The solutions to the three main menaces of
information security--A machine learning approach to spam mails, spywares and
web intrusion detections
Shi-Jinn Horng,
National Taiwan University of Science and Technology, Taiwan, China
Fault-Tolerant Routing
in Dual-Cube
Yamin Li, Hosei University, Tokyo, Japan
Interaction design and
usability for ubiquitous, mobile and everyday devices
Frode Eika Sandnes, Oslo University College, Norway
Failure Detection in
Distributed Systems: Retrospective and Recent advances
Xavier DEFAGO, Japan Advanced Institute of Science and
Technology, Japan
Mobile Ad Hoc Networks
and Smart Antenna Technology: New Challenges at the MAC Layer
Jacir Bordim, Brasilia University, Brazil
The solutions to the three main menaces of information security: A machine
learning approach to spam mails, spywares and web intrusion detections
Shi-Jinn Horng,
National Taiwan University of Science and Technology, Taiwan, China
TUTORIAL DESCRIPTION
The revolution of the network is expanding like a raging fire of the world. People’s life has a great of changes because a number of applications can be found in the network. For example, owing to the invention of E-mail, the interpersonal communication becomes shorter and faster; all kinds of websites created are used to accelerate the integrating and the spreading of information to everyone; the sharing and exchanging of application software makes life be more abundant and convenient too. The network has already become an indispensable part in our life.
But when we are enjoying all the advantages brought with the network, some new problems follow. As to users, the most influential one is no more than the threats on the information security. There are three main kinds of threats that were disturbing people; the spam mail overflow, threat with spyware and web intrusion. In the face of those threats, the traditional way can only defend us passively, it detects a known danger by pattern matching. Whenever there is a novel attack appeared, the traditional detecting method will find no way out. Especially, the traditional defending methods which always worked after a huge loses are not the good one that we really want. In order to improve such a helpless situation, a new way is to combine the concept of machine learning with behavior analysis and characteristic extraction to train out a brand new detecting model. The protecting mechanism not only can defend and stop up the known attacks effectively, but also intercept the novel attacks voluntarily. With such kind of mechanism, it can achieve the goal of ensure the information security against those threats in real time.
ABOUT THE PRESENTE
Dr Shi-Jinn Horng received the BS degree in electronics engineering from the National Taiwan Institute of Technology, the MS degree in information engineering from the National Central University, and the PhD degree in computer science from the National Tsing Hua University in 1980, 1984, and 1989, respectively. Currently, he is a full professor in the Computer Science and Information Engineering Department and also the director of Computer Center at the National Taiwan University of Science and Technology. He has published more than 100 research papers and received many awards especially the one named as distinguished research award of National Science Council 2004. His research interests include information security, VLSI design, multiprocessing systems, and parallel algorithms.
Interaction design and usability for
ubiquitous, mobile and everyday devices
Frode Eika Sandnes, Oslo University College, Norway
TUTORIAL DESCRIPTION
We are increasingly dependent on the mobile, miniature and
mobile devices that are surrounding us, from mobile phones, PDA's, mp3 players
to advanced digital wristwatches and other wearable computers. Other everyday
objects are constantly added to this list as they are gaining electronic
intelligence. Technological advances are making these consumer electronics
devices more computationally powerful and physically smaller. However, many
device designs are failing in the marketplace because they are designed by
engineers without proper
understanding of the end users, their needs and how they interact with these
devices. In this tutorial we will look at the human factors in device design.
Key principles that should drive a design will be covered with practical
examples. After completing the workshop the attendee should be able to evaluate
a given device and objectively compare different designs. Further, the
participants should be better equipped at making sensible design decisions that
adhere to state of the art in interaction design for ubiquitous and mobile
devices.
ABOUT THE PRESENTER
Frode Eika Sandnes received a B.Sc. in computing science from the University of Newcastle Upon Tyne, England, in 1993, and a Ph.D. in computer science from the University of Reading, England, in 1997. He has several years of experience from the space industry developing communications and onboard systems for low-earth orbit environmental satellites. He is currently an Associate Professor in the Department of Computer Science at Oslo University College, Norway. Dr. Sandnes' research interests include multiprocessor scheduling, error-correction and mobile human computer interaction.
Failure Detection in Distributed Systems:
Retrospective and Recent advances
Xavier DEFAGO, Japan Advanced Institute of Science and
Technology, Japan
TUTORIAL DESCRIPTION
The tutorial will present several aspects of failure detection in asynchronous distributed systems, including an introduction to the basic theoretical foundations, implementation issues, as well as architectural issues. Failure detection is a fundamental issue to building fault-tolerant distributed computer systems. In most systems, it is not possible to distinguish, with absolute certainty, a crashed node from one that is simply very slow. The tutorial will provide the attendance with an exhaustive view of the abundant researches, both practical and theoretical, that have been developed in this field, including very recent advances.
ABOUT THE PRESENTER
Xavier Defago is an associate professor at the Japan Advanced
Institute of Science and Technology (JAIST), as well as a PRESTO researcher for
the Japan Science and Technology Agency (JST). He is a member of the 21s century
Center of Excellence program on "Trustworthy e-Society" at JAIST. Xavier
obtained the Ph.D. in Computer Science in 2000 from the Swiss Federal Institute
of Technology in Lausanne (EPFL) under the supervision of Prof. Andr'e Schiper.
In the summer 2005, Xavier was an invited professor at the IRISA and the
University of Rennes 1 (France). He served on the program committee of several
international conferences on dependable computer systems, including DSN, SRDS,
ISORC, and LADC. He is also responsible for the track on ``dependable and
autonomic computing'' of the AINA 2006 conference. His research interests
include distributed algorithms, fault- tolerant distributed systems, mobile ad
hoc networks, as well as their application to cooperative autonomous mobile
systems.
Mobile Ad Hoc Networks and
Smart Antenna Technology: New Challenges at the MAC Layer
Jacir Bordim, Brasilia University, Brazil
TUTORIAL DESCRIPTION
Mobile ad hoc networks are envisioned as rapidly deployable, infrastructureless networks where each node is equipped with wireless capabilities and act as a mobile router. These characteristics make ad hoc networks suitable to support communications in urgent and temporary tasks, such as business meetings, disaster-and-relief, search-and-rescue, law enforcement, among other special applications. Despite its benefits and wide range of applications, ad hoc networks are known to suffer from poor network throughput due to the relay load imposed by distant nodes and the inefficient spatial reuse provided by omnidirectional antennas. Aiming to improve over this scenario, the research community begun considering the use of smart antenna technology in the context of ad hoc networks. The key benefits provided by smart antennas include reduced co-channel interference, transmission range extension, better spatial reuse and signal quality. However, these benefits can only be fully exploited when efficient medium access control (MAC) protocols, tailored for smart antennas, become available. This tutorial will present an overview of smart antenna technology, its benefits and suitability for use in ad hoc networks, as well as techniques and known problems in designing suitable MAC protocols.
ABOUT THE PRESENTER
Jacir L. Bordim received B.E. degree from Passo Fundo University (Brazil) in 1994, M.E. degree from Nagoya Institute of Technology (Japan) in 2000, and PhD degree from Japan Advanced Institute of Science and Technology in 2003. From 2003 to 2005, he worked as at ATR/ACR Labs (Japan), where he was enrolled in the design and test of MAC layer protocols tailored for smart antenna technology. Since March 2005, he has joined the Department of Computer Science at Brasilia University (Brazil), where he is an Associated Professor. His interests include mobile computing, sensor networks, computer networks, distributed systems, and graph theory.
Fault-Tolerant
Routing in Dual-Cube
Yamin Li, Hosei University, Tokyo, Japan
TUTORIAL DESCRIPTION
The binary hypercube, or n-cube, has been widely used as the
interconnection network in parallel systems, SGI Origin 2000 for instance.
However, the major drawback of the hypercube is the increase in the number of
communication links for each node with the increase in the total number of nodes
in the system. In order to connect more nodes, SGI Origin 2000 uses a special
router to link multiple hypercubes.
In this tutorial, we introduce a low-degree interconnection network, called
dual-cube, which can link much more nodes than hypercube. An n-cube connects 2^n
nodes where n is the number of links per node. In contrast, the dual-cube can
connect up to 2^{2n-1} nodes. This network mitigates the problem of increasing
number of links in the large-scale hypercube network while keeps most of the
topological properties of the hypercube network.
We investigate the topological properties of the dual-cube, compare them with
other hypercube-like networks, and establish the basic routing and collective
communication, including one-to-all, all-to-all broadcasting and one-to-all,
all-to-all personalized communication algorithms.
Then we present how to build the disjoint paths between any two nodes in
dual-cube and how to find a fault-free cycle containing as large number of nodes
as possible in dual-cube that may contain faulty nodes.
Finally, we give two efficient algorithms for finding a nonfaulty routing path
between any two nonfaulty nodes in dual-cube with a large number of faulty
nodes. One uses the adaptive-subcube technique and the other uses binomial-tree
technique. The experimental results show that, with high percentages of node
failures, the two algorithms can build routing paths with a very high
probability.
ABOUT THE PRESENTER
Yamin Li is a professor of the Department of Computer
Science, Hosei University, Tokyo, Japan. He received his PhD degree in computer
science from Tsinghua University, Beijing, China, in 1989. His research
interests include computer architecture, parallel and distributed computing,
interconnection networks and fault-tolerant computing. Dr. Li is a senior member
of the IEEE and a member of the IEEE Computer Society.