Scalability is one of the main challenges for kernel-based methods and support vector machines (SVMs). The quadratic demand in memory for storing kernel matrices makes it impossible for training on million-size data. Sophisticated decomposition algorithms have been proposed to efficiently train SVMs using only important examples, which ideally are the final support vectors (SVs). However, the ability of the decomposition method is limited to large-scale applications where the number of SVs is still too large for a computer's capacity. From another perspective, the large number of SVs slows down SVMs in the testing phase, making it impractical for many applications. In this paper, we introduce the integration of a vector combination scheme to simplify the SVM solution into an incremental working set selection for SVM training. The main objective of the integration is to maintain a minimal number of final SVs, bringing a minimum resource demand and faster training time. Consequently, the learning machines are more compact and run faster thanks to the small number of vectors included in their solution. Experimental results on large benchmark datasets shows that the proposed condensed SVMs achieve both training and testing efficiency while maintaining a generalization ability equivalent to that of normal SVMs.

Scalability is one of the main challenges for kernel-based methods and support vector machines (SVMs). The quadratic demand in memory for storing kernel matrices makes it impossible for training on million-size data. Sophisticated decomposition algorithms have been proposed to efficiently train SVMs using only important examples, which ideally are the final support vectors (SVs). However, the ability of the decomposition method is limited to large-scale applications where the number of SVs is still too large for a computer's capacity. From another perspective, the large number of SVs slows down SVMs in the testing phase, making it impractical for many applications. In this paper, we introduce the integration of a vector combination scheme to simplify the SVM solution into an incremental working set selection for SVM training. The main objective of the integration is to maintain a minimal number of final SVs, bringing a minimum resource demand and faster training time. Consequently, the learning machines are more compact and run faster thanks to the small number of vectors included in their solution. Experimental results on large benchmark datasets shows that the proposed condensed SVMs achieve both training and testing efficiency while maintaining a generalization ability equivalent to that of normal SVMs.

IP multicasting is an efficient and scalable way to deliver high quality audio and video streams for group communication. However, widespread deployment of IP multicast is unlikely in near future and it requires substantial infrastructure modifications. For rapid deployment of multicast application, Application Layer Multicast (ALM) is an alternate choice to circumvent this situation, as it supports group communication independent of network layer support. Previous studies on ALM mainly focused on achieving high throughput for bandwidth intensive applications and unfairness of communication quality is still a serious issue in multi-source communication. In this paper, we present an autonomous distributed scheme to construct an ALM tree to improve fairness and overall communication quality among the participants in a multi-source multimedia communication. Extensive simulation and experimental results show that our proposed scheme increases the incoming bitrate of the lowest recipient among all the nodes, as well as improving fairness for all receivers.

In ubiquitous environments, there are a huge number of computer nodes which provide a lot of information services via the Internet. Therefore, a secure communication system is required for ubiquitous environments. In this paper, we propose a new architecture of distributed public key infrastructure, named as Hash-based Distributed Public Key Infrastructure (HDPKI). A HDPKI system achieves a scalable management of public key certificates by using Distributed Hash Table. Additionally, it is easy to use a HDPKI system, because we can register our public key certificates automatically. In this paper, we show scalability of HDPKI through computer simulation. According to these results, the total amount of required messages for certificate management in a HDPKI system is O(log n) where n is the number of nodes.

Recent use of web services has spread over a wide variety of application areas. Users are requested to disclose private information for the use of web services. Such a disclosure is facing the risk of leaking private information. Privacy leakage is becoming one of the serious social problems. For the better use of web services, it is necessary to provide a reliable scheme for privacy management. There are many cases of private information leakage which are caused by user's misuse or unintended disclosure. In order to prevent users from unnecessarily leaking of private information, the users will be asked to clearly define information disclosure criteria. This paper discusses a privacy management architecture to prevent users from unnecessary leaking of private information. We demonstrate a prototype implementation how the proposed architecture effectively works in the initial setting of web services. © 2010 IEEE.

In this paper, we propose an overlay authentication network for active utilization of private information. The proposed network realizes safe registration and usage of users' private information. In addition, the proposed network realizes the distributed management and utilization of users' private information. These features enable services which utilize users' private information. In this paper, we present implementation of the prototype system based on the proposed network, and confirm the effectiveness of the proposed network. © 2010 IEEE.

IP multicasting is an efficient and scalable way to deliver high quality audio and video streams for group communication. However, widespread deployment of IP multicast is unlikely in near future and it requires substantial infrastructure modifications. For rapid deployment of multicast application, Application Layer Multicast (ALM) is an alternate choice to circumvent this situation, as it supports group communication independent of network layer support. Previous studies on ALM mainly focused on achieving high throughput for bandwidth intensive applications and unfairness of communication quality is still a serious issue in multi-source communication. In this paper, we present an autonomous distributed scheme to construct an ALM tree to improve fairness and overall communication quality among the participants in a multi-source multimedia communication. Extensive simulation and experimental results show that our proposed scheme increases the incoming bitrate of the lowest recipient among all the nodes, as well as improving fairness for all receivers.

In ubiquitous environments, there are a huge number of computer nodes which provide a lot of information services via the Internet. Therefore, a secure communication system is required for ubiquitous environments. In this paper, we propose a new architecture of distributed public key infrastructure, named as Hash-based Distributed Public Key Infrastructure (HDPKI). A HDPKI system achieves a scalable management of public key certificates by using Distributed Hash Table. Additionally, it is easy to use a HDPKI system, because we can register our public key certificates automatically. In this paper, we show scalability of HDPKI through computer simulation. According to these results, the total amount of required messages for certificate management in a HDPKI system is O(log n) where n is the number of nodes.

Recent use of web services has spread over a wide variety of application areas. Users are requested to disclose private information for the use of web services. Such a disclosure is facing the risk of leaking private information. Privacy leakage is becoming one of the serious social problems. For the better use of web services, it is necessary to provide a reliable scheme for privacy management. There are many cases of private information leakage which are caused by user's misuse or unintended disclosure. In order to prevent users from unnecessarily leaking of private information, the users will be asked to clearly define information disclosure criteria. This paper discusses a privacy management architecture to prevent users from unnecessary leaking of private information. We demonstrate a prototype implementation how the proposed architecture effectively works in the initial setting of web services. © 2010 IEEE.

In this paper, we propose an overlay authentication network for active utilization of private information. The proposed network realizes safe registration and usage of users' private information. In addition, the proposed network realizes the distributed management and utilization of users' private information. These features enable services which utilize users' private information. In this paper, we present implementation of the prototype system based on the proposed network, and confirm the effectiveness of the proposed network. © 2010 IEEE.

Using Web-based content management systems such as Blog, an end user can easily publish User Generated Content (UGC). Although publishing of UGCs is easy, controlling access to them is a difficult problem for end users. Currently. most of Blog sites offer no access control mechanism, and even when it is available to users, it is not sufficient to control users who do not have an account at the site, not to mention that it cannot control accesses to content hosted by other UGC sites. In this paper, we propose new access control architecture for UGC in which third party entities can offer access control mechanism to users independently of UGC hosting sites. With this architecture, a user can control accesses to his content that might be spread over many different UGC sites, regardless of whether those sites have access control mechanism or not. The key idea to separate access control mechanism from UGC sites is to apply cryptographic access control and we implemented the idea in such a way that it requires no modification to UGC sites and Web browsers. Our prototype implementation shows that the proposed access control architecture can be easily deployed in the current Web-based communication environment and it works quite well With Popular Blog sites.

Using Web-based content management systems such as Blog, an end user can easily publish User Generated Content (UGC). Although publishing of UGCs is easy, controlling access to them is a difficult problem for end users. Currently. most of Blog sites offer no access control mechanism, and even when it is available to users, it is not sufficient to control users who do not have an account at the site, not to mention that it cannot control accesses to content hosted by other UGC sites. In this paper, we propose new access control architecture for UGC in which third party entities can offer access control mechanism to users independently of UGC hosting sites. With this architecture, a user can control accesses to his content that might be spread over many different UGC sites, regardless of whether those sites have access control mechanism or not. The key idea to separate access control mechanism from UGC sites is to apply cryptographic access control and we implemented the idea in such a way that it requires no modification to UGC sites and Web browsers. Our prototype implementation shows that the proposed access control architecture can be easily deployed in the current Web-based communication environment and it works quite well With Popular Blog sites.

This paper studies two problems on compressed strings described in terms of straight line programs (SLPs). One is to compute the length of the longest common substring of two given SLP-compressed strings, and the other is to compute all palindromes of a given SLP-compressed string. In order to solve these problems efficiently (in polynomial time w.r.t. the compressed size) decompression is never feasible, since the decompressed size can be exponentially large. We develop combinatorial algorithms that solve these problems in O(n(4) log n) time with O(n(3)) space, and in O(n(4)) time with O(n(2)) space, respectively, where n is the size of the input SLP-compressed strings. (C) 2008 Elsevier B.V. All rights reserved.

This paper studies two problems on compressed strings described in terms of straight line programs (SLPs). One is to compute the length of the longest common substring of two given SLP-compressed strings, and the other is to compute all palindromes of a given SLP-compressed string. In order to solve these problems efficiently (in polynomial time w.r.t. the compressed size) decompression is never feasible, since the decompressed size can be exponentially large. We develop combinatorial algorithms that solve these problems in O(n(4) log n) time with O(n(3)) space, and in O(n(4)) time with O(n(2)) space, respectively, where n is the size of the input SLP-compressed strings. (C) 2008 Elsevier B.V. All rights reserved.

Recently, P2P networks have been evolving rapidly. Efficient authentication of P2P network nodes remains a difficult task. As described herein, we propose an authentication method called Hash-based Distributed Authentication Method (HDAM), which realizes a decentralized efficient mutual authentication mechanism for each pair of nodes in a P2P network. It performs distributed management of public keys using Web of Trust and a Distributed Hash Table. The scheme markedly reduces both the memory size requirement and the overhead of communication data sent by the nodes. Simulation results show that HDAM can reduce the required memory size by up to 95%. Furthermore, the results show that HDAM is more scalable than the conventional method: the communication overhead of HDAM is O(log p).

Recently, P2P networks have been evolving rapidly. Efficient authentication of P2P network nodes remains a difficult task. As described herein, we propose an authentication method called Hash-based Distributed Authentication Method (HDAM), which realizes a decentralized efficient mutual authentication mechanism for each pair of nodes in a P2P network. It performs distributed management of public keys using Web of Trust and a Distributed Hash Table. The scheme markedly reduces both the memory size requirement and the overhead of communication data sent by the nodes. Simulation results show that HDAM can reduce the required memory size by up to 95%. Furthermore, the results show that HDAM is more scalable than the conventional method: the communication overhead of HDAM is O(log p).

Due to the dynamic topology and non infrastructure, network participants cooperate with their neighbors to route packets. The lack of centralized services allows mobile ad hoc networks to be easily and swiftly deployed, but make it difficult to check others' identities on the other hand. Cryptographic tools have been introduced to secure group communications, such as Private and Public Key Infrastructure. The autonomous and distributed nature of mobile ad hoc network demands a decentralized authentication service, where Public Key Infrastructure is considered a better solution. Public Key Infrastructure can ensure both confidentiality and authenticity, but it is impractical to provide an online trusted third party as Certificate Authority (CA) for mobile ad hoc network. In this paper, we proposed a new key management protocol which utilizes certificate graphs and distributed Certificate Authorities. Certificate graph maintained by each user represents the trust among his neighbors, then the maximum clique of certificate graph is selected to be CAs. Based on the assumption that initial certificate graph building is secure [11], good users have more friends while bad ones have less, thus a reliable group can be constructed. The most trustful subset of these good users - the maximum clique - is elected as the governor of this group, which takes the responsibility of certificate authentication.

Recent trends in the telecommunication industry have been moving toward the development of ubiquitous information systems, where the provision of a plethora of advanced multimedia services should be possible, regardless of time and space limitations. An efficient and seamless delivery of multimedia services over various types of wireless networks is still a challenging task. The underlying difficulty consists of the disparity in the bandwidth availability over each network type. Indeed, the fundamental challenge upon a handoff phenomenon in a heterogeneous wireless network consists of an efficient probing of the bandwidth availability of the new network, followed by a prompt adjustment of the data delivery rate. This paper presents a cross-layer approach that involves five layers, namely, physical, data link, application, network, and transport layers. The three former layers are used to anticipate the handoff occurrence and to locate the new point of attachment to the network. Based on their feedback, the transport layer is used then to probe the resources of the new network using low-priority dummy packets. Being the most widely used protocol for multimedia delivery, this paper addresses multimedia applications based on the Transmission Control Protocol (TCP) and the Real-time Transport Protocol (RTP). The design of the whole cross-layer architecture is discussed, and enhancements to the two protocols are proposed. The performance of the enhanced TCP and the RTP are evaluated and compared against existing schemes through extensive simulations. The obtained results are encouraging and promising for the delivery of multimedia services in heterogeneous wireless networks.

Recent trends in the telecommunication industry have been moving toward the development of ubiquitous information systems, where the provision of a plethora of advanced multimedia services should be possible, regardless of time and space limitations. An efficient and seamless delivery of multimedia services over various types of wireless networks is still a challenging task. The underlying difficulty consists of the disparity in the bandwidth availability over each network type. Indeed, the fundamental challenge upon a handoff phenomenon in a heterogeneous wireless network consists of an efficient probing of the bandwidth availability of the new network, followed by a prompt adjustment of the data delivery rate. This paper presents a cross-layer approach that involves five layers, namely, physical, data link, application, network, and transport layers. The three former layers are used to anticipate the handoff occurrence and to locate the new point of attachment to the network. Based on their feedback, the transport layer is used then to probe the resources of the new network using low-priority dummy packets. Being the most widely used protocol for multimedia delivery, this paper addresses multimedia applications based on the Transmission Control Protocol (TCP) and the Real-time Transport Protocol (RTP). The design of the whole cross-layer architecture is discussed, and enhancements to the two protocols are proposed. The performance of the enhanced TCP and the RTP are evaluated and compared against existing schemes through extensive simulations. The obtained results are encouraging and promising for the delivery of multimedia services in heterogeneous wireless networks.

We propose iSVM - an incremental algorithm that achieves high speed in training support vector machines (SVMs) on large datasets. In the common decomposition framework, iSVM starts with a minimum working set (WS), and then iteratively selects one training example to update the WS in each optimization loop. iSVM employs a two-stage strategy in processing the training data. In the first stage, the most prominent vector among randomly sampled data is added to the WS. This stage results in an approximate SVM solution. The second stage uses temporal solutions to scan through the whole training data once again to find the remaining support vectors (SVs). We show that iSVM is especially efficient for training SVMs on applications where data size is much larger than number of SVs. On the KDD-CUP 1999 network intrusion detection dataset with nearly five millions training examples, iSVM takes less than one hour to train an SVM with 94% testing accuracy, compared to seven hours with LibSVM - one of the state-of-the-art SVM implementations. We also provide analysis and experimental comparisons between iSVM and the related algorithms. ©2008 IEEE.

Hoeffding Tree Algorithm is known as a method to induce decision trees from a data stream. Treatment of numeric attribute on Hoeffding Tree Algorithm has been discussed for stationary input. It has riot yet investigated, however, for non-stationary input where the effect of concept drift is apparent. This paper identifies three major approaches to handle numeric values, Exhaustive Method, Gaussian Approximation, and Discretizaion Method, and through experiment shows the best suited modeling of numeric attributes for Hoeffding Tree Algorithm. This paper also experimentaly compares tire performance of two known methods for concept drift detection, Hoeffding Bound Based Method and Accuracy Based Method.

We propose iSVM - an incremental algorithm that achieves high speed in training support vector machines (SVMs) on large datasets. In the common decomposition framework, iSVM starts with a minimum working set (WS), and then iteratively selects one training example to update the WS in each optimization loop. iSVM employs a two-stage strategy in processing the training data. In the first stage, the most prominent vector among randomly sampled data is added to the WS. This stage results in an approximate SVM solution. The second stage uses temporal solutions to scan through the whole training data once again to find the remaining support vectors (SVs). We show that iSVM is especially efficient for training SVMs on applications where data size is much larger than number of SVs. On the KDD-CUP 1999 network intrusion detection dataset with nearly five millions training examples, iSVM takes less than one hour to train an SVM with 94% testing accuracy, compared to seven hours with LibSVM - one of the state-of-the-art SVM implementations. We also provide analysis and experimental comparisons between iSVM and the related algorithms. ©2008 IEEE.

Hoeffding Tree Algorithm is known as a method to induce decision trees from a data stream. Treatment of numeric attribute on Hoeffding Tree Algorithm has been discussed for stationary input. It has riot yet investigated, however, for non-stationary input where the effect of concept drift is apparent. This paper identifies three major approaches to handle numeric values, Exhaustive Method, Gaussian Approximation, and Discretizaion Method, and through experiment shows the best suited modeling of numeric attributes for Hoeffding Tree Algorithm. This paper also experimentaly compares tire performance of two known methods for concept drift detection, Hoeffding Bound Based Method and Accuracy Based Method.

There are numerous research challenges that need to be addressed until a wide deployment of vehicular ad hoc networks (VANETs) becomes possible. One of the critical issues consists of the design of scalable routing algorithms that are robust to frequent path disruptions caused by vehicles' mobility. This paper argues the use of information on vehicles' movement information (e.g., position, direction, speed, and digital mapping of roads) to predict a possible link-breakage event prior to its occurrence. Vehicles are grouped according to their velocity vectors. This kind of grouping ensures that vehicles, belonging to the same group, are more likely to establish stable single and multihop paths as they are moving together. Setting up routes that involve only vehicles from the same group guarantees a high level of stable communication in VANETs. The scheme presented in this paper also reduces the overall traffic in highly mobile VANET networks. The frequency of flood requests is reduced by elongating the link duration of the selected paths. To prevent broadcast storms that may be intrigued during path discovery operation, another scheme is also introduced. The basic concept behind the proposed scheme is to broadcast only specific and well-defined packets, referred to as "best packets" in this paper. The performance of the scheme is evaluated through computer simulations. Simulation results indicate the benefits of the proposed routing strategy in terms of increasing link duration, reducing the number of link-breakage events and increasing the end-to-end throughput.

There are numerous research challenges that need to be addressed until a wide deployment of vehicular ad hoc networks (VANETs) becomes possible. One of the critical issues consists of the design of scalable routing algorithms that are robust to frequent path disruptions caused by vehicles' mobility. This paper argues the use of information on vehicles' movement information (e.g., position, direction, speed, and digital mapping of roads) to predict a possible link-breakage event prior to its occurrence. Vehicles are grouped according to their velocity vectors. This kind of grouping ensures that vehicles, belonging to the same group, are more likely to establish stable single and multihop paths as they are moving together. Setting up routes that involve only vehicles from the same group guarantees a high level of stable communication in VANETs. The scheme presented in this paper also reduces the overall traffic in highly mobile VANET networks. The frequency of flood requests is reduced by elongating the link duration of the selected paths. To prevent broadcast storms that may be intrigued during path discovery operation, another scheme is also introduced. The basic concept behind the proposed scheme is to broadcast only specific and well-defined packets, referred to as "best packets" in this paper. The performance of the scheme is evaluated through computer simulations. Simulation results indicate the benefits of the proposed routing strategy in terms of increasing link duration, reducing the number of link-breakage events and increasing the end-to-end throughput.

Along with recent advances in mobile networking and portable computing technologies, there is a trend in the telecommunications industry toward the development of efficient ubiquitous systems that can provide a set of bandwidth-intensive and real-time services to multiple users while supporting their full mobility. Largescale deployment of 802.11-based technologies will play an integral part in the construction of such ubiquitous wireless mobile systems. A challenging task in the development of such networks is efficient provisioning of QoS-enabled services for mobile users. In this context, we propose a scheme that constantly monitors the overall network performance to perform admission control and traffic conditioning at the 802.11-based access points and mobile terminals. The focus is on servicelevel fairness, where different flows from the same traffic class can still receive the same QoS level even if they have different bit rates. Furthermore, given the mobility of users, the success of any resource allocation and admission control model depends on the continuity of QoS guarantees across different WLANs. This article proposes a dynamic service level agreement negotiation protocol that allows mobile terminals to perform handoffs between different WLANs while maintaining the agreed level of service. End users also can change their service levels in response to changes in network conditions. © 2007 IEEE.

Attacks against encrypted protocols are becoming increasingly popular. They pose a serious challenge to the conventional Intrusion Detection Systems (IDSs) which heavily rely on inspecting the network packet fields and are consequently unable to monitor encrypted sessions. IDSs can be broadly categorized into two types: signature-based and anomaly-based IDSs. The signature-based IDSs rely on previous attack signatures but are often ineffective against new attacks. On the other hand, anomaly-based detection systems depend on detecting the change in the protocol behavior caused by an attack. The latter can be employed to detect novel attacks, and therefore are often preferred over their signature-based counterpart. In this paper, we envision an anomaly-based IDS which can detect attacks against popular encrypted protocols, such as SSH and SSL. The proposed system creates a normal behavior profile and uses nonparametric Cusum algorithm to detect deviation from the normal profile. Upon detecting an anomaly, the proposed mechanism generates an alert, sets a delay to the protocol response, and traces back the attacker. The effectiveness of the proposed detection scheme is verified via simulations. © 2007 IEEE.

Attacks against encrypted protocols have become increasingly popular and sophisticated. Such attacks are often undetectable by the traditional Intrusion Detection Systems (IDSs). Additionally, the encrypted attack-traffic makes tracing the source of the attack substantially more difficult. In this paper, we address these issues and devise a mechanism to trace back attackers against encrypted protocols. In our efforts to combat attacks against cryptographic protocols, we have integrated a traceback mechanism at the monitoring stubs (MSs), which were introduced in one of our previous works. While we previously focused on strategically placing monitoring stubs to detect attacks against encrypted protocols, in this work we aim at equipping MSs with a traceback feature. In our approach, when a given MS detects an attack, it starts tracing back to the root of the attack. The traceback mechanism relies on monitoring the extracted features at different MSs, i.e., in different points of the target network. At each MS, the monitored features over time provide a pattern which is compared or correlated with the monitored patterns at the neighboring MSs. A high correlation value in the patterns observed by two adjacent MSs indicates that the attack traffic propagated through the network elements covered by these MSs. Based on these correlation values and a prior knowledge of the network topology, the system can then construct a path back to the attacking hosts. The effectiveness of the proposed traceback scheme is verified by simulations. Copyright 2007 ACM.

Users in next generation wireless networks are expected to be highly dynamic while maintaining connectivity through different devices with different processing and communication capabilities. In wireless environments, bandwidth is scarce and channel conditions are time-varying. To guarantee Quality of Service (QoS) to users roaming between heterogeneous wireless networks, a dynamic QoS negotiation mechanism, which allows users to dynamically negotiate their service-levels with the network, is required. Several protocols for dynamic service level negotiation have been proposed, each focusing on a particular mode. This paper presents an overview of these protocols and discusses their limitations. To alleviate these shortcomings, a dynamic QoS negotiation scheme to allow users to change their service levels in response to changes in both network conditions and their own resource requirements is proposed. In the proposed scheme, upon an intra-domain handoff of a mobile node, the visited access point consults the previously used access point to confirm the legitimacy of the service negotiation request issued by the mobile node. The performance of the proposed scheme has been investigated and compared with other dynamic negotiation approaches. It was demonstrated that the proposed scheme outperforms the state-of-the-art method, in terms of the signaling overhead and data storage, at the expense of a slight increase in the overall negotiation delay. © 2007 IEEE.

Attacks against encrypted protocols are becoming increasingly popular. They pose a serious challenge to the conventional Intrusion Detection Systems (IDSs) which heavily rely on inspecting the network packet fields and are consequently unable to monitor encrypted sessions. IDSs can be broadly categorized into two types: signature-based and anomaly-based IDSs. The signature-based IDSs rely on previous attack signatures but are often ineffective against new attacks. On the other hand, anomaly-based detection systems depend on detecting the change in the protocol behavior caused by an attack. The latter can be employed to detect novel attacks, and therefore are often preferred over their signature-based counterpart. In this paper, we envision an anomaly-based IDS which can detect attacks against popular encrypted protocols, such as SSH and SSL. The proposed system creates a normal behavior profile and uses nonparametric Cusum algorithm to detect deviation from the normal profile. Upon detecting an anomaly, the proposed mechanism generates an alert, sets a delay to the protocol response, and traces back the attacker. The effectiveness of the proposed detection scheme is verified via simulations. © 2007 IEEE.

In multi-hop sensor networks, information obtained by the monitoring nodes need to be routed to the sinks. If we assume that the transmitter power level can be adjusted to use the minimum energy required to reach the intended next hop receiver, the energy consumption rate per unit information transmission depends on the choice of the next hop node. In a power-aware routing approach, most proposed algorithms aim at minimizing the total energy consumption or maximizing network lifetime. In this paper, we propose a new routing algorithm with two goals: minimizing the total energy consumption and ensuring fairness of energy consumption between nodes. We formulate this as a nonlinear programming problem and use a sub-gradient algorithm to solve the problem. We also evaluate the proposed algorithm via simulations at the end of this paper. © 2007 IEEE.

The transmission of high quality video requires high bandwidth. Ensuring constantly high bandwidth in wireless environments is a challenging task given constraints in the current wireless network resources. Current mobile computers are equipped with multiples wireless interfaces that can be used to improve the video quality by aggregating the bandwidth of these interfaces. Such Bandwidth Aggregation (BAG) approach involves multiple paths in communication and gives rise to a number of issues related to the management of the Service Level Agreement (SLA) and packet reordering. To guarantee an efficient and fair management of SLA, this paper presents a bandwidth aggregation-aware QoS negotiation mechanism that enables users to dynamically negotiate their desired service levels and to reach them through the use of bandwidth aggregation. This operation is performed while ensuring a fair use of the network resources among all competing users. To cope with packet reordering, a new scheduling strategy is presented. The performance evaluation of the proposed bandwidth aggregation-aware QoS negotiation scheme and the proposed scheduling algorithm are conducted via simulations and the results are discussed. © 2007 IEEE.

This paper presents a novel data transmission protocol "SVFTP," which enables high-speed and error-free video data transmission over IP networks. A video transmission system based on SVFTP is also presented. While conventional protocols are designed for file transmission, SVFTP focuses on video data as a continuous media. In order to fit a flexible video transmission system, SVFTP achieves higher throughput on the long distance link as well as transmission interruption/resumption and progressive download and play back. In addition, a rate shaping mechanism for SVFTP is introduced in order to control greediness and burst traffic of multiple-TCP sessions. Laboratory and field transmission experiments show that SVFTP achieves high performance and functionality.

This paper presents a novel data transmission protocol "SVFTP," which enables high-speed and error-free video data transmission over IP networks. A video transmission system based on SVFTP is also presented. While conventional protocols are designed for file transmission, SVFTP focuses on video data as a continuous media. In order to fit a flexible video transmission system, SVFTP achieves higher throughput on the long distance link as well as transmission interruption/resumption and progressive download and play back. In addition, a rate shaping mechanism for SVFTP is introduced in order to control greediness and burst traffic of multiple-TCP sessions. Laboratory and field transmission experiments show that SVFTP achieves high performance and functionality.

The wide spread of the internet and its evolution as a new social infiastructure increase the importance of applications which best use internet potential. The intelligent functions which can be realized with resent computer technology are indispensable to best benefit from the internet. In this' pap. er we, overview the current situation together with the issues which we have to solve to realize the new infrastructure of networked society.

This report concerns a system for monitoring the quality of connections in the gateway switches that interconnect telecommunication networks. Although the time series to be monitored is clearly nonstationary, movements that are almost periodic can be seen if a seven-day interval is set as the period. Therefore, this system introduces a new technique, which is called "partition model optimization," for using the Akaike information criterion (AIC) to determine appropriate groups of local stationary models from a quasi-periodic time series. Also, to perform quality management in real time, a high-speed decision technique called the sequential probability ratio test (S) is introduced. The effectiveness of the partition model optimization technique was systematically investigated in advance by using actual traffic data. Not only was decision parameter maintenance unnecessary, but deterioration could also be detected more accurately than with the conventional system. (C) 2003 Wiley Periodicals, Inc. Electron Comm Jpn Pt 2, 86(6): 73-82, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjb.10157.

The wide spread of the Internet and its evolution as a new social infrastructure increase the importance of applications which best use its potential. The intelligent functions which can be realized with computer technology are indispensable to best benefit from the Internet. We overview the current situation together with issues which we have to solve to realize the new infrastructure of networked society.

This report concerns a system for monitoring the quality of connections in the gateway switches that interconnect telecommunication networks. Although the time series to be monitored is clearly nonstationary, movements that are almost periodic can be seen if a seven-day interval is set as the period. Therefore, this system introduces a new technique, which is called "partition model optimization," for using the Akaike information criterion (AIC) to determine appropriate groups of local stationary models from a quasi-periodic time series. Also, to perform quality management in real time, a high-speed decision technique called the sequential probability ratio test (S) is introduced. The effectiveness of the partition model optimization technique was systematically investigated in advance by using actual traffic data. Not only was decision parameter maintenance unnecessary, but deterioration could also be detected more accurately than with the conventional system. (C) 2003 Wiley Periodicals, Inc. Electron Comm Jpn Pt 2, 86(6): 73-82, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjb.10157.

It is important to give useful clues for selecting desired content from a number of retrieval results obtained (usually) from a vague search request. Compared with monolingual retrieval, such a support framework is inevitable and much more significant for filtering given translingual retrieval results. This paper describes an attempt to provide appropriate translation of major keywords in each document in a cross-language information retrieval (CLIR) result, as a browsing support for users. Our idea of determining appropriate translation of major keywords is based on word co-occurrence distribution in the translation target language, considering the actual situation of WWW content where it is difficult to obtain aligned parallel (multilingual) corpora. The proposed method provides higher quality of keyword translation to yield a more effective support in identifying the target documents in the retrieval result. We report the advantage of this browsing support technique through evaluation experiments including comparison with conditions of referring to a translated document summary, and discuss related issues to be examined towards more effective cross-language information extraction.

It is important to give useful clues for selecting desired content from a number of retrieval results obtained (usually) from a vague search request. Compared with monolingual retrieval, such a support framework is inevitable and much more significant for filtering given translingual retrieval results. This paper describes an attempt to provide appropriate translation of major keywords in each document in a cross-language information retrieval (CLIR) result, as a browsing support for users. Our idea of determining appropriate translation of major keywords is based on word co-occurrence distribution in the translation target language, considering the actual situation of WWW content where it is difficult to obtain aligned parallel (multilingual) corpora. The proposed method provides higher quality of keyword translation to yield a more effective support in identifying the target documents in the retrieval result. We report the advantage of this browsing support technique through evaluation experiments including comparison with conditions of referring to a translated document summary, and discuss related issues to be examined towards more effective cross-language information extraction.

When the Internet is used for school education, students in a classroom access the WWW under the supervision of their teacher. In this situation, many request packets for WWW objects may flood the Internet, triggered by the instructions of the teacher, and may cause heavy congestion in access lines or servers. To avoid traffic congestion, caching and mirroring techniques will be required. In this paper, we study caching and mirroring with pre-fetch methods suitable for the educational use of the Internet.

When the Internet is used for school education, students in a classroom access the WWW under the supervision of their teacher. In this situation, many request packets for WWW objects may flood the Internet, triggered by the instructions of the teacher, and may cause heavy congestion in access lines or servers. To avoid traffic congestion, caching and mirroring techniques will be required. In this paper, we study caching and mirroring with pre-fetch methods suitable for the educational use of the Internet.