http://dspace.dtu.ac.in:8080/jspui/handle/123456789/54 2026-04-28T04:03:52Z 2026-04-28T04:03:52Z SHARMA, SHASHI KANT http://dspace.dtu.ac.in:8080/jspui/handle/repository/13890 2014-12-19T11:03:55Z 2011-12-15T00:00:00Z

Title: A FRAGMENTATION TECHNIQUE IN DISTRUBUTED DATABASES & ITS IMPLEMENTATION Authors: SHARMA, SHASHI KANT Abstract: The distributed database design is an optimization problem that includes various subproblems: data fragmentation, allocation and local optimization. Each of these can be solved in various ways. In context of this research we address fragmentation and allocation problems simultaneously for distributed databases management systems (DDBMS) for enhancing the process of server load balancing using Cisco Global Site Selector (GSS). Fragmentation is mainly of three types namely: horizontal, vertical and hybrid (grid) fragmentation. Although there is large work carried out on the design of data fragmentation but most of them are either horizontal or vertical. The core of this thesis defines a new type of grid fragmentation technique in the context of relational databases. Our proposal combines horizontal fragmentation (HF) on the basis of attribute locality precedence and cost model of vertical fragmentation (VF) to generate grid fragments. Both of these two techniques also address the problem of fragment allocation in distributed databases. As part of whole process we discuss and compare various other algorithms for generating candidate vertical and horizontal partition schemes. In the end, we compare our approach with one of the graphical grid fragmentation methods by implementing these techniques as part of Cisco GSS, a product for enhancing the DNS resolution process. Results shows that proposed grid fragmentation is a superior approach and it can solve fragmentation and allocation problem of relational database systems properly. Description: M.TECH

2011-12-15T00:00:00Z KUMAR, RAMAN http://dspace.dtu.ac.in:8080/jspui/handle/repository/13598 2014-12-19T10:58:04Z 2009-11-26T00:00:00Z

Title: IMPLEMENTATION OF TINY ENCRYPTION ALGORITHM FOR SECURING SYSTEM FILE Authors: KUMAR, RAMAN Abstract: The file securing system requires a Tiny Encryption Algorithm (TEA). It basically consist of using a combination of XOR, ADD and SHIFT operations on file blocks to encrypt is using key bits and some junk bits. This algorithm is highly efficient and can satisfy need for a user-friendly encryption system that can be used at homes and offices. The Tiny Encryption Algorithm (TEA) is a block cipher encryption algorithm that is very simple to implement, has fast execution time, and takes minimal storage space. The Tiny Encryption Algorithm (TEA) is a cryptographic algorithm designed to minimize memory footprint and maximize speed. It is a Feistel type cipher that uses operations from mixed (orthogonal) algebraic groups. This thesis presents the cryptanalysis of the Tiny Encryption Algorithm. In this thesis proposed to inspect the most common methods in the cryptanalysis of a block cipher algorithm. TEA seems to be highly resistant to differential cryptanalysis, and achieves comple... Description: ME THESIS

2009-11-26T00:00:00Z BHATT, SHOBHA http://dspace.dtu.ac.in:8080/jspui/handle/repository/13546 2014-12-19T11:22:18Z 2007-07-25T00:00:00Z

Title: HMM BASED HINDI SPEECH RECOGNITION Authors: BHATT, SHOBHA Abstract: Many times key board acts as a barrier between computer and the user. This is true especially for rural areas. We face big challenge in applying this high level specialist knowledge of computers to Community development projects like “Sarva Siksha Abhiyan” so as to quickly bridge the “digital divide”.This barrier can be major hindrance in delivering computer literacy efficiently. Fascinated by the “Hole in the wall” initiative by NIIT, I decided to work on Hindi Speech Recognition so that the users can explore a computer’s potential by interacting with it in their own language thus unlocking the flood gates of development especially for our teeming rural population. This led to start of my work on “HMM BASED HINDI SPEECH RECOGNITION”. The main purpose of this study is to perform speech recognition for isolated words in Hindi language. The railway speech corpus has been included for the study. The speech corpus consists of 207 words spoken by 40 different male, fe... Description: ME THESIS

2007-07-25T00:00:00Z MISHRA, S P http://dspace.dtu.ac.in:8080/jspui/handle/repository/13529 2011-03-26T21:11:28Z 2008-05-15T00:00:00Z

Title: EXPLORING WEAKNESSES OF RC4 IN DIFFERENT USAGE MODES Authors: MISHRA, S P Abstract: Cryptography is an important building block used i n communication systems to provide confidentiality and authenticity. Stream ciphers, which encrypt data one bit or byte at a time have been used in applications sherec low delay and hiogh speed are essential requirements.RC4, a stream cipher designed by Rivest for Data Security Inc. has been used in several commercial applications. In this thesis RC4 is described and its various weaknesses have been exploited. It is shown that second binary derivative of the least significant bit of sequence is correlated to 1 with correlation coefficient close to 152 -3n. The result can be used to distinguish RC4 from other key-stream generators. Digraph probabilities gives another method of distinguishing RC4 from random sequence requiring lesser amount of output key-stream bits. 40-bit RC4 was recommended for WEP, the link-layer security protocol for 802.11 networks. An attack was proposed by Fluhrer, Shamir and Mantin based upon non-random be... Description: ME THESIS

2008-05-15T00:00:00Z