M Tech Dissertations

Permanent URI for this collectionhttp://drsr.daiict.ac.in/handle/123456789/3

Browse

Subject: search.filters.subject.Computer science

Search Results

Now showing 1 - 10 of 21
  • Thumbnail Image
    ItemOpen Access
    Formal semantic analysis and modeling of natural language Wh-question
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2013) Patel, Rupali; Dasgupta, Sourish
    The problem of Natural Language Query Formalization is to understand the semantics of a user given query in natural language (NL) and then translating the query into a formal language (FL) such that the FL semantic interpretation has equivalence with the NL interpretation. Such linguistic analysis based formalization can be used as more accurate query analyzer when compared to statistical analyzers. In this thesis work we have proposed a linguistic analysis based query model called Description Logic based Wh-Query Modelthat syntactically characterizes wh-queries in English and has a complete semantic equivalency to Description Logics (DL). This work also includes a rules to identify desire depedency in case of complex and compound query. We evaluate the query characterization coverage using Microsoft Encarta query dataset and OWLS-TC V4.0 service query dataset.
  • Thumbnail Image
    ItemOpen Access
    Multipath verification defense against SSL stripping attack
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2013) Arora, Sunil; Mathuria, Anish M.
    SSL stripping attack is a man- in- the- middle attack which poses a serious threat to the security of secure socket layer protocol. In SSL stripping attack the attacker has ability to downgrade security of SSL protected connection, and view web traffic of the user in clear text. The attack is based on the fact that user rarely request for secure connection explicitly but rely on the web server to redirect them to secure version of the particular website. An attacker, after becoming man- in- the- middle can suppress such messages and provide the user with stripped version of the requested website and forcing him to communicate over insecure HTTP channel. There are several solutions recently proposed to solve the problem of SSL stripping attack, however all solutions have some limitations. In this thesis work we address the limitations of the existing solutions and proposed a new method using idea of multipath verification to detect SSL stripping attack. We establish multiple connection with the remote server using alternate paths, and compare security of them (server support HTTP or HTTPS). We accept the connection with the remote server if securities of the connection established over various paths match, otherwise we block the connection.
  • Thumbnail Image
    ItemOpen Access
    Migration in initial and dynamic virtual machine placement algorithms
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2013) Reddy, Narender A.; Chaudhary, Sanjay
    Cloud computing provides a computing platform for the users to meet their demands in an efficient way. Virtualization technologies are used in the clouds to aid the efficient usage of hardware. Virtual machines are utilized to satisfy the user needs and are placed on physical machines of the cloud for effective usage of hardware resources and electricity in the cloud. Optimizing the number of physical machines used helps in cutting down the power consumption by substantial amount. An optimal technique is to map virtual machines to physical machines such that the number of required physical machines is minimized. The virtual machine placement problem with the target of minimizing the total energy consumption by the running of physical machines, which is also an indication of increasing resource utilization and reducing cost of a data center. Due to the multiple dimensionality of physical resources, there always exists a waste of resources, which results from the imbalanced use of multi-dimensional resources. To characterize the multi-dimensional resource usage states of physical machines, a multi-dimensional normalized resource cube is presented. Based on this model, we propose a virtual machine placement algorithm with migration support which can balance the utilization of multi-dimensional resources, reduce the number of running physical machines and thus lower the energy consumption. We also evaluate our proposed algorithm via extensive simulations and experiments on Cloudsim. Experimental results show, over the long run, proposed algorithm can save as much as 15% energy than the other algorithms.
  • Thumbnail Image
    ItemOpen Access
    Resource allocation on infrastructure as a service cloud using policies
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2010) Nathani, Amit; Divakaran, Srikrishnan
    Conventionally, a cloud refers to an Infrastructure as a service cloud. Infrastructure as a service cloud providers manage a large set of computing resources. These resources can be provided to cloud users on demand in the form of virtual machines. Cloud consumers do not need to manage resources and be worried about the performance issues because they are handled by cloud providers. Resource allocation in the context of infrastructure as a service cloud means allocating virtual resources namely computing capacity, storage etc. to competing requests based on pre-defined resource allocation policies. In real world most of the Infrastructure as a service clouds rely on simple resource allocation policies like immediate and best effort. Immediate means the resources are allocated if they are available or the request is rejected and best effort means the requested resources are allocated if they are available or the request is placed in first come first serve queue. Sometimes it is not possible for a cloud provider to satisfy all the requests which come to them immediately because of lack of resources. In this case cloud providers can benefit from more complex resource allocation policies. Haizea is a resource lease manager that tries to address above issues. It uses resource leases as resource allocation abstraction and implements these leases as virtual machines. Currently, it supports four kinds of resource allocation policies: immediate, best-effort, advance reservation and deadline sensitive. The aim of thesis is to extend the current scheduling algorithm of Haizea to support deadline leases in an efficient manner. A dynamic planning based scheduling algorithm is proposed which will admit new leases and prepare the schedule whenever a new lease can be accommodated. The proposed algorithm is implemented in Haizea. Experiments are performed to demonstrate the effectiveness of it. The results show that it maximizes resource utilization and acceptance of leases compared to the existing algorithm of Haizea.
  • Thumbnail Image
    ItemOpen Access
    Eye localization in video: a hybrid approach
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2010) Kansara, Bena; Mitra, Suman K.
    Location of eyes is an important process for operations such as orientation correction, which are necessary pre-processes for face recognition. As eyes are one of the main features of the human face, the success of facial feature analysis and face recognition depends greatly on eye detection. It is advantageous to detect eyes before other facial features because the position of other facial features can be estimated using eye position and golden ratio. Since relative position of eyes and interocular distance are nearly constant for different individuals, eye localization is also useful in face normalization. Hence, eye localization is a very important component for any face recognition system. Various approaches to eye localization have been proposed and can be classified as feature based approaches, template based approaches and appearance based approaches. Feature based methods explore eye characteristics - such as edge and intensity of iris - to identify some distinctive features around the eyes. In template based methods, a generic model of eye shape is designed; this template is then matched to the face image pixel by pixel to find the eyes. Appearance based methods detect eyes based on their photometric appearance. Template based and appearance based methods can detect eyes accurately but they are not efficient when considering time factor while feature based methods are efficient but do not give accurate results. So, by combining feature based method with template based or appearance based method, we can get better results. In the proposed algorithm, we have combined feature based eye LEM approach proposed by Mihir Jain, Suman K. Mitra, Naresh D. Jotwani in 2008 and appearance based Bayesian classi_er approach proposed by Everingham M., Zisserman A. in 2006 to achieve eye localization. The work of localizing eyes in a video is motivated by some of the applications where eye localization can serve very useful purpose such as to find drowsiness of a person driving a car, eye based control of computer systems for people with motor difficulties. To carry out eye localization, after doing some preprocessing which include frame separation from video and to convert it into gray-scale images, the proposed algorithm is applied on each of these frames. For the experimentation, we have taken videos of few people in the normal blink condition as well as in the sleepy condition. All the videos have been taken in the lab environment. To check the accuracy of the proposed algorithm, we have performed various tests, namely, Wilcoxon signed rank test, Mann-Whitney U test, Kolmogorov-Smirnov test, Sensitivity and False alarm rate tests. And the results of these tests show that the proposed algorithm proves to be quite accurate in localizing the eyes in a video. All the experiments have been carried out in MATLAB.
  • Thumbnail Image
    ItemOpen Access
    Negotiation for resource allocation on infrastructure as a service cloud
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2010) Akhani, Janki; Divakaran, Srikrishnan
    The Cloud is a computing platform that provides dynamic resource pools, virtualization, and high availability. Cloud computing infrastructures can allow enterprises to achieve more efficient use of their IT hardware and software investments. Infrastructure As A Service (IAAS) cloud providers manage a large set of computing resources. These resources can be provided to cloud consumers on demand in the form of virtual machines. Cloud consumers do not need to manage resources and be worried about the performance issues because they are handled by cloud providers. Open Nebula is an open source cloud toolkit which can be used to setup an IAAS cloud. It has three components: Open Nebula Core, Virtual Machine Scheduler and Cloud Drivers. Haizea is an open-source resource lease manager, and can act as a virtual machine scheduler for Open Nebula or used on its own as a simulator to evaluate the performance of different scheduling strategies. Haizea supports four kinds of resource allocation policies: immediate, best-effort, advance reservation and deadline sensitive. To reserve resources in advance using Haizea, consumer submits parameters like amount of resources, start time and duration of a reservation as a request. If one or more parameters can not be satisfied, then Haizea will reject the request. This method is very rigid method because it does not allow negotiation of any parameter. Consumer can resubmit new requests by modifying previously submitted request parameters. Consumer will not be aware of the current resource allocation on provider side so, the chances of new requests getting rejected are more. Thus, it will increase communication overhead between cloud provider and consumer as well as it will decrease resource utilization on provider’s side. It will also degrade the performance of a provider in managing many incoming requests due to previously rejected ones. To overcome the above problems, negotiation can be provided. Negotiation process consists of three components which are negotiation protocol, negotiation objectives and agents’ decision making algorithm. The proposed algorithm to generate set of counter offers is a part of decision making model at provider side. It provides set of counter offers to consumer when his advance reservation request gets rejected. It provides set of counter offers considering parameters’ flexibilities to maximize the chances of their acceptance. The proposed algorithm for User selection policy is a part of decision making model at consumer side. Consumer can get best suitable offer from set of counter offers using the algorithm of user selection policy. Ranking algorithm is a partof algorithm for user selection policy. Using this ranking algorithm, consumers will get suitable offers sorted according to their needs. It will reduce consumers’ efforts to go through all the provided counter offers and choose best suitable one. These algorithms are implemented in Haizea. Experiments are performed to demonstrate the effectiveness of algorithms. The results show that the proposed algorithm to generate counter offers maximizes resource utilization and acceptance of requests compared to rigid and exact methods.
  • Thumbnail Image
    ItemOpen Access
    Bidirectional service composition
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2010) Khakhkhar, Sandip; Chaudhary, Sanjay
    Service is a network addressable software component to perform a specific task. A service consumes given input parameters, performs specific task based on input parameters and returns the result in terms of output parameters. A service request specifies required task in terms of input parameters that can be provided and output parameters that are required. A service discovery mechanism can be used to find services that can be executed to satisfy service request. Service and service request is match by comparing their input/output parameters. A service request may be complex enough that it can not be satisfied by an individual service. It might be possible to execute a chain of services in a particular order to satisfy service request. This chain of services is referred as composition plan and service offered by executing this composition plan is referred as composite service. The aim of service composition algorithm is to generate a composition plan and generate composite service to satisfy service request. Services involved in composition plan are selected manually while designing composite service in static composition process. This process consumes considerable amount of time and effort. It is also vulnerable to changes in input/output of services. A dynamic composition algorithm is required that can automatically select services involved in composite plan and generate a composite service on-the-fly. Main issue with dynamic composition algorithms is composition time taken by algorithm to generate a composite service. Composition time indicates duration of the time at which the service request was submitted to the algorithm till the algorithm generate a composite service that can satisfy service request. Composition time depends upon the number of services required to explore in order to find services that can take part in composite plan. Dynamic composition algorithms presented in previous work mainly follows either forward chaining approach or backward chaining approach to find a composite service. Performance of algorithms based on forward chaining approach or backward chaining approach suffers for certain cases to generate a composite service where number of services explored increases exponentially as number of iterations increases. This work proposes a dynamic composition algorithm that gives consistent performance across all the cases. Proposed algorithm approaches from two directions alternatively, one follows forward chaining approach and other follows backward chaining approach. Composition algorithm following only forward chaining approach or backward chaining approach performs all the iterations in one direction only where as proposed algorithm requires only half number of iterations in both directions. Algorithm uses two types of matching strategy to compare input/output parameters. First one is based on keyword matching and second one based on semantic matching strategy. Performance of proposed algorithm is evaluated by performing relevant experiments and results are compared with algorithms based on only forward chaining approach or backward chaining approach. Proposed algorithm explores less number of services and takes less composition time compared to algorithms based on only forward chaining approach or backward chaining approach.
  • Thumbnail Image
    ItemOpen Access
    Particle swarm optimization based synthesis of analog circuits using neural network performance macromodels
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2009) Saxena, Neha; Mandal, Sushanta Kumar
    This thesis presents an efficient an fast synthesis procedure for an analog circuit. The proposed synthesis procedure used artificial neural network (ANN) models in combination with particle swarm optimizer. ANN has been used to develop macro-models for SPICE simulated data of analog circuit which takes transistor sizes as input and produced circuit specification as output in negligible time. The particle swarm optimizer explore the specfied design space and generates transistor sizes as potential solutions. Several synthesis results are presented which show good accuracy with respect to SPICE simulations. Since the proposed procedure does not require an SPICE simulation in the synthesis loop, it substantially reduces the design time in circuit design optimization.
  • Thumbnail Image
    ItemOpen Access
    Path complexity of the class binary search tree
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2009) Doshi, Nishant; Amin, Ashok T.
    Path complexity of a program is defined as the number of program execution paths as a function of input size n. This notion of program complexity has been extended to complexity of a class as follows. Class had data members and data operations. The notion of state for the class is defined based on structural representation of a class. We are assuming only those data operations that change state of a class. The path complexity of a class is defined to be the number of valid input sequences, each of them containing n data operations. We have analyzed the path complexity of the class Binary Search Tree (BST) based on the algorithms for insert and delete data operations. Later we modify program for delete operation to facilitate determination of path complexity for the class BST. The bounds for the path complexity of the class BST are determined. A program is developed to obtain path complexity of the class BST.
  • Thumbnail Image
    ItemOpen Access
    Path complexity of maximum segment sum problem
    (Dhirubhai Ambani Institute of Information and Communication Technology, 2009) Mishra, Devesh; Amin, Ashok T.
    Various software complexity metrics have been proposed in literature. A program complexity measure called path complexity is proposed in [1]. Path complexity P(A,n) of an algorithm A is defined to be the number of program execution paths of A over all inputs of size n. It defines a partition of input space of program A into equivalence classes on the basis of different program execution paths. All the inputs belonging to an equivalence class are equivalent to each other in a sense that they follow same execution path. We present path complexity analysis of four different algorithms for one-dimensional maximum segment sum problem which shows that algorithms with different computational complexity may be equivalent to each other in their path complexity. We also present lower bounds on one dimensional as well as two dimensional maximum segment-sum problems. A different perspective and several observations on one dimensional problem are given