercise server-side resources such as access bandwidth, CPU, and disks in unanticipated ways. Administrators today do not have the requisite tools to fully understand the effect that flash crowds can have on server-side resources. As a result, most Web-servers today rely on significant over-provisioning, strict admission control, or alternatively use potentially expensive solutions like CDNs to provide high availability under load. A more fine-grained understanding of the performance of servers under emulated but controlled flash crowd like conditions can guide administrators to make more efficient provisioning and resource management decisions.We present the initial design of Mini-Flash Crowds (MFC) -- a light-weight wide-area profiling service that reveals resource bottlenecks in a Web-server infrastructure, including access bandwidth, processing resources, and back-end data management. The MFC approach is based on a set of controlled measurements in which an increasing number of distributed clients make synchronized requests to exercise specific resources of a remote server. Using a number of wide-area experiments and controlled lab tests, we show that our approach can faithfully track the impact of request loads on different server resources. Our approach is non-intrusive and thus we can use it to actively probe numerous live Web servers. We present the results from a preliminary measurement study of resource provisioning on public Web servers.

Despite the wide and intensive research efforts focused on Web prediction and prefetching techniques aimed to reduce user's perceived latency, few attempts to implement and use them in real environments have been done, mainly due to their complexity and supposed limitations that low user available bandwidths imposed few years ago. Nevertheless, current user bandwidths open a new scenario for prefetching that becomes again an interesting option to improve web performance. This paper presents Delfos, a framework to perform web predictions and prefetching on a real environment that tries to cover the existing gap between research and praxis. Delfos is integrated in the web architecture without modifying the standard HTTP 1.1 protocol, and acts inserting predictions in the web server side, while prefetchs are carried out by the client. In addition, it can be also used as a flexible framework to evaluate and compare existing prefetching techniques and algorithms and to assist in the design of new ones because it provides detailed statistics reports.

In this paper we present results from a measurement campaign evaluating the web performance of different mobile Internet access technologies with and without caching. We measured in GPRS, EDGE, UMTS and HSDPA radio access networks of a mobile operator in Austria. The operator has implemented a web-proxy for performance improvement. To evaluate the impact on the different technologies we ran the setup with and without the web-proxy. We defined the download time of web pages as the performance index for our work. We developed a tool to capture this parameter and derived the index for a set of web sites frequently used by the customers in the network. Beginning with an over-all comparison we will present detailed analysis for special cases. We evaluated the impact of the proxy system to different site configurations using the performance index. Finally we compared the performance numbers for HTTP with results collected from FTP transfers.

This chapter describes the acceleration of web performance with ISA 2006 caching capabilities. Forward Web caching has the advantage of creating a faster access for internal users as they retrieve the Web objects over a fast Local Area Network (LAN) connection. Forward caching takes place when a user on a network protected by the ISA 2006 firewall makes a request for Web content. The primary bottom line benefit of ISA 2006's forward caching is cost savings realized by reduced bandwidth usage on the Internet connection. Reverse caching reduces traffic on the internal network and speeds access for external users when the company hosts its own Web sites. Reverse caching reduces bandwidth usage on the internal network when cached content is served directly from the ISA 2006 machine. It is observed that when a request for an object on the company Web server comes in from a user over the Internet, the ISA firewall checks its cache for the object. It is found that the content download feature is used to schedule ISA 2006 to download new content from the Internet at predefined times so that, when Web Proxy clients request those objects, updated versions will be in the cache.

This paper presents a complete solution of a scalable content-aware load balancing solution for cluster-based Web information systems that includes service differentiation and quality of service. The architecture is based on a farm of Web servers managed by a content-aware load balancer that differentiates the service required by the workload and implements a control admission under extreme overload. The results show that the burstiness in the workload has to be controlled on the Web servers in order to obtain the best response in terms of performance and overhead

Proxy caching is a commonly accepted methodology used to reduce Internet traffic, decrease back-end related user delays, and generally improve Web performance. Nevertheless, it seems to be inadequate when it comes to caching dynamic Web pages, due to their low degree of reusability and strong dependency on the back-end site infrastructure. In order to overcome these drawbacks we present DOMProxy (dynamic objects manager proxy), a new proxy architecture that enables front-end Web caching for dynamically generated Web pages. Instead of caching Web pages produced dynamically or their fragments, the proposed architecture provides methods for "caching" their generation process. This is achieved by caching both the applications and the back-end data content that are used to create dynamic Web pages. We discuss the new challenges and issues raised by this approach, with special emphasis on the caching and manipulation of different types of Web objects at a front-end caching point.

In this paper, we propose data mining analysis of Internet performance data. The research is aimed at characterization of Web performance as seen by the end-users. Our strategy involves discovering knowledge that characterizes Web performance perceived by end-users and then making use of this knowledge to guide users in further Internet usage. Our two-phase data mining decision model can predict RTT and throughput behavior when downloading Web objects. The data is collected by the Wing measurement infrastructure. Real measurements data is used in model evaluation. We show that the number of correct classifications is high (76%-93%).

Web user behavior has widely changed over the last years. To perform precise and up-to-date Web user behavior characterization is important to carry out representative Web performance studies. In this sense, it is valuable to capture detailed information about the user's experience, which permits to perform a fine grain characterization. Two main types of tools are distinguishable: complex commercial software tools like workload generators and academic tools. The latter mainly concentrate on the development of windows applications which gather Web events (e.g., browser events) or tools modifying a part of the web browser rode. In this paper, we present CARENA, a client-side browser-embedded tool to capture and replay user navigation sessions. Like some commercial software packages our tool captures information about the user session, which can be used later to replay or mimic the gathered user navigation. Nevertheless, unlike these software packages, our tool emulates the original user think times since these times are important to obtain precise and reliable performance results. Among the main features of CARENA are: multiplatform, open source, lightweight, standards based, easily installable and usable, programmed in JavaScript and XUL.

Quality of service(QoS) is so important for content Providers that they constantly face the challenges of adapting their web servers to support rapid growth and customer’s demand for more reliable and differentiated services. A web QoS measurement and evaluation system(WQMES) was designed and implemented based on in-depth research on the key techniques of web QoS. The prototype implementation of WQMES and the performance evaluation criteria based on performance aggregation are carefully introduced. Our contribution is presenting a single and quantitative result combining several web performance metrics. Experiment results indicate the scalable WQMES can do the real-time detection on web QoS from the end user’s perspective. The performance aggregation approach is a bran-new idea, and of a definite practicability.

Microsoft's marketing emphasis with ISA Server 2004 is on its firewall and virtual private networking (VPN) gateway functionality. It also provides companies with a viable Web caching solution that can save hundreds or thousands of dollars that would have to be spent for a separate caching product if one implements competing firewall products that do not include caching functionality. The ISA Server 2004 caching capabilities enhance network's productivity by providing acceleration of access to external Websites by internal users, via the forward caching feature. It can also accelerate the access of external users who connect to internal Websites, via the reverse caching feature. ISA Server 2004 uses the cache array routing protocol (CARP) for communications among Web caching servers. CARP is a hash-based protocol that allows multiple caching proxies to be arrayed as a single logical cache and uses a hash function to ascertain which cache a request should be sent to.

A key aspect in managing resources for customer sites is to predict and assess the load associated with a site in order to figure out how best to allocate resources for the site over time and to efficiently schedule tasks. The cost associated with the site and return on investment are also key parameters. This paper describes work we have done in developing tools for answering these critical questions. The tools use both analytical models and discrete event simulations to predict performance and analyze costs needed for handling a customer workload while satisfying the service level objectives. These tools provide capacity and load planning, performance simulation, and cost and financial analyses. Our tools have been used successfully by several major customers, and those experiences have shaped how the tools have evolved over time.

In this paper, the Multi Agent System (MAS) Architecture, GPS, GIS, and Wireless Communication technologies were discussed. New application architecture of complex vehicle Location Based Service (LBS), and Navigation and Intelligent Transportation systems based on the MAS architecture were proposed. The implementation and Web performance of this methodology on the urban garbage trucks management in the Shanghai Putuo district were introduced. The final system testing results were evaluated. And the future potentials of the MAS based approach to solve complex urban management problems and monitoring information network systems were prospected.

New connection-oriented networking technologies can provide quality-of-service guaranteed network connectivity required by some web-based applications. In this paper, we present a solution for leveraging these new networking technologies without requiring modifications to existing web server and client software. We implemented a CGI-based software package called WebFT. We demonstrate how it can be used along with existing web servers and clients to enable fast web page downloads across a wide-area experimental connection- oriented network called CHEETAH, which we deployed as part of an NSF-sponsored project. WebFT is integrated with the CHEETAH end-host software modules to provide deterministic data transfer services transparently to users. The "add-on" design of the CHEETAH network, whereby connection-oriented services are provided in addition to the connectionless service offered by the Internet, allows the WebFT package to provide normal web server access to non- CHEETAH clients through the Internet while simultaneously serving CHEETAH clients on dedicated circuits. The experiments conducted on the CHEETAH testbed show that WebFT can achieve low-variance end-to-end transfer delays at different circuit rates and low transfer delays when high-speed circuits are possible. 1. Introduction

Web application access on mobile wireless PDAs is becoming increasingly popular. However, web browsing on these systems can be quite slow. An alternative approach is handheld thin-client computing, in which the web browser and associated application logic run on a server, which then sends simple screen updates to the PDA for display. To assess the viability of this thin-client approach, we compare the web browsing performance of thin clients against fat clients that run the web browser locally on a PDA. Our results show that thin clients can provide better web browsing performance compared to fat clients, both in terms of speed and ability to correctly display web content. Surprisingly, thin clients are faster even when having to send more data over the network. We characterize and analyze different design choices in various thin-client systems and explain why these approaches can yield superior web browsing performance on mobile wireless PDAs.

This chapter discusses performance issues that arise for Web content delivery in wireless local area network (WLAN) environments. Two experiments are conducted using IEEE 802.11 b WLAN technology, The first experiment considers client-side Web browsing performance for a mobile user with a wireless PDA in an infrastructure-based WLAN. The second experiment considers server-side performance for a wireless Web server delivering Web content in a wireless ad hoc network environment. The results show that the wireless network bottleneck can lead to inefficient HTTP performance, inefficient TCP performance, packet losses, and network thrashing, depending on the characteristics of the Web traffic generated. Solving these issues is important to improve wireless Web performance.