Seminar report on "Routing In Mobile Ad Hoc Networks"

Developing support for routing is one of the most significant challenges in ad hoc networks and is critical for the basic network operations. Certain unique combinations of characteristics make routing in ad hoc networks interesting.

First, nodes in an ad hoc network are allowed to move in an uncontrolled manner. Such node mobility results in a highly dynamic network with rapid topological changes causing frequent route failures. A good routing protocol for this network environment has to dynamically adapt to the changing network topology.

Second, the underlying wireless channel provides much lower and more variable bandwidth than wired networks. The wireless channel working as a shared medium makes available bandwidth per node even lower. So routing protocols should be bandwidth-efficient by expending a minimal overhead for computing routes so that much of the remaining bandwidth is available for the actual data communication.

Third, nodes run on batteries which have limited energy supply. In order for nodes to stay and communicate for longer periods, it is desirable that a routing protocol be energy-efficient as well. This also provides also another reason why overheads must be kept low. Thus, routing protocols must meet the conflicting goals of dynamic adaptation and low overhead to deliver good overall performance.

Types of Routing in MANETs

• Flooding
• Proactive Routing
• On-Demand Routing
• Location-Based Routing

Book: Mobile Ad Hoc Networking by: Stefano Basagni

Web Resource: MANET Routing Protocols

Download Power Point Presentation: Routing in MANETs

Seminar report on "Expert Systems and decision making"

One of the most successful applications of artificial intelligence reasoning techniques using facts and rules has been in building expert systems that embody knowledge about a specialized field of human endeavour, such as medicine, engineering, or business.

Knowledge Base and the Inference Engine are two major parts of the expert system. The knowledge base consists of facts and rules about the subject at hand. The inference engine consists of all processes that manipulate the knowledge base to deduce information requested by the user.

Knowledge acquisition is a crucial stage in the development of expert systems. As a process, it involves eliciting, interpreting and representing the knowledge from a given domain. Knowledge acquisition for expert systems (from domain experts) is time consuming, expensive and potentially unreliable.

In the construction of an expert system, a ‘knowledge engineer’ (usually a computer scientist with artificial intelligence training) works with an expert (or experts) in the field of application in order. to represent the relevant knowledge of the expert in a form that can be entered into the knowledge base. This process is often aided by a knowledge acquisition subsystem that, among other things, checks the growing knowledge base for possible inconsistencies and incomplete information. These are then presented to the expert for resolution.

Book: Expert Systems: The Technology of Knowledge Management for the 21st Century, by: Cornelius T. Leondes.
Web Resources: Expert Systems and Artificial Engineering 
Download Power Point Presentation (PPT): Knowledge Engineering and Expert Systems

Seminar report on "Load Balancing in Parallel Computers"

A parallel compute is a collection of processing elements that communicate and cooperate to solve large problems efficiently.

Parallel computers vary in two fundamental architecture facets, (i) Single Instruction Multiple Data (SIMD) Vs Multiple Instruction Multiple Data (MIMD) and (ii) Shared memory Vs Distributed memory. A parallel computer with a logically shared memory system provides a single global address space to all processors, and hence a shared programming paradigm to the users. Such systems ae referred as distributed shared memory (DSM) machines.

Load balancing on DSM machines is a challenging task, even though the shared global address space may be used as a common pool for work-loads awaiting as in centralized memory systems. Accessing remote memory banks are very expensive, an appropriate distribution of work-loads across physically distributed memories helps reduce such costly remote access.

Creating parallel programs involves first decomposing the overall computation into tasks and then assigning the tasks to the processors, this step is also called as partitioning. The optimization objective for partitioning is to balance the work-load among processors and to minimize the inter process communication needs. The number of processes generated by the partitioning step may not be equal to the processors, thus a processor may be idle or loaded with multiple processes. The primary optimization objective of mapping is to balance the workload of processors and to minimize the inter-processor communication cost. Collectively, the problem of load balancing is to develop partitioning and mapping algorithm for the purpose of achieving their respective optimization objectives.

Load balancing algorithms can be broadly categorized as static or dynamic. Static load balancing algorithms distribute the processes to processors at compile time, while dynamic algorithms bind processes to processors at run time. Static load balancing algorithms rely on the estimate execution times of the processes and inter-process communication requirement. It is not satisfactory for parallel programs that are of the dynamic and/or unpredictable kind. Consequently in dynamic load balancing, processes are generated and destroyed without a pattern at run time. A dynamic load balancing algorithm consists of four components, Load Measurement rule, an Information Exchange rule, an Initiation rule and a Load Balancing Operation.

Book: Scheduling and Load Balancing in Parallel and Distributed Systems, Editors, Behrooz A. Shirazi, Krishna M. Kavi and Ali R. Hurson

Web Resource: Parallel Computing Tutorial

Download Power Point Presentation: Parallel Computing 

Seminar report on "Load Balancing in Internet"

The Internet is a network of networks in which users at any one computer can get access to the information from any other computer. It was conceived by the Advanced Research Projects Agency (ARPA) of the U.S. government in 1969 and was first known as the ARPANET. The main aim was to allow the researchers of different universities be able to talk each other. Today, the Internet is a public, cooperative, and self-sustaining facility accessible to hundreds of millions of people worldwide. Technically, the Internet uses a set of protocols called TCP/IP (for Transmission Control Protocol/Internet Protocol).

As the Internet operates without a central governing body, the most critical consideration in Internet is Load Balancing. Load balancing aims at improving the performance of the Internet by distributing the workload evenly across the network, in order to get optimal resource utilization, maximize throughput, minimize response time, avoid overload and increase bandwidth. To scale performance, Network Load Balancing distributes IP traffic across multiple cluster hosts. It also ensures high availability by detecting host failures and automatically redistributing traffic to the surviving hosts. Network Load Balancing brings special value to enterprises deploying TCP/IP services, such as e-commerce applications, that link clients with transaction applications and back-end databases.

Network Load Balancing servers (also called hosts) in a cluster communicate among themselves to provide key benefits, including:
• Scalability. Network Load Balancing scales the performance of a server-based program, such as a Web server, by distributing its client requests across multiple servers within the cluster. As traffic increases, additional servers can be added to the cluster, with up to 32 servers possible in any one cluster.
• High availability. Network Load Balancing provides high availability by automatically detecting the failure of a server and repartitioning client traffic among the remaining servers within ten seconds, while providing users with continuous service.

Network Load Balancing, load-balances incoming client requests by directing a selected percentage of new requests to each cluster host; the load percentage is set in the Network Load Balancing Properties dialog box for each port range to be load-balanced. The algorithm does not respond to changes in the load on each cluster host (such as the CPU load or memory usage). However, the mapping is modified when the cluster membership changes, and load percentages are renormalized accordingly.

Book: Server Load Balancing by Tony Bourke

Web Resource: Server Load Balancing: Algorithms 
                      Load balancing (computing)

Download Power Point Presentation: Load Balancing Parallel Applications on Heterogeneous Platforms

Seminar report on"TCP/IP on Gigabit Networks"

Abstract

The volume of traffic being offered to the Internet continues to increase. Many of the other presentations at this conference describe applications which will add to this volume. (Include pointers to other sessions, especially if there is one which gives statistics on traffic growth trends).Currently the fastest communications in general widespread use are FDDI for the local area (100 Mbps) and T3 (45 Mbps) for the wide area. There is a great deal of marketing promotion and limited actual experience of the use of ATM at 155 Mbps for both local and wide area connections. My purpose is to discuss technologies which can be used to carry TCP/IP at 1 Gbps - as one possible view of the technologies in widespread use before the year 2000. Using combinations of fibre optic cable, amplification and electronic switching, gigabit speeds are achievable in local, campus and world wide networks. Technologies like HIPPI and fibre channel support 1 Gbps data streams over limited distances. SDH/Sonet is required for long distance multicarrier connections. By "gigabit networks", I mean sending data at approximately 1,000 Mbits/s (ie 10 times FDDI, or 1000 times conventional Ethernet speeds).
HIPPI
HIPPI is currently in widespread use as a connection method for supercom puters, normally at 800 Mbps. There are also HIPPI interfaces available for HP, DEC Alpha, SGI Indigo 2, RS 6000, and Sun workstations, and even a HIPPI card to fit into a PCI bus. There is a specification for HIPPI at 1600 Mbps, though this is rarely used. RFC 1374 specifies IP and ARP over HIPPI. The European Centre for Medium Range Weather Forecasting has documented 57 Mbytes/s (456 Mbps) transfer rates using a ver sion of ttcp between two Cray Research machines connected using an IP router over HIPPI. There are other connections to an IBM 3090, a Silicon Graphics Challenge, and RS6000 workstations. (This installation can be described to whatever level of detail required - network diagrams, packet formats, packet sequences, etc). Sample pricing for host interface cards and IP routers at June 95 can be provided.
HIPPI advantages
Technology has been in use for several years There is an established standard Excellent for transferring large volumes of data Maximum packet size is 64K bytes
Disadvantages
One connection excludes any other traffic until first transfer complete Inefficient for small packet sizes On copper, maximum distance 25 meters. Connectors fragile. Serial HIPPI runs at much longer distances over fibre.
Fibre channel
Fibre channel is ANSI standard X3T11 for the physical and signalling interface. Also refer to the Internet Draft "IP and ARP on fibre channel" It is much more complex than HIPPI, having a connectionless, a connection oriented and a `mixed' mode. It can be used in `direct channel' mode, to support TCP/IP, or for peripheral connection, in a switched configuration, in a routed environment, or in a loop. `Quarter speed' fibre channel products are in use (at 266 Mbps) and 1 Gbps fibre channel host interface cards and switches will be available by June 95. One customer running TCP/IP between two RS6000s connected via a fibre channel switch at quarter speed measured a transfer rate of more than 125 Mbps using TCP/IP, at which point the bottleneck in the file transfer was the disk access speed of the RS6000s, not the network transfer rate. (By June I expect to have implemetation experience with fibre channel at 1 Gbps, both host interface cards and switches). Sample pricing valid at that date can be provided, and compared to HIPPI prices.
SDH/Sonet
SDH and Sonet are similar but significantly different systems used for modern trunk digital communictions network infrastructures by long distance carriers. They offer substantial improvements over the PDH for sustainable service, given adequate network design, both because to the fundamental architecture (dual rings) and the SS7 signalling. In the UK the new carriers (Mercury Communications, Energis, Scottish Power, United Artists, Videotron) have installed it as they have built their infrastructures. BT are retrofitting it to their existing network. Certain customers have been already been able to have SDH connections at 155 Mbps installed on their premises. RFC 1619 specifies PPP encapsulation for IP over SDH/Sonet. If router manufacturers were to implement this, then it would be possible to use existing and developing IP routing and mulitcasting protocols (BGP4, DVMRP, PIM, and mobile IP) directly on the SDH framing. ATM then becomes just one of the many protocols which can be carried in IP. The SDH hierarchy starts at 155 Mbps, and goes up through 622 Mbps, to 2.4 Gbps.
ATM Equipment is available but very little of it is in actuall service use. Much of the necessary standardisation work has not been completed. Customers are beginning to realise that ATM is in a similar state to ISDN five years ago. ATM is specified at a range of speeds, the highest of which is 620 Mbps. Operated at this speed, there is a very substantial throughput loss due to overhead, giving an effective rate of approximately 450 Mbps - a long way short of 1 Gbps.
Conclusion : This has been a brief discussion of possible technologies for running IP at speeds higher that the currently widely available fast ethernet or 155 Mbps ATM speeds.

Seminar report on "Biometrics Authentication systems"

Reliable user authentication is becoming an increasingly important task in the Web-enabled world. The consequences of an insecure authentication system in a corporate or enterprise environment can be catastrophic, and may include loss of confidential information, denial of service, and compromised data integrity. The value of reliable user authentication is not limited to just computer or network access. Many other applications in everyday life also require user authentication, such as banking, e-commerce, and physical access control to computer resources, and could benefit from enhanced security.

A wide variety of systems require reliable personal recognition schemes to either confirm or reject the identity of an individual requesting their services. The purpose of such schemes is to ensure that the services are accessed only by a legitimate user and no one else. Examples of such applications include secure access to buildings, computer systems, laptops, cellular phones, and ATMs.

In the absence of robust personal identification schemes, these systems are vulnerable to the undesired access. Biometric recognition refers to the automatic recognition of individuals based on their physiological and/or behavioral characteristics. With the help of biometrics it is possible to confirm or establish an individual’s identity based on “who he/she is”, rather than by “what he/she possesses” (e.g., an ID card) or “what he/she remembers” (e.g., a password). The prevailing techniques of user authentication, which either passwords and user Ids (identifiers), or identification cards and PINs (personal identification numbers), suffer from several limitations. Passwords and PINs can be acquired by direct covert observation. Once an intruder acquires the user ID and the password, the intruder has total access to the user’s resources. Biometric systems gives a better alternative to the existing systems.

The bioentities that can be used for authentication are Deoxyribonucleic Acid (DNA), Face, Fingerprint, Iris (the colour pattern in eyes bounded by pupil and limbus), Hand geometry, Palmprint, Signature, Voice, Other biometrics such as gaits, lip prints, brain signals, ears, teeth, retinas, odor, keystrokes, heights, weights and genders have been proposed.

Book: Biometric Systems: Technology, Design and Performance Evaluation by James Wayman

Web resources: Biometrics: A Look at Facial Recognition 

Download Power Point Presentation (PPT): Biometrics

Seminar report on "Next Generation Protocol (IPv6)"

One of the greatest revolutionary innovations of the twentieth century is Internet. The Internet is a global system of interconnected computer networks that use the standard Internet Protocol Suite (TCP/IP) to serve billions of users worldwide. It is a network of networks that consists of millions of private and public, academic, business, and government networks of local to global scope that are linked by a broad array of electronic and optical networking technologies.

Rapidly, there is a dramatic change in the way we communicate, do business, educate ourselves and entertain ourselves. Perhaps even the architects of Internet would not have foreseen the tremendous growth rate of the network being witnessed today. It made the ‘global village utopia ‘ a reality in a rather short span of time.

Using the internet it is easy for the people to exchange ideas, resources and information. Likewise there are millions of users to the Internet and the drawback of the Internet is that it cannot adequately support many services being imagined, such as interconnection of gigabit networks with lower bandwidths, high security applications and interactive virtual reality applications.

A more serious problem with today’s Internet is that it can interconnect a maximum of four billion systems only, which is a small number as compared to the projected systems on the Internet in the twenty-first century.

As IPv4 support 32 bit address, which is given to each machine on the net and unique. With 32 bits, a maximum of about four billion addresses are possible. Even though this is a large number, but the latest trends of accessing Internet from mobile phones, TV sets, and even pizza machines will lead to the shortage of addressing bits. Since each of them must have an IP address, this number becomes too small.
The revision of IPv4 “Internet Protocol version 4” was taken up mainly to resolve the address problem, but in the course of refinements, several other features were also added to make it suitable for the next generation Internet. This refined version was initially named IPng “IP next generation” and is now officially known as IPv6.

IPv6 supports 128-bit addresses, the source address and the destination address, each being, 128 bits long. Presently, most routers run software that support only IPv4. To switch over to IPv6 overnight is an impossible task and the transition is likely to take a very long time. However to speed up the transition, an IPv4 compatible IPv6 addressing scheme has been worked out.

Books: IPv6 Essentials by Silvia Hagen

Web Resources: IPv6: The Next Generation Internet 
                        The Source for IPv6 Information

Download Power Point File (PPT): Next Generation Protocol