Sridhar Iyer's Project Students

Sridhar Iyer: Project Students

This page is perpetually under construction. More information will be added as and when new students join and/or old students graduate.
Reports/Slides of each project will be available only after the abstract appears on this page.

Post-Doctoral Fellows

Year Student Topic

[2007+] Dr. Farida Umrani
(PhD in Psychology from IIT Bombay) Impact assessment of eLearning technologies.
[Publications: CM1; DL 08; ]
(Fellowship supported by Project OSCAR)

[2001-2003] Dr. Leena Chandran-Wadia
(PhD in Physics from IISc Bangalore) Medium access control in wireless networks.
[Publications: ICCC 02; ICCCN 03; ICDCS 03; CIT 03.]
(Fellowship supported by Infosys)

PhD Projects Back

Year Student Topic

[2009+] Sameer Sahasrabudhe Design considerations for e-learning animations.

[2003-2008] Saraswathi Krithivasan Adaptive mechanisms for multimedia dissemination

[2003-2009] Raghuraman Rangarajan Design of multi-tier wireless mesh networks

[2002-2006] Vijay Raisinghani Cross layer feedback architecture for mobile device protocol stacks

[2001-2007] Srinath Perur Connectivity properties for Topology design in sparse multi-hop wireless networks

[2001-2006] Vikram Jamwal BoBs: Breakable objects - Building blocks for flexibile application architectures

Master's Projects Back

Year Student Topic

-

[2009-2010] N Ganesh Adaptation of CDEEP lecture videos to mobiles

[2009-2010] J Manoj System for transmitting CDEEP videos over low bandwidth networks

[2009-2010] Mohammed Nazeem Unicast-Multicast gateway for tunneling lectures from any CDEEP studio to VSAT network

[2009-2010] Harshad Inarkar Application layer multicast for reducing load on CDEEP VOD server

[2009-2010] Rohit Gujrati Modeling the CDEEP system using system dynamics

-

[2008-2009]
Sabbatical Kapil Kadam
[external student] Translation of OSCAR applets to midlets for mobiles

-

[2007-2008] Nidhi Malik Discovering dependencies in courseware repositories

[2007-2008] Janak Chandrana Implementation of WiFiRe MAC: Framing, Memory and Wireless modules

[2007-2008] M Ranjith Kumar Implementation of WiFiRe MAC: Ranging, Registration and Classification modules

[2007-2008] Sreedhar Reddy M Implementation of WiFiRe PHY sectorization in OPNET

[2007-2008] Kedar Rudre MH-WiFiRe: Multi-hop extension for WiFiRe

-

[2006-2007] Moniphal Say Shikav extensions to support networking animation

[2006-2007] H Shravan Kumar Design and implementation of MAC layer of WiFiRe protocol

[2006-2007] Sameer Kurkure Design and implementation of WiFiRe MAC layer protocol

[2006-2007] Ch Sudheer Keshav Formal specification and verification of WiFiRe protocol

[2006-2007] Venkat Reddy M Performance evaluation of WiFiRe using OPNET

[2006-2007] Priyesh Wadhwa Kumar Design of PSTN-VoIP gateway with inbuilt PBX & SIP extensions for wireless medium

[2006-2007] K Sravana Kumar Design of PSTN-VoIP gateway for rural environments

-

[2005-2006] Annanda Rath HSM: A hybrid streaming mechanism for delay-tolerant multimedia applications

[2005-2006] Laxmikant Patil Video transmission over varying bandwidth links

[2005-2006] Anshu Veda Application partitioning - A dynamic, runtime, object-level approach

[2005-2006] Anshuman Tiwari Proxy-AODV: Extension of AODV for partially connected Adhoc networks

[2005-2006] Anirudha Bodhankar Implementation of WiFiRe model in OPNET

[2005-2006] Nitesh Gupta RFIDplanner: A coverage planning tool for RFID networks

[2005-2006] Naval Bhandari Improving RFID systems to read tags efficiently

-

[2004-2005] Kiran Kumar Gavini PLUS-DAC: An admission control scheme for IEEE 802.11 wireless LANs

[2004-2005] Supriya Maheshwari An efficient QoS scheduling architecture for IEEE 802.16 wireless MANs

[2004-2005] Shailesh Birari Mitigating the Reader Collision Problem in RFID networks

[2004-2005] Anusha S. RFIDcover: A coverage planning tool for RFID networks with mobile readers

[2004-2005] Chaitanya Krishna B. M2MC: Middleware for Many to Many Communications over broadcast networks

-

[2003-2004] Kiran Diwakar QoS guarantees for real-time applications in 802.11 WLANs

[2003-2004] Kalpesh Patel Lookup service for peer-to-peer systems in mobile adhoc networks

[2003-2004] Kalyan Rao D. Performance Comparison of DAMA MAC schemes over satellite networks

[2003-2004] Ranjeeth Kumar A. MSIP: A protocol for efficient handoffs of real-time multimedia in mobile wireless scenarios

-

[2002-2003] Anupam Goyal Design and deployment of file transfer protocol over asymmetric satellite networks

[2002-2003] Abhishek Goliya Dynamic adaptation of the DCF and PCF mode in IEEE 802.11 WLANs

[2002-2003] Satyajit Rai Design and evaluation of IEEE 802.11 based Dual MAC for mobile adhoc networks

[2002-2003] Deepanshu Shukla Tackling the exposed node problem in IEEE 802.11 MAC

-

[2001-2002] Manoj Kumar C. Video streaming in mobile environments

[2001-2002] Abhilash P. Route repair in mobile adhoc networks

[2001-2002] Vijil E.C. Security issues in mobile agents

[2001-2002] Ajay Kumar Singh ATCP: Adapted TCP for mobile wireless environments

[2000-2002] Mamata Desai Distributed intrusion detection systems

-

[2000-2001] Richa Jain Extending the ns2 simulator for GPRS support

[2000-2001] Dr. Sheikh Mahmood Distributed archival and retrieval of medical images using DICOM

[2000-2001] Srinath Perur Routing in mobile adhoc networks

[2000-2001] Vikram Jamwal Mobile agents for distance evaluation

[2000-2001] Rahul Jha Mobile agents for e-commerce

[2000-2001] Venkat G. Automated testing of UML behavioral descriptions

-

[1999-2000] Raghuraman Rangarajan Sequential and parallel reachability analysis of concurrent Java programs

[July 2009+] Design considerations for animations in e-learning: Sameer Sahasrabudhe
Note: Registered with: YCMOU; Working at IIT Bombay; [Co-guide: Prof. Sahana Murthy]

Download: Thesis: Slides:
Abstract:

[July 2003 - Oct 2008 (defended June 2009)] Adaptive mechanisms for multimedia dissemination: Saraswathi Krithivasan

Download: Thesis: PDF file [1.5 M]. Slides: PPT [3.8 M].
Abstract: Delay-tolerant multimedia applications, where clients specify the time when playout must start, fit the profile of many emerging applications, including distance education and corporate training. Such applications typically depend on a Content Delivery Network (CDN), where the Content Service Provider (CSP) disseminates multimedia content to geographically dispersed clients through a distribution network from servers located on the CDN. In these applications, a client Ci connects to the source S at time t0 and requests for contents with base encoding rate Γ and playout duration T ; Ci specifies the time it is willing to wait for the playout to start, its delay tolerance δi ; Playout at Ci starts at (t0 +δi ) and ends at ((t0 + δi ) + T ). Note that t0 may also be the start time of a scheduled streaming session.
This thesis deals with the issue of maximizing the quality of the multimedia contents delivered at the clients even when link bandwidths are constrained in the path from the source to the clients. To achieve this, we develop a suite of algorithms that can exploit clients’ delay tolerance considering the following parameters: (i) Service type: whether contents are streamed according to a schedule or occur on-demand and (ii) Bandwidth: whether link capacity is static or variable, using appropriate resources at the nodes – Transcoders, Layer encoders, or Streaming servers with transcoding or layering functionality. In particular, we consider the following three cases: (i) Scheduled streaming when bandwidths are static, (ii) Scheduled streaming when bandwidths are varying and predicted, and (iii) On-demand streaming when bandwidths are static.
The algorithms developed are a result of formulating the objectives in an optimization framework or by an analysis of properties of the network topology and client requirements. Furthermore, where an optimal algorithm is computationally expensive, we develop algorithms based on heuristics as practical alternatives. The approaches are validated through extensive simulations, using topologies derived from real-world scenarios.
The algorithms developed in this thesis would help a CSP serving clients in a subscription based network in: (i) improving the quality of reception at the clients by leveraging their delay tolerance values, (ii) estimating the resources required to provide the best delivered rates to clients, and (iii) determining placement of resources to maximize the delivered rates at clients. Thus, using the analysis presented and algorithms developed in this thesis, a CSP can deploy resources in order to ensure effective quality of service to clients and efficient resource utilization by leveraging clients’ delay tolerance.

[Jan 2003 - 2009; (defended July 2009)] Design of multi-tier wireless mesh networks: Raghuraman Rangarajan

Download: Thesis: PDF file [2.7 M]. Slides: PDF [2.7 M].
Abstract: In this thesis, we investigate the issue of automated design of capacity-constrained Wireless Mesh Networks (WMN). We argue for the necessity of applying network design methodologies from wired and cellular network fields in wireless network design scenarios and present algorithms for Wireless Local Area Networks (WLANs) and backbone topology design. The deployment scenario we envision is a campus of office buildings requiring wireless connectivity. The client nodes to be deployed in each office, their application traffic requirements and the deployment layout are given. We identify three main stages in the design of such wireless networks: 1) Association of clients to access points, 2) WLAN topology construction and 3) Backbone topology construction.
Capacity provisioning and network cost minimisation are the two constraints imposed on the design problem. In the first stage, we define the AP-assignment problem, that is, the problem of associating client nodes with the nearest Access Point (AP) and investigate various access point association scenarios. We compute the performance of 802.11 WLANs under homogeneous realtime application deployments with theoretical and OPNET simulation results for various voice and video codecs. We capture the performance, in terms of number of flows supported for an application, as the capacity of a WLAN. We then examine heterogeneous application deployments and show the inability of 802.11 DCF mechanism to handle them. We propose an extended DCF for handling scenarios where delay-sensitive and delay-tolerant applications are deployed together.
Next, we propose a novel approach called sub-optimal AP-assignment and show that the utilisation of 802.11 DCF can be increased by up to 75\%. We show how the solutions to the AP-assignment problem can then be used for abstract representations of the 802.11 DCF MAC in wireless design problems. In the second stage, we define a network design problem for constructing WLAN topologies. The scenario we consider is intra-office connectivity for client nodes. We present a recursive bottom-up algorithm for capacity-constrained topology construction. The topology construction algorithm considers the deployment scenario, the client nodes deployed and their application scenarios as inputs. The AP-assignment solutions are used as a construction mechanism in the form of link specification functions. We introduce a new object, called composite unit, as an abstract building block for network topology construction. The generated topology is then validated with simulations (OPNET Modeler).
In the third stage, we define a wireless mesh network (WMN) design problem for constructing a mesh topology for a campus-like scenario. The design problem is defined as a case of traditional network design problem for optimal node location and topology construction. These problem form Mixed Integer Linear Programming (MILP) formulations and are solved with an MILP solver (CPLEX). We have built a tool to implement our multi-tier wireless design solution. The tool, the Wireless Infrastructure Deployment tool (WIND), designs topologies for both WLANs as well as WMNs. WIND takes information about nodes deployed, their properties and deployment layouts to construct logical network topologies. The input parameters and output topology use XML schemas compatible with data formats of the OPNET Modeler network simulator. This allows the topology to be input directly to the simulator for validation. Using simulation we show that the constructed topologies satisfy the given constraints on application scenarios, protocols and deployment scenario. We present case studies of constructing topologies for both WLAN and WMN scenarios.

[Jan 2002 - Jun 2006 (defended Dec 2006)] Cross layer feedback architecture for mobile device protocol stacks: Vijay Raisinghani

Download: Thesis: PDF file [1.0 M]. Slides: PDF [1.3 M].
Abstract: Applications using traditional protocol stacks (for example, TCP/IP) from wired networks do not function efficiently in mobile wireless scenarios. This is primarily due to the layered architecture and implementation of protocol stacks.
Cross layer feedback is one of the mechanisms to improve the performance of a layered stack in mobile wireless environments. For example, transport layer retransmissions could be reduced by making it aware of network disconnections or hand-off events. One such optimization is Receiver Window Control. Since the protocol stack is an integral part of the operating system, any cross layer modification to the stack should not impact its correctness, efficiency and maintainability. An appropriate architecture would help ensure that cross layer modifications confirm to these requirements.
We define the design goals for a cross layer architecture based on our study of the pros and cons of existing approaches to cross layer feedback. We present our architecture ECLAIR which addresses these design goals. In ECLAIR we exploit the fact that stack behavior is determined by the values stored in various protocol data structures. Our architecture facilitates easy manipulation of these values stored in the protocol data structures. ECLAIR requires minimal or no modification to the existing protocol stack.
We validate and evaluate ECLAIR through a prototype implementation, of receiver window control, and experiments. To evaluate a cross layer architecture we identify metrics for evaluation against each of the design goals. We identify time and space overhead, user-kernel crossing, data path delay, number of changes to protocol stack and number of changes to cross layer optimization as metrics for the design goals. Our results and analyses show that ECLAIR is an efficient cross layer architecture and is easily maintainable. To further enhance the efficiency of ECLAIR we propose a core sub-architecture. Finally, we also present a design guide for cross layer optimizations using ECLAIR.

[Jul 2001- Apr 2007 (defended July 2008)] Connectivity properties for Topology design in sparse multi-hop wireless networks: Srinath Perur

Download: Thesis: PDF file [840 K]. Slides: PPT [670 K].
Abstract: Multi-hop Wireless Networks (MWNs) are decentralised, infrastructure-less networks enabled by cooperative multi-hop routing among the participating nodes. In this work, we deal with topology design with respect to connectivity properties for sparse MWNs.
In existing work, MWN topology design has primarily focused on one metric: connectivity. Connectivity is the probability that all the nodes of a network form a single connected component. Most related work consists of asymptotic analyses dealing with finding the values of network parameters that ensure that the MWN is connected with high probability. The parameters defining the network are usually the number of nodes, their transmission ranges, and the dimensions of the deployment area.
In this work, we deal with sparse MWNs, which are unlikely to be completely connected. We argue that sparse networks can form during the functioning of MWNs, and further, that networks can be designed to be sparse in order to facilitate tradeoffs between network parameters. Since much existing work on connectivity is asymptotic, and since it focuses only on the operating point at which the network becomes connected, we provide a finite-domain, empirical model for connectivity. However, we find that connectivity is not ideal for dealing with sparse MWNs because it is i) not indicative of the extent to which the network supports communication; and ii) it is unresponsive to fine changes in network parameters. We introduce a connectivity property called reachability, defined as the fraction of connected node pairs in the network, which we claim is more appropriate for topology design in sparse MWNs. We define and prove properties of reachability, and illustrate its application in performing fine-grained tradeoffs in network parameters through a case study. We also provide a finite-domain, empirical characterisation of reachability, and a tool called Spanner (Sparse Network Planner) to help apply this model. Given three values from side of the deployment area, number of nodes, uniform transmission range of the nodes, and reachability, Spanner computes the fourth. Our empirical charecterisations of connectivity and reachability are for static networks with up to 500 nodes uniformly distributed at random in a square area. These are also applicable to networks with mobile nodes where the mobility model preserves the uniform distribution of nodes.
Much work in the area, including our characterisations of connectivity and reachability, are for networks operating in a square area of deployment. We show that results obtained for a square area do not necessarily apply even to similar rectangular areas. We ascribe this to the edge effect by which nodes located near the boundaries of the area of operation cannot utilise their entire transmission coverage for communication. We quantify analytically the expected coverage for a single node in a rectangle and describe how this can be applied in extending results obtained for square areas to rectangular areas.
We have also developed a simulator, Simran, for studying topological properties of MWNs. Simran takes as input a scenario file with initial positions and movement traces of nodes, and generates a trace file containing metrics of topological interest such as average number of neighbors, averaged shortest path lengths over all pairs of nodes, reachability, connectivity, and number and size of connected components.

[Jan 2001 - Aug 2006 (defended July 2007)] BoBs: Breakable objects - Building blocks for flexibile application architectures: Vikram Jamwal

Download: Thesis: PDF file (old version) [3.6 M]. Slides: PDF [870 K].
Abstract: Computing systems are becoming increasingly varied in terms of computing and networking capabilities. A given application may often need to be deployed in diverse scenarios. Additionally, a software needs to evolve continuously, sometimes in disparate directions, to suit a diverse portfolio of user requirements. However, given an application defined for one scenario, it may not be possible to simply refactor the application for a direct reuse or an easy adaptation to another scenario. The application itself may need to be redesigned to enable such flexibility. Hence a pertinent problem is - How do we design and implement a software system such that it has a lesser compulsion to be redesigned, and has a greater flexibility to adaptation?
In this thesis we propose a novel way of constructing such applications - wherein an application is built using Breakable Objects or BoBs. BoBs are similar to the traditional objects or components, but have the property that they can be readily broken into sub-objects or sub-components. As such, BoBs naturally form the building blocks for flexible application architectures. We term this architecture as BODA - Breakable Object Driven Architecture. We claim that: (i) BODA provides an architecturally robust mechanism for flexible fine grained reuse, and (ii) BODA greatly facilitates automatic application translations for various deployment scenarios.
We apply BODA in the context of object-oriented programming systems. We present a programming model for BoBs in Java called JavaBoB . JavaBoB is a subset of Java language specification, and contains some small extensions to the present lan- guage. We also present mechanisms by which BoBs can be composed, decomposed and used in an application. Furthermore, the BODA process is illustrated and evaluated in the thesis using three main case-studies and some distilled-data from similar other application experiences.

[2010-2011] Development of a PBL system for e-learning: Souman Mandal

Download: Thesis: Slides:
Abstract:

[2010-2011] Automated extraction of concept maps and ontologies from web-based courses: Neela Madhav G

Download: Thesis: Slides:
Abstract:

[2010-2011] Automated insertion of index tags into CDEEP videos: K Vijay Download: Thesis: Slides:
Abstract:

[2009-2010] Adaptation of CDEEP lecture videos to mobiles: N Ganesh

Download: Thesis: Slides:
Abstract:

[2009-2010] System for transmitting CDEEP videos over low bandwidth networks: J Manoj

Download: Thesis: Slides:
Abstract:

[2009-2010] Unicast-Multicast gateway for tunneling lectures from any CDEEP studio to VSAT network: Mohammed Nazeem [Guide: Prof. Puru Kulkarni]

Download: Thesis: Slides:
Abstract:

[2009-2010] Application layer multicast for reducing load on CDEEP VOD server: Harshad Inarkar [Guide: Prof. Puru Kulkarni]

Download: Thesis: Slides:
Abstract:

[2009-2010] Modeling various aspects of CDEEP operations using system dynamics: Rohit Gujrati [Guide: Prof. Sahana Murthy]

Download: Thesis: Slides:
Abstract:

[2008-2009 (Sabbatical)] Translation of OSCAR applets to Midlets for mobiles: Kapil Kadam [Student from Walchand College of Engineering, Sangli]

Download: Thesis: Slides:
Abstract:

[2007-2008] Discovering dependencies in courseware repositories: Nidhi Malik

Download: Thesis: PDF file [2.1 M]. Slides: PDF [135 K].
Abstract: Nowadays elearning has become popular, especially with the availability of large courseware repositories such as NPTEL and OCW. A variety of e-learning tools and systems are also available. However, suppose a user wants to learn about a particular subject, the search tools typically just return a large number of links to the user in response to his/her query. Many of these are not directly relevant, so the user does not know which links to follow in order to enhance his knowledge.
In this project we have built a system which not only provides the user with the most relevant learning module for his query, but also provides him/her with the relevant pre-requisite and follow-up modules also. This is done by creating a dependency graph of the courseware modules in the repository. We have implemented and tested our system on six courses taken from the NPTEL repository. For effective evaluation of our solution approach, we have not only compared the results for four different heuristics but also compared them with the dependencies determined by an expert in the subject area.

[2007-2008] Implementation of WiFiRe MAC: Framing, Memory and Wireless modules: Janak Chandrana [co-guide: Prof. Anirudha Sahoo]

Download: Thesis: PDF file [1.3 M]. Slides: PDF [835 K].
Abstract: Long range wireless for data and voice connectivity is being considered as viable and affordable solution for rural India for few years now. Numerous solutions were presented to bridge the digital divide; WiFiRe is one of them. Basic idea is to change CSMA/CA MAC of 802.11 with more efficient TDMA MAC. WiFiRe promises higher data-rate, longer range and cost-effective solution as compared to WiFi based solution.
Preliminary implementation was started last year with design phase and we are extending that work. This year, WiFiRe MAC team at IIT Bombay has implemented most of the WiFiRe MAC modules mentioned in the WiFiRe draft and they can be integrated with 802.11b PHY easily. Our MAC implementation is able to support web and voice traffic for multiple subscriber terminals and their clients. It also supports end-to-end connectivity for all the clients, while they are unaware of underlying layer-2 protocol.
We have come across numerous implementation issues and design decisions with respect to WiFiRe MAC testbed. While developing WiFiRe MAC, we got insights into various system and networking issues with hands-on experience. We have also suggested various extensions possible to this work in PHY integration and QoS area. Initial results achieved by this project are encouraging and will lead to full deployment of system soon.

[2007-2008] Implementation of WiFiRe MAC: Ranging, Registration and Classification modules: M Ranjith Kumar [co-guide: Prof. Anirudha Sahoo]

Download: Thesis: PDF file [1.1 M]. Slides: PDF [600 K].
Abstract: WiFiRe is an extension of the existing WiFi protocol for providing broadband access for rural areas. It provides good bandwidth at ow cost by making use of licence free spectrum at 2.4GHz band and also cost effective chipsets of WiFi (PHY). It mainly replaces the MAC of existing WiFi (802.11b) with a MAC of WiMAX (802.16). It follows Time Division Duplex-Multi Sector TDM (TDD-MSTDM) which uses same channel for both UL (up-link) and DL (down-link). It has star topology, and divides whole area into sectors; each sector has one base station (BS) with sectorized antenna, each village has one subscriber terminal (ST) with directional antenna and a System (S) located at fiber Point-of-Presence (PoP), which controls the whole network. WiFiRe has one single MAC for all sectors (PHY) in the System, which helps to co-ordinate the medium access.
Design of WiFiRe MAC was started last year and emulation over Ethernet was proposed. We implemented most of the functionalities for single sector WiFiRe MAC. Implementation is done using C sockets and it allows user to test basic MAC functions such as connection establishment, packet flow and header construction, in absence of WiFiRe hardware. Our code can be easily ported onto hardware. Our implementation supports various kind of applications like HTTP, FTP, VoIP etc. for Single BS, multiple Subscriber Terminals (ST) and their clients. Challenges encountered in implementation and their solutions are covered as well.

[2007-2008] Implementation of WiFiRe PHY sectorization in OPNET: Sreedhar Reddy

Download: Thesis: PDF file [1.0 M]. Slides: PDF [910 K].
Abstract: WiFiRe - WiFi Rural extension, is an extension of the existing WiFi protocol for providing long range broadband access for rural areas in cost effective way. As 802.11 MAC is not suitable for long range communications, WiFiRe replaces the MAC of existing 802.11b with MAC of 802.16 while retaining 802.11b PHY layer to extend the range. WiFiRe uses star topolgy and divides the region in to sectors, with each sector having one base station with sectorized antenna and one or more subscriber terminal with directional antenna. WiFiRe system has single MAC for all sectors, which helps to co-ordinate the medium access using multisector TDM.
In this thesis, we discuss the WiFiRe model design in OPNET. The PHY part of WiFiRe is implemented with OPNET using pipeline stages which describes the physical behaviour of wireless medium. The directional antennas in a six sectored system are modelled using OPNET antenna pattern editor. Some experiments are performed for range of sectorization and number of VoIP calls supported by ST in six sectored system to validate the model. Results shows the model is working as expected.

[2007-2008] MH-WiFiRe: Multi-hop extension for a WiFiRe system: Kedar Rudre [co-guide: Prof. Puru Kulkarni]

Download: Thesis: PDF file [2.2 M]. Slides: PDF [870 K].
Abstract: Cost effectiveness is an important criterion for any technology/system to be successful in rural regions. WiFiRe: WiFi Rural Extension based on 802.11 provides low cost broadband Internet access for rural regions and supports real time traffic like VoIP and Video. WiFiRe assumes star topology with its cell covering a circular region of 15-20kms in radius. Due to its single hop nature, it suffers from drawbacks like line of sight requirement and fixed coverage area. In this project, we extend WiFiRe to multiple hops and present multi-hop WiFiRe architecture which alleviates the drawbacks of WiFiRe.
We first present different architectures for multi-hop WiFiRe and compare them on different dimensions important from the persepective of feasibility of the system in rural region. We also present detailed cost and coverage analysis for some of these architectures. On the basis of this comparative analysis, we select an architecture based on tree topology for further studies. Our next contribution is in providing detailed description of the MAC protocol for the multi-hop system. We consider TDMA based MAC protocol influenced by 2P. We present MAC level frame structure and further give detailed procedures required to perform tasks of the MAC protocol. To perform time slots allocation, we propose three different scheduling schemes. We then perform comparative analysis for these three scheduling schemes.
The next contribution of our project is a tool to evaluate the performance of the multi-hop WiFiRe system. We compare performance of our system for different codecs. We present delay analysis, coverage analysis and capacity analysis of the system. Through this analysis we show that its not only feasible to have multi-hop extension but its also possible to perform trade-off between coverage and number of VoIP calls supported. The trade-off depends on type of codec and scheduling scheme. The final contribution of the project is the future work that can be done in the same domain.

[2006-2007] Shikav extensions to support networking animation: Moniphal Say [co-guide: Prof. Abhiram Ranade]

Download: Thesis: PDF file [900 K]. Slides: PDF [900 K].
Abstract: There is a growing interest in e-learning tools. One category of e-learning tools is electronic lesson. In addition to the lectures in classroom, electronic lessons can also be used, which help students in understanding the topics more completely by the use of animation as interactive visual aids. For this purpose, many authoring tools have been developed. Shikav 2.0 is one among them.
Shikav is a framework for constructing interactive electronic lessons, including animations using its own high level script language. The framework lets us create lessons in the field of Mathematics, Algorithms, Biology, Physics, and Economics. Shikav is developed in Java and has support for constructs such as point, line, arc, circle, triangle, rectangle, anglecurve, edge, vertex. However these are not sufficient to easily create animations of network protocols. Additional constructs need to be supported such as node, packet, beacon, repeat loops, if else conditional statement.
We have added modules, and made modifications to Shikav to support lessons explaining the basic concept of network protocols. In this thesis, a detailed description of the concepts, design, features of Shikav are presented. The design and the newly added features to Shikav have also been thoroughly described. We also explain a new defined script called "network script" language.

[2006-2007] Design and implementation of MAC layer of WiFiRe protocol: H Shravan Kumar [co-guide: Prof. Anirudha Sahoo]

Download: Thesis: PDF file [830 K]. Slides: PDF [300 K].
Abstract: WiFiRe (Wireless Fidelity Rural Extension) is an extension of the existing WiFi (802.11) protocol. The main aim of WiFiRe is to provide long range communications with high bandwidth, with low cost and easy availability of the chipsets. It mainly replaces the MAC mechanisms of existing WiFi (802.11b) so that it can be used for long range communication for about 15 – 20km, in contrast to existing technology which can support only upto a few hundreds of meters. It continues to use the existing Physical layer of WiFi (802.11b). WiFiRe MAC has many of the features similar to WiMAX (802.16).
WiFiRe provides long range communication by dividing the whole area into sectors, each sector having one Base Station (BS), which is a sectorized antenna. At each Sub- scriber Terminal (ST), a directonial antenna is used to connect ST to BS. WiFiRe uses only one channel for both uplink and downlink, in which each sector is allocated slots based on Time Division Multiplexing-Multi-sector TDM(TDM-MSTDM) mechanism.
In this report we have designed and implemented the WiFiRe protocol and emulated in over a LAN. Problems associated with design and implementation and their plausible solutions are covered as a part of this report. Additionally, it also comprises of sequence diagrams, flow diagrams and state diagrams of working components of WiFiRe. We have emulated the protocol using C sockets and this report covers the details of the same. This report also contains how the slots are being scheduled to STs by BS.

[2006-2007] Design and implementation of WiFiRe MAC layer protocol: Sameer Kurkure [Guide: Prof. Anirudha Sahoo]

Download: Thesis: PDF file [895 K]. Slides: PPT [235 K].
Abstract: Wireless Fidelity for Rural Extension (WiFiRE) is a MAC layer protocol especially designed to provide internet broadband facility in rural area in India. It supports long ranged communication using inherited 802.16d MAC and 802.11b PHY layer. WiFiRE uses low cost chip sets with capability to communicate on unlicensed spectrum to transmit over 15-20km.
The report covers the background knowledge of WiFiRe protocol with basic working of the protocol and its implementation in C. Problems associated with design and implementation and their plausible solutions are covered as a part of report. Additionally, it also comprises of sequence diagrams, flow diagrams, state diagrams etc. of working components of WiFiRE along with design model in C sockets and describes the issues and challenges involving implementation of the projects.

[2006-2007] Formal specification and verification of WiFiRe protocol: Ch Sudheer Keshav [co-guide: Prof. Krishna S]

Download: Thesis: PDF file [750 K]. Slides: ??
Abstract: WiFiRe stand for WiFi-Rural Extension[1]. There has been a lot of growth in Internet as well as cellular telephony, but it is mostly restricted to only cities in developing countries like India. A number of alternative solutions have been proposed which in theory are suitable for rural areas, but because of their high cost WiFiRe is the best option. WiFiRe is a mixture of the advantages of both 802.11[2] and 802.16[3]. The cost affordability of 802.11 and QoS capabilities of 802.16 are merged to obtain WiFiRe. The WiFiRe MAC is time-division duplex (TDD) over a single 802.11b channel along with a multi-sector TDM mechanism. WiFiRe will be able to provide long range communication by dividing the whole area into sectors, each sector will be having one base station which is directional so that it covers more distance.
Formal methods may be used to give a description of the system to be developed, at whatever level(s) of detail desired. This formal description can be used to guide further development activities; additionally, it can be used to verify that the requirements for the system being developed have been completely and accurately specified. Formal methods are mathematically-based techniques for the specification, development and verification of software and hardware systems. Formal methods and testing are a perfect couple, testing and formal verification are both necessary. A formally verified specification is a good starting point for testing. So a formal specification and verification of WiFiRe is required.

[2006-2007] Performance evaluation of WiFiRe using OPNET: Venkat Reddy M [co-guide: Prof. Varsha Apte]

Download: Thesis: PDF file [760 K]. Slides: PDF [700 K].
Abstract: The goal of WiFiRe is to provide broadband Internet services in the rural areas. The key idea in WiFiRe is to replace the 802.11b MAC mechanisms, with something more suitable for long-range communication, while using 802.11b PHY. WiFiRe network topology is a star topology, a System (S) with set of Base Stations (BS) with sectorized antennas at the fiber Point of Presence (PoP) and Subscriber Terminals in the surrounding villages with directional antennas. WiFiRe Cell is divided into sectors and each sector will be covered by a BS . WiFiRe has one common medium access (MAC) controller for all the sectors, to co-ordinate the medium access among them. There are multiple STs in each sector. The multiple access mechanism is time division duplexed, multi-sector TDM (TDD-MSTDM) scheduling of slots.
This protocol is modeled in OPNET. In this model we have implemented a Round Robin slot scheduling algorithm and a Connection Admission Control mechanism. We have simulated this protocol in different scenarios with different MAC configurations. We have simulated UGS, rtPS, nrtPS and BE service classes with different configurations, like changing polling time, changing slot length, etc. We have compared throughput and service delays of different service classes in different scenarios. By simulation results, we have determined the best slot length and maximum number of users supported by each BS for different type of service classes.

[2006-2007] Design of PSTN-VoIP gateway with inbuilt PBX & SIP extensions for wireless medium: Priyesh Wadhwa

Download: Thesis: PDF file [770 K]. Slides: PDF [630 K].
Abstract: VoIP gateway enables voice communication between users of the IP network and the Public Switched Telephone Network(PSTN). The system setup requires a PC installed with Asterisk, and a Gateway to integrate with PSTN. The problem with this solution is the high cost, power consumption, and the involved setup of the system. We have designed a single box solution for the PSTN-VoIP integration system. We studied detailed architecture of the gateway, the protocols used in the VoIP call setup and communication, the software used for the PBX systems, and the internal parts of the SPA3000 gateway. We used the Via motherboard, flash memory, and a normal data modem to create a improved and cost-effective system.
Next, we performed various studies on the Asterisk’s response time in wired and wireless medium. We found a remarkable difference in the response time of Asterisk for both the mediums. The reason for the high response time of Asterisk in wireless medium is the slow call setup. SIP being a text-based protocol, is engineered for high data rate links, and so SIP message’s size have not been optimized. With low bit rate IP connectivity of signaling channel, the transmission delays for call setup and feature invocation are significant.
We have extended the Session Initiation Protocol for improving its efficiency in wire- less medium. We have implemented the compression and decompression mechanisms according to the SigComp standard, and integrated them to the Asterisk server. We have also developed a SigComp enabled client to run with Asterisk. We performed extensive testing of the system and obtained upto 90% compression using SigComp and Deflate compression algorithm.

[2006-2007] Design of PSTN-VoIP gateway for rural environments: K Sravana Kumar

Download: Thesis: PDF file [1.2 M]. Slides: PDF [1.3 M].
Abstract: A Voice over IP(VoIP) gateway is one of the various components used in realizing the convergence of data and telephone networks. It enables voice communication between callers on the IP network and the Public Switched Telephone Network(PSTN). There are many vendors that provide VoIP-PSTN gateways although most of them are very expensive. Soft PBX is a software-based PBX solution using VoIP technology. Asterisk is an example of a Soft PBX. If we want to maintain a VoIP-based telephone exchange, the cost of the entire system proves to be very expensive and it is not suitable for rural deployment. Also most of the devices require AC power which is not easily available in rural areas. There is currently no device in the market which provides the functionalities of both gateway and PBX and yet operates on DC power. Our aim is to design such a box which performs as a VoIP gateway with inbuilt Asterisk PBX and at the same time consume less power and be relatively inexpensive.
Asterisk has its own authentication for SIP users. Generally, large organizations and universities maintain some external authentication for their employees/students such as LDAP. Many of these organizations provide VoIP telephony facilities to their people. So, for unique authentication of users, there is a need to provide an external authentication mechanism in Asterisk, which can interface with the authentication system in use at the organization. So, one more aim of this project is to integrate Asterisk with LDAP.

[2005-2006] HSM: A hybrid streaming mechanism for delay-tolerant multimedia applications: Annanda Rath

Download: Thesis: PDF file [928 K]. Slides: PPT [1.5 M].
Abstract: Streaming is a technique for transferring data such that it can be processed as a steady and continuous stream. In a content dissemination network, typically a central server at the source streams the content in response to client requests. We term it as Pure Streaming Mechanism (PSM). Considering that in a dissemination network controlled by a Content Service Provider (CSP), the backbone links are highly provisioned, using a streaming server at the source leads to underutilization of these links. Also the links are occupied for the duration of play out of the multimedia content. This is because according to the streaming property, the streaming server sends only the amount of data equivalent to the streaming encoded rate to the client irrespective of the available bandwidth in the path. Delay-tolerant applications are a special class of on demand multimedia applications where clients request the start of play back at a time specified by (t +dti) where t is the current time and dti is the delay tolerance acceptable to client i.
In this thesis report, we present a novel Hybrid Streaming Mechanism (HSM) for Delay-Tolerant Multimedia Applications to enhance the following performance parameters: (i) number of serviced clients, and (ii) delivered stream rate at clients. HSM allows streaming from strategically chosen intermediate nodes to which the content is dynamically transferred from the source, using FTP (File Transfer Protocol). As FTP uses the entire bandwidth to transfer the data, it frees up the high bandwidth links faster for serving requests from other clients sharing these links, increasing the efficiency of the service. In HSM, transferred contents are temporally cached at the intermediate node (Streaming Point). Such temporary caching further enhances the performances of HSM as requests for the same content are serviced from the cache.

[2005-2006] Video transmission over varying bandwidth links: Laxmikant Patil

Download: Thesis: PDF file [480 K]. Slides: PPT [415 K].
Abstract: Internet today is a best-effort network, with time-varying bandwith characteristics. A system for multicasting video over the Internet has to deal with heterogeneity of the receiver’s capability and/or requirements. So adaptive mechanisms are necessary. Con- ventionally multi-layered transmission of data is preferred to solve the problem of varying bandwidth in multimedia multicast application. In todays scenario majority of the flows are highly bandwidth consuming and are bursty in nature. Consequently we observe sudden changes in available bandwidth, leading to lot of variations in received video quality.
Here our aim is to minimize these variations in video quality. In this thesis we propose traffic pattern based adaptive multimedia multicast (TPAMM) scheme.In this scheme, link bandwidth values are available for prediction window length. Bandwidth prediction model is refined periodically using feedback from receiver. TPAMM scheme also maxi- mizes the minimum quality video playout.

[2005-2006] Application partitioning - A dynamic, runtime, object-level approach: Anshu Veda

Download: Thesis: PDF file [1.2 M]. Slides: ??
Abstract: With the advent of increase in the computational complexity of the programs, it seems conducive to distribute a centralized program written for a standalone system onto a network of nodes. Application partitioning is one such technique, that aims at division (allocation) of application components amongst different hosts in the network, thereby getting a standalone application executed in a distributed setting.
Most of the existing work in application partitioning, uses classes as the basic component for partitioning. However, the behavior of an application is determined by the interaction of its entities at run-time. In an object-oriented program, the run-time entities are principally objects. We believe that, a more effective and relevant partitioning mechanism can be created by considering objects as the basic components for distribution. We also believe that, a truly flexible partitioning system should as far as possible, postpone the decision regarding the placement of each component to run-time. Any such decision should try to optimize on both - the number of remote invocations that a distribution strategy would result in, as well as the load distribution on the available hosts.
We thus propose a working architecture that exploits the dependency relationships between the components, to build a model prior to the program execution. At run-time, the model progressively incorporates the information about already allocated components and helps in deciding the position of new component.

[2005-2006] Proxy-AODV: Extension of AODV for partially connected Adhoc networks: Anshuman Tiwari

Download: Thesis: PDF file [520 K]. Slides: PPT [500 K].
Abstract: Ad hoc on-demand Distance Vector (AODV) is a routing scheme for delivering messages in a connected Mobile Ad hoc Network (MANET). In MANETs, a set of nodes are used to route the data on behalf of other nodes. This scheme relies on the assumption that nodes are distributed over the entire region and there exists connectivity between any source-destination pair in the network at all times. This scheme fail when the network is partially connected, i.e. when there is no single-hop or multi-hop path from source to destination.
The existing schemes for routing in partially connected ad hoc networks make assumptions like, source and destination never have a connected path, a set of mobile nodes with fixed route deliver the data, large storage space at nodes, etc. All these assumptions may not hold true in a resource constrained ad hoc network. We propose an extension to AODV, named proxy-AODV to remove the above assumptions. In situations where the network is connected our protocol behaves like normal AODV. When there is no connected path, we exploit mobility of nodes and use “store and forward” approach to deliver the data.
In proxy-AODV when source and destination are not connected. Then some of the nodes called proxy nodes are selected by source to hold the data on behalf of destination. Proxy nodes acts as source and try to deliver data to the destination. Using extensive simulations in QualNet simulator, we show that our protocol provides good message delivery ratio while keeping the buffer occupancy at nodes under check.

[2005-2006] Implementation of WiFiRe model in OPNET: Anirudha Bodhankar

Download: Thesis: PDF file [630 K]. Slides: ??
Abstract: The 802.11(WiFi) family of wireless communication is extremely popular for indoor wireless networking. The chipsets are designed so that PHY and MAC layers are separate and they are available off-the-shelf at low cost. Also it operates in 2.5GHz ISM band which is license free spectrum avoiding licensing cost, hence is an attractive option for long range communication in rural areas of developing countries.
However many problem occur when using 802.11 for outdoor long range (15-20 Km) communication. The DCF mode does not support any quality of service, while PCF is not suitable for large number of stations and long distance. The MAC is based on CSMA/CA and is not efficient when the number of stations increases. Hence it is necessary to redesign MAC while retaining the PHY. One approach for this is WiFiRe.
WiFiRe replaces the MAC of 802.11 with the MAC similar to that of 802.16 MAC which is suitable for long range communication. It also uses directional antennas and is meant for a star topology. WiFiRe MAC at the Base Station is multi-sector MAC which controls more than one 802.11 directional PHY.
The goal of this project is to implement the WiFiRe model in OPNET and analyse its performance. The code developed is to be such that, it can be easily extended to a real implementation.

[2005-2006] RFIDplanner: A coverage planning tool for RFID networks: Nitesh Gupta

Download: Thesis: PS file [16 M]. PDF file [1.6 M]. Slides: PPT [11 M].
Abstract: Radio Frequency Identification or RFID systems are emerging as economical solutions for fast identification of objects. One of the important aspects of setting up a RFID network is deploying RFID readers to ensure complete coverage of RFID tags in a given area. This task is usually accomplished by conducting site surveys and then deploying RFID readers in different configurations to determine the optimal one. As the size of the given area increases, time taken to setup a RFID network increases thus increasing the cost of deployment.
We present RFIDPlanner, a coverage planning tool for RFID networks, which attempts to tackle the issue using simulation. Given the details of an area, the properties of obstacles present in the area, and the specifications of the RFID reader and tags used, RFIDPlanner simulates the coverage pattern of a reader giving a graphical view of the RFID reader's interrogation range. Through its interactive GUI the layout can be modified during run-time, by moving and deleting existing RFID readers and adding new ones. The zoom feature allows different levels of visualizations ranging from a view of the entire layout to focused views of particular sections and RFID readers.

[2005-2006] Improving RFID systems to read tags efficiently: Naval Bhandari [Guide: Prof. Anirudha Sahoo]

Download: Thesis: PDF file [6.3 M]. Slides: ??
Abstract: Radio Frequency IDentification (RFID) is slated to become a standard for tagging various products. As more and more products become RFID enabled, fast tag identification mechanisms will become important. Various tag identification (or anti-collision) algorithms have been proposed for RFID systems.
This work focuses on methods to improve tag read efficiency in RFID Systems. We propose an Intelligent Query Tree (IQT) Protocol for tag identification that exploits specific prefix patterns in the tags and makes the identification process more efficient. IQT is a memoryless protocol that identifies RFID tags more efficiently in scenarios where tag IDs have some common prefix (e.g., common vendor ID or product ID). IQT is most suitable for readers deployed in exclusive showrooms, and shipment points of big malls, where the products may come from same manufacturers and may have same product type. We provide the worst case complexity analysis of IQT and show the performance improvement of this protocol over traditional Query Tree protocol in different scenarios. For other cases we show the improvement using simulation results.

[2004-2005] PLUS-DAC: An admission control scheme for IEEE 802.11 Wireless LANs: Kiran Kumar Gavini [Guide: Prof. Varsha Apte]

Download: Thesis: PDF file [460 K]. Slides: PDF [640 K].
Abstract: With increasing demand of support for realtime applications, there is a compelling need for Quality of Service (QoS) in present day wireless LANs. IEEE 802.11 Medium Access Control (MAC) has become a defacto standard for wireless LANs, but there are many inherent QoS limitations in the base standard, as it was basically developed for best effort data services.
The 802.11 Task Group E (TGe), is about to ratify a QoS extension to the base 802.11 standard namely 802.11e. The 802.11e standard provides many mechanisms for QoS support at the MAC layer level. However, even the service differentiation provided in 802.11e is not enough to meet the QoS requirements of time bounded multimedia traffic at high load. These can be better satisfied, if we employ Admission Control and Bandwidth Reservation mechanisms.
Another important concern in WLANs is channel utilization. Generally, partitioning based reservation schemes do static division of bandwidth. When bandwidth is divided statically, often, more bandwidth can get allocated to a category which is currently not offering much traffic to the network, resulting in under-utilization of the bandwidth resources. More over, the bandwidth partitioning should not be purely based on the priority of the traffic.
We propose a measurement based distributed admission control mechanism, for the 802.11e Wireless Local Area Network (WLAN) functioning in infrastructure mode. We call the scheme PLUS-DAC (Priority, Load and Utilization based Scheme for Distributed Admission Control). PLUS-DAC measures the load and utilization in the BSS and adapts the Transmission Opportunity (TXOP) reservation dynamically. Our results show that, PLUS-DAC can achieve quasi-optimal utilization and continue to satisfy QoS guarantees given to multimedia flows.

[2004-2005] An Efficient QoS Scheduling Architecture for IEEE 802.16 Wireless MANs: Supriya Maheshwari [co-guide: Prof. Krishna Paul]

Download: Thesis: PDF file [590 K]. PS file [2128 K]. Slides: PPT [1116 K].
Abstract: IEEE 802.16 WirelessMAN standard specifies air interface of a fixed point-to-multipoint Broadband Wireless Access. IEEE 802.16 MAC provides extensive bandwidth allocation and QoS mechanisms for various types of applications. However, details of packet scheduling mechanisms for both downlink and uplink direction are left unspecified in the standard.
We propose an efficient QoS scheduling architecture for IEEE 802.16 WirelessMANs. Our main design goals are to provide delay and bandwidth guarantees for various kinds of applications and to maintain fairness among various flows while still achieving high bandwidth utilization. Our architecture supports diverse QoS requirements of all four kinds of service flows specified in IEEE 802.16 standard.
We have implemented IEEE 802.16 MAC integrated with our architecture in Qualnet 3.6 network simulator. We also present the simulation analysis of our architecture. We have shown, through simulation, that our architecture is capable of achieving high bandwidth utilization. Simulation results also show that sufficient bandwidth is allocated to high priority flows so that their QoS guarantees are always met.

[2004-2005] Mitigating the Reader Collision Problem in RFID Networks: Shailesh Birari

Download: Thesis: PDF file [1204 K]. Slides: PPT [1756 K].
Abstract: Radio Frequency Identification (RFID) is a means to identify and track objects using radio frequency transmission. An RFID system consists of readers and tags. Readers use radio signals to communicate with the tags. Tags may be active (battery powered) or passive (powered by the reader's signals). RFID is increasingly being used in many applications such as inventory management, object tracking, retail checkout etc. The reader collision problem occurs when the signal from one reader interferes with the signal from other readers. Such interference can result in lack of communication between the readers and some of the tags in the vicinity leading to incorrect and inefficient operation of an RFID system. This problem is further aggravated when mobile/hand-held readers are used in the system. Hence efforts are required to minimize this interference.
We describe Pulse, a distributed protocol to reduce these reader collisions in the RFID systems. The operation of the Pulse protocol is based on periodic beaconing on a separate control channel by the reader, while it is reading the tags. The protocol functions effectively not only with fixed RFID readers but also with mobile RFID readers. We show, using simulation in QualNet, that using Pulse protocol, the throughput (overall read rate) is increased by 98%(with 49 readers) as compared to ``Listen Before Talk''(CSMA) and by 337% as compared to Colorwave (with 9 readers). We also present an analytical model for our protocol in a single hop scenario.

[2004-2005] RFIDcover: A coverage planning tool for RFID networks with mobile readers: Anusha S.

Download: Thesis: PDF file [1444 K]. Slides: PPT [1227 K]. Demo: ZIP [80 K].
Abstract: Radio Frequency Identification (RFID) finds use in numerous applications involving item identification and tracking. In a typical application, RFID tags are attached to the items and are periodically queried by readers. These applications require that a sufficient number of readers be deployed in order to provide complete coverage of all the tags in the given area.
Using a fixed placement of readers to guarantee complete coverage at all times increases the deployment costs. Also, most practical applications do not need complete coverage at all times. It is enough to provide complete coverage periodically, say each tag being covered every T seconds. For such applications, using mobile readers to cover the area would be more cost-effective.
Given an area to be covered completely within a period T, determining the number of mobile readers required, their placement and movement pattern, is a difficult problem. We have designed and developed RFIDcover, an automated coverage planning tool, that addresses this problem. Given an application scenario and reader specifications, RFIDcover determines an optimal number of readers required to guarantee complete coverage within the specified period T. It also generates a layout giving the placement and movement pattern of the readers. In this thesis report, we present RFIDcover and its implementation for a retail inventory tracking application scenario and evaluate its effectiveness.

[2004-2005] M2MC: Middleware for Many to Many Communications over Broadcast Networks: Chaitanya Krishna B.

Download: Thesis: PDF file [2272 K]. Slides: PPT [544 K].
Abstract: M2MC is a new distributed computing middleware designed to support collaborative applications running on devices connected by broadcast networks. Examples of such networks are wireless ad hoc networks of mobile computing devices, or wired devices connected by a local area network. M2MC is useful for building a broad range of multi$-$user applications like multiplayer games, conversations, group ware systems.
M2MC architecture consists of Messages Ordering protocol, Member Synchronization protocol, and protocols for processes to join and leave the groups. We emphasized on Message Ordering protocol as it is a key component in developing group communication applications. Hence we proposed a new message ordering called $S_b$ ordering that orders the messages based on their semantic relationship as specified by the users.
Some salient features of M2MC are: Unlike existing middleware architectures that rely on central servers, the M2MC is truly distributed protocol and hence application developed using M2MC does not require central servers. Being broadcast oriented, M2MC does not require any resource consuming routing protocols. Distributed applications development is simplified by M2MC APIs.

[2003-2004] QoS Guarantees for Real-Time Applications in 802.11 WLANs: Kiran Diwakar

Download: Thesis: PDF file [815 K]. Slides: PPT [415 K].
Abstract: IEEE 802.11 standard specifies the MAC and PHY layers of the OSI Seven Layer Model for Wireless LANs. 802.11 is by far the most widely used and popular of the suite of WLANs. Channel access is both distributed as well as centralized called DCF and PCF respectively.
The expansion of IEEE 802.11 based WLANs has created interest in providing Quality of Service (QoS) guarantees in such networks. 802.11e suggests a flow-based solution based on priority queues at the node, for providing QoS guarantees. We propose a different approach, using a variant of TDMA, called Dynamic Time-division Multiple Access (DTMA). It is based on the observation that the ratio of data transmission time to control packet (poll or acknowledgement) transmission time is typically 6:1, and drops down almost to 2:1 when the data packet size becomes smaller than 500 bytes. DTMA focusses on reducing control information and thereby increasing time available for data transmission. More time available for data transmission implies increased throughput.
We expunge the overhead of control packets in the HCF by using cummulative acknowledgements and piggybacking.Taking advantage of 802.11's inherent limited range, DTMA, the modified TDMA is used for data transfer in HCF, without encountering the problem of distributed time synchronization. In this dissertation, we do the analysis of 802.11e and DTMA using probability models and also simulations to support these models. Thus, aided by simulations and analytical methods, we prove that DTMA has stricter and lesser delay bounds than 802.11e, for real-time and QoS sensitive applications. We also show that DTMA enhances the overall throughput by almost 20\%, thereby implying better QoS guarantees in the 802.11 domain.

[2003-2004] Lookup service for peer-to-peer systems in mobile adhoc networks: Kalpesh Patel [co-guide: Prof. Krishna Paul]

Download: Thesis: PDF file [448 K]. PS file [572 K]. Slides: PPT [477 K].
Abstract: A Peer-to-Peer(P2P) system consists of a set of nodes in which every node shares its own resources and services among other nodes in the system. Any node can make query for data, currently available in the network. These peer-to-peer systems are basically overlay networks, work on top of fixed network infrastructure. With the increasing popularity of Mobile Ad-hoc Networks (MANET), it becomes necessary to think of P2P systems which can efficiently work in mobile environments.
Current P2P systems do not fit well into mobile environments where each node have to follow multi-hop path routing and have to face unpredictable mobility. If current P2P systems are deployed on the MANET, application layer and network layer both perform different routing mechanism independent of each other, resulting in multiple layer routing redundancy. Also, routes at the application layer may not be necessarily optimal at the network layer. Thus, we bring down all routing and query lookup mechanism to network layer. Further, current ad-hoc network protocols do not fulfill the P2P requirements as basically they all are flooding based protocols. Here, we address this problem and propose a novel, controlled-flooding based approach which works at network layer and helps P2P systems to work in mobile environment efficiently.
RINGS is a protocol which implements this approach.RINGS performs query lookup at network layer rather than at application layer, thus eliminating application layer routing overhead. It also reduces query lookup cost by evenly distributing data indices throughout the network, thus reducing network layer routing overhead.

[2003-2004] Performance Comparison of DAMA MAC Schemes over Satellite Networks: Kalyan Rao D.

Download: Thesis: PDF file [943 K]. Slides: PDF [141 K].
Abstract: Satellite networks provides wide coverage of geographic area and high bandwidth. As satellite capacity is often limiting resource which must be used efficiently. Internet traffic is highly bursty in nature and Demand Assignment Multiple Access (DAMA) techniques are suitable provide bandwidth to match instantaneous requirements and provide significant improvements in the delay/utilization performance of Geo-stationary Earth Orbit (GEO) satellite channels supporting a finite number of users with bursty data traffic. So, the performance of various DAMA MAC schemes is important.
In this thesis, we investigate the performance of CFDAMA and BTDAMA protocols on satellite networks and proposes an extention to BTDAMA, User Prioritized-BTDAMA (BTDAMA-UP) which implements Quality of Service (QoS) by user prioritization so that different group of users will get different level of service.

[2003-2004] MSIP: A protocol for efficient handoffs of real-time multimedia in mobile wireless scenarios: Ranjeeth Kumar A.

Download: Thesis: PDF file [367 K]. PS file [2296 K]. Slides: ??
Abstract: The support for IP mobility has become very important with the increasing growth of mobile applications. One of the key challenges for the deployment of such wireless Internet infrastructure is to efficiently manage user mobility. Mobility is handled by Mobile IP at the network layer and Session Initiation Protocol (SIP) at the application layer.
The main aim of these two schemes is to provide seamless connectivity for ongoing communications, that is, to keep the handoff latency delay as less as possible. The delay constraint becomes even more crucial for real-time applications. The Mobile IP scheme has drawbacks like triangular routing, longer delays and the need for tunneling management. SIP solves many of the problems in Mobile IP, but it involves the time consuming process of obtaining a new IP address from the DHCP server.
In this report, we propose a hybrid scheme, MSIP, that is the integration of Mobile IP and SIP. MSIP reduces the handoff delay for real-time applications. It avoids the need for tunneling the packets throughout the handoff period and the communication need not be suspended till a new IP address is allocated by the DHCP server. Thus MSIP gives better performance than Mobile IP and SIP, for real-time applications.

[2002-2003] Design and deployment of file transfer protocol over asymmetric satellite networks: Anupam Goyal

Download: Thesis: PDF file [1479 K]. Slides: PPT [1.5 M].
Abstract: Satellite networks are the most widely deployed networks for audio and video transmissions. Though the initial installation cost is more, the benefit of reaching users spread over very large areas makes it a viable and very effective option for sending information in which some loss is tolerable. Some of the main drawbacks of satellite networks are limited bandwidth availability, large delay involved in transmission and their asymmetric nature. This leads to still higher delay in correcting the losses occurred during transmission making them difficult to use for reliable transmission. But when the need for sending data reliably to multiple users spread across a wide region arises, the long reach and inherent multicast nature of satellite networks make them an interesting option.
Our work explores the possibility of using asymmetric networks like satellite networks for reliable data transmission to multiple users. An asymmetric network consists of a comparatively large bandwidth unreliable communication channel coupled with a low bandwidth reliable communication channel. We have designed a protocol which can be used to transmit data reliably over above defined asymmetric networks. We have implemented the protocol in developing a file transfer utility for Distance Education Program, IIT Bombay which uses a satellite network of the above kind. The algorithm is general enough to be employed for reliably transferring data over any asymmetric network though the change in network parameters will affect the protocol efficiency.

[2002-2003] Dynamic adaptation of the DCF and PCF mode in IEEE 802.11 WLANs: Abhishek Goliya

Download: Thesis: PDF file [671 K]. Slides: PPT [1.6 M].
Abstract: IEEE 802.11 specifies the most famous family of WLANs. It features two basic mode of operation: Distributed Coordinating Function (DCF) and Point Coordinating Function (PCF). Both PCF and DCF mode of IEEE 802.11 do not perform equally well under all traffic scenarios. Their behavior varies depending upon current network size and traffic load. It is useful to use the DCF mode for low traffic and small network size, and the PCF mode for high traffic loads and to reduce contention in large size network.
In this thesis, we have designed three protocols to dynamically adapt IEEE 802.11 MAC under varying load. One of them is designed to dynamically switch between either mode. Our Dynamic Switching Protocol (DSP) observes network traffic to decide switching point and switches dynamically to suit current traffic load and network size.
PRRS is our second contribution that aims to reduce polling overheads. A major drawback of polling scheme in PCF, is their inefficiency when only a small number of nodes have data to send. Unsuccessful polling attempts cause unnecessary delays for station with data. We have presented network monitoring based scheme that replaces simple Round Robin scheduling in PCF with our Priority Round Robin Scheduling (PRRS). Result shows considerable increase in throughput especially when small fraction of node has data to transmit.
In addition, we have presented the need to dynamically adapt various configuration parameters in both PCF and DCF. Statically configured values results in degraded performance under varying scenarios .We have showed the performance variation of PCF with PRRS by using different CFP repetition intervals. Our proposed CFP repetition interval adaption algorithm dynamically adjusts the value of CFP repetition interval, depending upon last CFP usage.

[2002-2003] Design and evaluation of IEEE 802.11 based Dual MAC for mobile adhoc networks: Satyajit Rai

Download: Thesis: PDF file [350 K]. Slides: PDF [77 K].
Abstract: Multihop ad-hoc wireless networks offer great challenges for protocol designers. Stations in such networks are constrained by factors like low power, limited bandwidth, link errors, and collisions. Changes are needed at various levels of the protocol stack, most importantly at the medium access layer (MAC). The medium access mechanism in multihop wireless networks should minimize collisions, and take care of the hidden and exposed node problems. The IEEE 802.11 MAC with Distributed Coordination Function (DCF) does not scale well in such networks. We introduce Point Coordination Function (PCF) in the region of high traffic areas, and discuss its effect on network performance.
To improve network scalability and throughput, we propose the design of a new MAC called Dual MAC. This work discusses architecture and working of the dual MAC in detail. Performance results of the network using dual MAC are presented, and compared with that of pure DCF operation.

[2002-2003] Tackling the exposed node problem in IEEE 802.11 MAC: Deepanshu Shukla

Download: Thesis: PDF file [1456 K]. PS file [1611 K]. Slides: PPT [338 K].
Abstract: Ad hoc wireless networks promise convenient infrastructure-free communication. An efficient MAC protocol through which mobile stations can share a common broadcast channel is essential in an ad-hoc network because the medium or channel is a scarce resource. The basic medium access mechanism is basically a Carrier Sense Multiple Access with Collision Avoidance mechanism (CSMA/CA). The CSMA/CA protocol is designed to reduce the collision probability between multiple stations accessing a medium. 802.11 MAC forestalls the possibility of feasible parallel communication by two neighboring nodes that are either both senders or both receivers giving rise to exposed Node problem.
Exposed Nodes are the nodes which are forced to defer the transmission of data as it is already exposed to an ongoing tranmission. This work adds enhancements to IEEE 802.11 MAC which enables it to schedule concurrent transmissions, thereby improving the channel utilization and solving the Exposed Node problem

[2001-2002] Video streaming in mobile environments: Manoj Kumar C.

Download: Thesis: PDF file [481 K]. PS file [559 K]. Slides: PDF [225 K].
Abstract: Video Streaming refers to the real-time transmission of stored video. In order to control network congestion, rate control schemes adjust the output rate of the video stream to the estimated available bandwidth in the network. Loss based rate control schemes like AIMD, TFRC employ the packet loss rate reported by the receiver as the principal feedback parameter to estimate the state of the network. In a heterogeneous network, where the streaming server is located on the wired network and the client/receiver is located on the wireless network, due to the high-burst error rates of the wireless channels, the loss rate reported by the receiver may not be a correct indicator of congestion in the network. Especially during bad wireless channel conditions, when the channel error rate is high, the loss rate reported will be high. Hence, loss based rate control schemes may inaccurately estimate the state of the network and respond by decreasing the output rate which will affect the quality of video received by the clients.
Two schemes have been proposed to alleviate the above problem. In ROLC Report-congestion-losses scheme, the receiver is made to report only fraction of the packets lost due to congestion. Due to this, the errors in the wireless channel cannot affect the rate control and quality of video as these errors are not reported to the sender. In Report-correlation-of-loss-and-delay, besides loss rate, the correlation of the packet loss and delay during a feedback interval is reported. This parameter is used by the sender in addition to the loss rate to respond to only congestion.

[2001-2002] Route repair in mobile adhoc networks: Abhilash P

Download: Thesis: PDF file [345 K]. PS file [998 K]. Slides: PDF [343 K].
Abstract: Mobile Adhoc Networks (MANET) are distributed, mobile, wireless, multihop networks that operate without pre-existing communication infrastructure, except for the mobile devices themselves. Several routing protocols both reactive and pro-active have been proposed to provide the self starting behavior needed for adhoc networks.
The nature of the network coupled with the mobility of the devices, result in a large number of route breakages. The current approach in case of broken routes is to flag an error and re-initiate route discovery either at the source or at the intermediate node. Repairing these broken links is a costly affair in terms of routing overhead and delay involved.
In this report, we propose a proactive approach called Routing Handoff, to repair broken routes, using the mobile devices in the vicinity of the broken link. The idea is incorporated into the AODV routing protocol. The results of the simulation indicate an increase in throughput under certain conditions. The improvement is a result of smaller overhead and delay. The approach may also be applied to other routing protocols with appropriate modification.

[2001-2002] Security issues in mobile agents: E.C. Vijil

Download: Thesis: PDF file [233 K]. Slides: PDF [100 K].
Abstract: An autonomous mobile agent is an executing program that can migrate from machine to machine in a heterogeneous network under its own control. An agent can either follow a pre-assigned path on the network or determine its itinerary based on the data collected from the network. Facilities for highly dynamic movement of code and data enables a program to take advantage of the locality of data. It also allows one to optimize between the requirements of low bandwidth, high latency and disconnected network connections.
This computing paradigm which exploits code, data and state mobility raises many new security issues, which are quite different from conventional client/server systems. Agent servers which provide an execution environment for the agents to execute can be attacked by malicious agents. Similarly agents could be carrying sensitive information about their owners and should be protected from tampering by malicious hosts. Also, the data collected by the agent from one host should be protected from tampering by another host in the itinerary.
In this report, we examine the various security issues that arise in mobile agents in general with special reference to data collection agents. We propose an algorithm to identify the malicious host modifying the data in data collection agents. Multiple hosts can collude to remove the data collected by the agent from previous hosts. We give a probabilistic collusion detection algorithm to detect deletion of data by colluding malicious hosts.

[2001-2002] ATCP: Adapted TCP for mobile wireless environments: Ajay Kumar Singh

Download: Thesis: PDF file [284 K]. PS file [357 K]. Slides: PDF [190 K].
Abstract: Transmission Control Protocol (TCP) is tuned to perform well in wired networks, but suffers from performance degradation over mobile wireless networks. This is due to distinct characteristics of these networks that packet losses also occur by high bit error rate and intermittent disconnection induced by mobility, and TCP misinterprets these losses as a indication of congestion.
Several approaches have been put forth by research community for alleviating the detrimental effect of poor characteristics of wireless channel, but only few approaches are presented for solving the mobility related issues. Even in these attempts, mostly requires support from base-station like some state per TCP connection. This requirement of these approach make them undesirable as these schemes introduced scalability difficulty, hinder the inter-operability of the mobile host with different networks, and also could not support encrypted IP traffic, different acknowledgement path.
All these observations motivated us to design of a new approach called ATCP. ATCP falls into the category of those approaches which mitigates the degrading effect of mobility while requiring modification only at mobile host like 3DA, Freeze TCP approaches. We also noticed that the authors of 3DA and Freeze TCP approach have not given any specific action to be taken when mobile host is sender. ATCP is designed for enhancing the performance of both direction data transfer i.e. mobile host (MH) to fixed host (FH) and MH to FH. ATCP scheme modified the TCP at mobile host end and requires feedback about the ongoing mobility from network layer in terms of connection and disconnection signal.
ATCP has been compared with 3DA, Freeze TCP and TCP Reno over network simulator ns-2. It improves the performance of TCP data transfer in both direction i.e. FH to MH and MH to FH. For data transfer from FH to MH, ATCP is shown to reduce response time as 3DA approach does, but it also achieves higher TCP throughput in most cases. It has been shown that for WLAN environments ATCP and Freeze TCP performs equally well, while in WWAN environments, ATCP is achieving same throughput as Freeze TCP for short disconnection period. ATCP has a significant advantage over Freeze TCP that it does not require prediction of impending disconnections. Another drawback of Freeze TCP shown that the throughput over the connection is sensitive to the prediction period. Simulations show that an improvement of up to 40\% is achieved in WLAN environments, and up to 150\% in WWAN environments.

[2000-2002] Distributed intrusion detection systems: Mamata Desai

Download: Thesis: PDF file [420 K]. PS file [951 K]. Slides: PPT [291 K].
Abstract: As more and more data goes online, there is a pressing need to secure the dissemination of a large amount of information. Because of the effort required to monitor networks and systems manually, it is not easy to detect attempts at misuse or successful attacks without the help of intelligent Intrusion Detection Systems (IDS). IDS, much like the security industry, has grown rapidly over the past few years. These tools have become essential security components - as valuable to many organizations as a firewall. However, as in any environment, things change. Networks and crackers are evolving fast, demanding that security tools keep up. Intrusion Detection Systems face several daunting, but exciting challenges in the future and are sure to remain one of our best weapons in the arena of network security.
The modern day Network IDS faces some very challenging problems, like switched environments, increased network traffic, and encryption. Add to that, the performance considerations of an IDS, such as false positives and missed attacks, and the mole hill does become a mountain! The way to go seems to be analysis and data correlation, in which, host IDSs also play an important role. The concept of a management console dedicated to the task of correlating abnormal event notifications, with relevance measures is an emerging one. One can picturize many distributed elements performing specific jobs, each passing the results onto a higher level for correlation and analysis. In an environment where many machines have similar configurations, a complete portscan on one machine may trigger alarms but slow scans across ports of different machines might go unnoticed and will result in the intruder gaining all the information about the services running on each machine, thus successfully performing a "distributed portscan".
We focus on detecting a "distributed portscan", by sniffing packets on the network. Five types of TCP portscans, performed by "nmap" are successfully detected, in scan sweeps of one-to-one, one-to-many, many-to-one and many-to-many hosts. Our approach also manages to detect slow scans which are typically missed by available commercial packages, because of the features that we select to examine.

[2000-2001] Extending the ns2 simulator for GPRS support: Richa Jain

Download: Thesis: PDF file [742 K]. PS file [1668 K]. Slides: PPT [220 K].
Abstract: Enabling wireless Internet access at data rates comparable to wired networks, is a growing concern in recent years. One attempt at this, is the General Packet Radio Service (GPRS), a packet based extension to GSM. The packet switched radio transmission enables efficient multiplexing of radio resources and data rates of up to 170 kbps can be achieved.
GPRS (General Packet Radio Service) Networks are currently being deployed in the market, and there is a need to study performance and network related issues. Simulations can provide a basis for evaluation and key decisions for deployment. In this project we undertake design and implementation of a GPRS simulator. We focus on the handling of the air interface and the Link Layer, Radio Link Layer and the Medium Access Layer for the Mobile Station - Base Station interaction. We use the ns-Network Simulator as a base for the implementation.
Using the simulator, we then study the effect of load conditions on the average packet delay and the number of users supported by a GPRS system vs a GSM system.

[2000-2001] Distributed archival and retrieval of medical images using DICOM: Dr. Sheikh Mahmood

Download: Thesis: PDF file [1605 K]. PS file [5378 K]. Slides: PPT [970 K].
Abstract: The introduction of digital imaging devices in the medical world has made the management of images easier and more efficient. The Digital Imaging and Communications in Medicine (DICOM) standard promotes the communication of digital image information regardless of device manufacturer. Digital Images in a hospital are managed with Picture Archiving and Communication System (PACS), which receives, stores and makes these images available to Radiologists and Physicians for diagnosis and review. The interface provided to the user, by PACS is proprietary, with each vendor having separate user interface.
This report describes the implementation and usage of World Wide Web interface to a DICOM based PACS that allows the user to query, select, move and display the images available in DICOM archive. The system supports querying of archive from any workstation (such as Windows, Mac) that supports a WWW browser.
In order to use the system, the user runs a WWW browser (such as Netscape, Internet Explorer) and specifies the URL of the server machine. On receiving the request, the server sends a query form to be displayed on the client. The query form contains three text input fields for patient name, patient ID and birthdate. The user may specify any or all the fields as well as wildcards in the name field.
Once the form is completed, the user clicks a button to submit the request. The HTML form submits the query to a CGI program that executes on the Unix server. This program accepts as input the form field values that the user specified. It then communicates with the archive via DICOM requests to determine those patients that match the search criteria. The user may then choose a patient, which in turn causes the studies for this patient to be displayed.
Finally, the user may select a study that causes those images to be retrieved from the archive and displayed via the Web browser. The result of this system is an easy to use interface to a DICOM PACS with the option to query or move images from PACS.

[2000-2001] Routing in mobile adhoc networks: Srinath Perur

Download: Thesis: PDF file [474 K]. PS file [3314 K]. Slides: PPT [160 K].
Abstract: Mobile ad hoc networks are infrastructure-less, wireless networks of mobile nodes that are formed dynamically without much set up time or cost. They are used where communications infrastructure is unavailable or infeasible to use, as in battlefield or disaster relief operations. Routing in such ad-hoc networks is an issue since node mobility makes the network topology highly dynamic.
This thesis surveys some of the work done in the area of ad hoc routing and proposes Kelpi, a new routing algorithm. Kelpi is based on the observation that with most routing algorithms, an intermediate node's movement could cause routes between distant nodes to be affected. Kelpi is designed to maintain stable routes that form a dynamic virtual backbone in an ad hoc network. The principle behind this is to keep routing information confined to a geographical area despite nodes being mobile. This is done by imposing a cellular structure on the ad hoc network using an accurate positioning service at every node. Kelpi also aims to increase link-layer throughput by using multiple levels of transmission power to reduce radio interference.
An implementation of Kelpi for the Network Simulator ns2 is also described.

[2000-2001] Mobile agents for distance evaluation: Vikram Jamwal

Download: Thesis: PDF file [452 K]. Slides: PPT [320 K].
Abstract: New models for education dissemination have emerged with the growth of distributed systems, especially with the widespread penetration of Internet. This has made it possible to impart education on a larger scale. Distance evaluation of students constitutes a crucial factor for success of Distance Education initiatives. Such large distributed systems also raise a number of challenges in terms of design, technologies and their implementations.
Most of the present day systems have client-server architectures. The client-server model though powerful, has scalability limitations for distance evaluation systems. Over the past few years, the mobile agent paradigm, which has emerged as a new approach for structuring distributed applications, attempts to address many of these concerns.
In this project, we survey the existing mobile agent frameworks to understand the state of art. We then use the mobile agent approach for designing, implementing and deploying a system for distance evaluation of students. We consider the examination process: starting from paper-setting, where the examiners spread over the internet collaborate to produce a question paper, to examination conduction, answer- paper evaluation, and ending with result compilation and publishing.
In this report, we detail our design, implementation and experimentations. We conclude by presenting our observations, experiences of using mobile agents for designing large distributed systems. We also list some the challenges that still need to be tackled and indicate the future directions of our work.

[2000-2001] Mobile agents for e-commerce: Rahul Jha

Download: Thesis: PDF file [440 K]. PS file [1038 K]. Slides: PPT [508 K].
Abstract: In the past few years, mobile agent (MA) paradigm has received a great deal of attention. While MAs have generated considerable excitement in the research community, they have not translated into a significant number of real-world applications. One of the main reasons for this is the lack of work that quantitatively evaluates the effectiveness of mobile agents versus traditional approaches. This project contributes towards such an evaluation and implements an e-commerce application using mobile agents.
The existing mobile agent applications in the domain of e-commerce are classified and the underlying patterns of mobility are identified. These in turn determine several implementation strategies using the traditional client-server and the mobile agent paradigms.
The project quantitatively evaluates various implementation strategies and identifies various application parameters that influence application performance. The project also provides qualitative and quantitative comparison across three Java based mobile agent framework viz. Voyager, Aglets, Concordia, for e-commerce applications.
Finally, we present the implementation and deployment issues of a complete B2C e-commerce application using mobile agent and messaging and discuss the software engineering aspects of mobile agent technology.

[2000-2001] Automated testing of UML behavioral descriptions: Venkat G. [co-guide: Prof. S.Ramesh]

Download: Thesis: PDF file [1075 K]. PS file [5119 K]. Slides: PPT [627 K].
Abstract: The increasing complexity of safety-critical software systems and the costs associated with software failure has given rise to a need for practical and effective testing strategies.
Software testing is a critical element of software quality assurance and represents the ultimate review of specification, design, and coding. Testing is a dynamic method for verification and validation, wherein the execution behavior of the given system is observed for several test cases. The time and cost devoted to testing needs to be managed accurately. Too often, lack of sufficient testing causes schedule and budget over-runs with insufficient guarantee of quality.
An effective software testing strategy starts with the implementation of a component test process. In component testing, critical components (shared, reusable, or complex) are subjected to functional or black box testing to ensure that they are functionally correct. Also, code coverage, i.e., the number of execution paths covered by a given test case, needs to be analyzed to determine the effectiveness of the testing process.
State machines or statecharts are a common approach for the abstraction of control and data flow in a wide range of application, such as real-time applications, networking protocols, GUI design and safety-critical systems. Statechart is a formalism for visual specification of reactive system behavior. The formalism extends traditional finite-state machines with notions of hierarchy and concurrency, and is used in many popular software design notations. A large part of the appeal of using statecharts for testing, is due to the intuitive operational interpretation of state machine behavior. Statecharts are hierarchical in nature and allow one to decompose states into sub-states.
For systems where tasks are performed under time-constraints, the misinterpretation of even a single event in state machine's behavior can lead to collapse the whole underlying system. Hence an efficient testing tool is mandatory for such systems.
In this report, we discuss various testing techniques, testing strategies and statechart semantics, using practical examples. We have developed StateTest, a tool that takes a statechart specification and the test scripts as the input, and generates the pass or fail report and the coverage metrics as the output. Presently, StateTest can test the behavior of traditional as well as hierarchical statecharts. The tool can also be extended to test the statecharts that have orthogonal (concurrent execution) as well as history states.

[1999-2000] Sequential and parallel reachability analysis of concurrent Java programs: Raghuraman Rangarajan [IIT-G, co-guide: Prof. G. Sajith]

Download: Thesis: PDF file [415 K]. PS file [481 K]. Slides: PPT [276 K].
Abstract: Concurrent programs are more difficult to test than sequential programs because of their non-deterministic behaviour. Reachability analysis is an important and well known tool for static analysis of concurrent programs. It involves the systematic enumeration of all possible global states of program execution. However, traditional algorithms to generate all reachable states of a concurrent program have exponential time and space complexities.
Apportioning is a technique based on the idea of classification of program analysis points as local(i.e. having influence within a class) and global (i.e. having influence outside a class). Apportioning uses this classification to abstract away some analysis points, thus reducing the size of the reachability graphs generated.
This report presents two algorithms for the generation of reachability graphs for a concurrent Java program. The algorithms are implemented for some apportioning-based reachability analysis tools. The results generated are used to verify the efficiency of apportioning as a tool for reachability analysis of concurrent Java programs.
The first algorithm is a sequential implementation of the apportioning technique, while the second generates the reachability graphs in parallel. While the sequential algorithm demonstrates the reduction in the space complexity of a reachability graph generation, the second algorithm attempts to mitigate the time complexity.

Bachelor's/Mini Projects Back

[2000-2001] Comparison of routing protocols for mobile adhoc networks Priyank Mishra IIT Bombay

[1998-1999] Process migration support in Linux Vishal Agrawal IIT Guwahati

[1998-1999] Reachability graph generation of concurrent Java programs Anirban Majumdar IIT Guwahati

[1998-1999] Generation of control-flow graphs for Java programs Ajay Bajpai IIT Guwahati

[1998-1999] Policy based management for enterprise networks Sameer S Chabungbam IIT Guwahati

In addition to the above, a large number of students from other colleges have worked with me for their final year projects. The numbers increased so rapidly that eventually I gave up trying to keep track of these projects. Approximately, it is an average of 10 students (3 group projects) per year.

Last Updated: June 2008

Back

Year	Student	Topic
[2007+]	Dr. Farida Umrani (PhD in Psychology from IIT Bombay)	Impact assessment of eLearning technologies. [Publications: CM1; DL 08; ] (Fellowship supported by Project OSCAR)
[2001-2003]	Dr. Leena Chandran-Wadia (PhD in Physics from IISc Bangalore)	Medium access control in wireless networks. [Publications: ICCC 02; ICCCN 03; ICDCS 03; CIT 03.] (Fellowship supported by Infosys)

[2000-2001]	Comparison of routing protocols for mobile adhoc networks	Priyank Mishra	IIT Bombay
[1998-1999]	Process migration support in Linux	Vishal Agrawal	IIT Guwahati
[1998-1999]	Reachability graph generation of concurrent Java programs	Anirban Majumdar	IIT Guwahati
[1998-1999]	Generation of control-flow graphs for Java programs	Ajay Bajpai	IIT Guwahati
[1998-1999]	Policy based management for enterprise networks	Sameer S Chabungbam	IIT Guwahati