Real-time content analysis is typically a bottleneck in Web filtering. To accelerate the filtering process, this work presents a simple, but effective early decision algorithm that analyzes only part of the Web content. This algorithm can make the filtering decision, either to block or to pass the Web content, as soon as it is confident with a high probability that the content really belongs to a banned or an allowed category. Experiments show the algorithm needs to examine only around one-fourth of the Web content on average, while the accuracy remains fairly good: 89% for the banned content and 93% for the allowed content. This algorithm can complement other Web filtering approaches, such as URL blocking, to filter the Web content with high accuracy and efficiency. Text classification algorithms in other applications can also follow the principle of early decision to accelerate their applications.

Educational websites contain rich knowledge components on a web page. Detecting student attention on web pages fulfills the recommendation of adequate knowledge components to students based on students' current interests. Previous studies have shown the application of learner attention in intelligent learning systems. This study proposes a methodology to analyze student on-line mouse movement patterns that indicate student attentions. The methodology can be combined with learning systems that implement pedagogical models such as inquiry-based learning and problem-solving learning activities. The feasibility and effectiveness of the proposed methodology have been evaluated by student mouse movements in problem-solving scenarios.

Peer-to-peer (P2P) web caching has been studied recently as it can exploit local caches of peers for web caching without additional infrastructure. However, dynamic join/leave behaviors or local caching strategies of the peers due to their autonomy in a P2P network may limit the performance of P2P web caching. To overcome these limitations, we propose an effective directory-based P2P web caching system under dynamic participation of peers. We introduce the object lifetime in a P2P network considering the lifetimes of both an object in the local cache of a peer and a peer who owns the object, and utilize this object lifetime information for neighbor selection and storage management in the P2P web caching framework. For the neighbor selection, the proposed system utilizes the object lifetimes in selecting accurately a neighbor who would still retain the requested object and still remain in the P2P network. To improve the storage management, the proposed system uses efficiently the entire cache storage of the P2P network in such a way that the object is stored selectively in the local cache of the peer who requested it, considering the object lifetime. The trace-driven simulation results show that the proposed system has higher accuracy and fewer redirection failures than the conventional directory-based P2P web caching system in the feasible P2P network.

Mining traversal patterns on the Internet is one of critical issues for exploring the user access behaviors. In this paper, we propose a new data mining scheme for mining frequent trip traversal patterns on the Internet. First, we define a trip traversal as a historical contiguous sequence of web sites or web pages, which were surfed or visited on an information-providing system by one user. Next, we derive all of the maximal trip traversals by analyzing and filtering these collected trip traversals. For mining the large trip traversals from the maximal trip traversals, we present a data mining scheme integrated with the schemes presented in. Here, the extracted large trip traversals can be thought of as the realistic frequent browsed behaviors for most of users either on a web site or on an information-providing system, such as a proxy server. Finally, we implement and design a data mining system to explore the large trip traversal patterns in order to capture user access patterns to some proxy server.

Recently, layer-4 (L4) switches have been widely used as load balancing front-end routers for Web server clusters. The typical L4 switch attempts to balance load among the servers by estimating load using the load metrics measured in the front-end and/or the servers. However, insufficient load metrics, measurement overhead, and feedback delay often cause misestimate of server load. This may incur significant dynamic load imbalance among the servers particularly when the variation of requested content is high. In this paper, we propose a new content sniffer based load distribution strategy. By sniffing the requests being forwarded to the servers and by extracting load metrics from them, the L4 switch with our strategy more timely and accurately estimates server load without the help of back-end servers. Thus it can properly react to dynamic load imbalance among the servers under various workloads. Our experimental results demonstrate substantial performance improvements over other load balancing strategies used in the typical L4 switch.

A content-based switch makes forwarding decisions (server selections) based on an application layer information and forwards data in the application layer. After making forwarding decisions, existing content-based switches improve their forwarding performance by TCP splicing, which releases them from maintaining TCP endpoints and allows them to forward data by packet forwarding. However, once content-based switches invoke TCP splicing, they are unable to use the application layer information for forwarding decisions. Thus the existing content-based switches cannot perform a handoff of pipelined HTTP transactions, which can greatly reduce client perceived latencies. This paper proposes an asymmetric TCP splicing and a method to perform the handoff of HTTP transactions between servers. Asymmetric TCP splicing allows the content-based switches to use all the application layer information in the TCP data stream from clients to servers, although it allows the switches to forward the TCP data stream from servers to clients by packet forwarding. The proposed handoff method content-based switches support pipelined HTTP transactions in combination with asymmetric TCP splicing. In the proposed method, the content-based switch utilizes the common function of TCP (TCP half-close) to detect the end of the series of responses from the currently selected server, and it changes the forwarding destination after the client finishes receiving a series of responses from the server. Our evaluation validates that the content-based switch which supports pipelined HTTP transactions by our method can reduce client-perceived latencies when there is a large correlation between destinations of any two consecutive requests.

This paper proposes an adaptive transmission scheme for web prefetching in wireless communication systems. The proposed adaptive transmission scheme controls the modulation format and the error control scheme according to the access probability of the web document being transmitted. In the proposed system, the actually requested documents and the documents which have high access probability are transmitted with a reliable transmission format, while the pages whose access probabilities are lower than a certain threshold are transmitted with a bandwidth efficient transmission format. The computer simulation results show that the proposed scheme drastically improves the latency performance.

Caching at the Web proxy server plays an important role in reducing the response time, the network traffic, and the load of Web servers. Many recent studies have proposed and examined the replacement and consistency policies for the proxy cache, which plays a central role in the performance of caching components. For better performance, they exploit various metadata of Web objects, such as the reference count, reference time, and modification time information of past behaviors, to estimate the re-reference likelihood and freshness of the objects. However, all of these known to the authors use the metadata only when the actual object is in the cache. We observed from various proxy traces that about 20-30% of clients' requests incurred only the validity checks of cached objects without transferring actual objects from the proxy server. In this case, only the metadata are necessary at the proxy server. This paper proposes a proxy cache consistency policy that uses the metadata even for absent objects. These include the time information of evicted objects from the cache and those out of the header-only replies from Web servers. Trace-driven simulations with public proxy cache traces show that our policy reduces the response time and the number of connections to Web servers significantly.

We describe an index for estimating the level of interest in Web pages. This "time-based interest" (TBI) index combinates an equation reflecting page accesses and an equation reflecting the decrease in interest over time. These equations work simultaneously by using a parameter that is based on the time since the last access. We experimentally estimated the decrease ratio of the TBI index and evaluated the characteristics of the TBI equation. We found that the index follows Zipf's distribution, indicating that reflects the change in popularity. We also introduce an access-log analysis system called CyberRanking that includes TBI analysis. CyberRanking analyzes the access logs of Web servers and presents the results in 2-D or 3-D graph on a Web browser.

Although there are many multimedia presentation systems on the market, they have some shortcomings and most of them only can work on one single computer, and few of them can work on Web. Thus, in the thesis we develop a network multimedia presentation system to let users easily design the multimedia presentation without restriction on technology or presentation time and place. Our system includes 3 main components: User Interface that includes temporal specification editor, spatial specification editor and multimedia object interface, Presentation Interface and Knowledge Base. There is a dynamic homepage generator in our system and we propose a displaying algorithm based on the Allen's theory, that there exist 13 temporal relationships between two intervals, for synchronizing the media objects.