MOBILE DATA COMPRESSION USING E-CLOUD

Abstract

Data processing has been existing as a field since the origin of computer science. However, the interest for data processing increased recently due to the present extension of Internet communication, and to the fact that nearly all texts produced today are stored on, or transmitted through a computer medium at least once during their lifetime. In this context, the processing of large, unrestricted texts written in various languages usually requires basic knowledge about words of these languages. These basic data are stored into large data sets called lexicons or electronic dictionaries, in such a form that they can be exploited by computer applications like spelling checkers, spelling advisers, typesetters, indexers, compressors, speech synthesizers and others. The use of large-coverage lexicons for data processing has decisive advantages: Precision and accuracy: the lexicon contains all the words that were explicitly included and only them, which is not the case with recognizers like spell. Predictability: the behavior of a lexicon-based application can be deduced from the explicit list of words in the lexicon. In this context, the storage and lookup of large-coverage dictionaries can be costly. Therefore, time and space efficiency is crucial issue. In mobile most of the words are repeating it again and again and there are lot of compression technique. It has been observed that LZ trie is best of them if the data is similar in most of the words or sentences. Trie data structure is a natural choice when we think about storing and searching over sets of strings or words. In the contemporary usage of the term, a trie for a set of words is a tree in which each transition represents one symbol (or a letter in a word), and nodes represent a word or a part of a word that is spelled by traversal from the root to the given node. The identical Mobile Data Compression using e-cloud Page 6 prefixes of different words are therefore represented with the same node and space is saved where identical prefixes abound in a set of words - a situation likely to occur with natural language data. The access speed is high, successful look up is performed in time proportional to the length of word since it takes only as many comparisons as there are symbols in the word. The unsuccessful search is stopped as soon as there is no letter in the trie that continues the word at a given point, so it is even faster. With the above technique we can compress the static data but data is changing continuously. In order to make it dynamic, we will compress the static data and will keep the separate database for the modify/delete/add entries and then send the compress data along with the database that stores the modified data to e-cloud in order to make it faster. At eloud, decompression take place for compressed data and appended/modified the list as per the modification and them compression take place and send it to mobile. Based on the results, it is concluded that LZ Trie is most suitable compression technique in terms of memory saving. It is having the time constraint for compressing the data if the data is very large which can be overcome by doing all these operations at e-cloud.