A Study Of Vulnerabilities Detection Technology Based On File Format

Hostile attacks exploiting software vulnerability turn out to be increasinglyrampant recent years, among which the malicious software have rendered seriousdamages to computer systems. Owing to the fact that all successful computer attackscould be attributed to successful exploitation of security flaws on the target computersystems, security vulnerability of network and computer software is seen as the primarycause of such attacks. Various videos、mages、audios and documents carried by webtech development not only bring considerable convenience to the common users, butalso provide plentiful targets for the web attackers. The severe threats posted by fileformat vulnerability thus make file format-based software vulnerability measurementtech one of the research hotspots of information security.This dissertation takes formats and software of Microsoft office, adobe and otherthird parties as the research subject and explores the generation mechanism of fileformat vulnerability as well as the triggering feature of such vulnerability whenspecific-format files are processed by focused software. Afterwards, it builds a test-useknowledge base and a model to develop a platform for vulnerability measurement of fileformat so as to offer a solution to system security.Concretely speaking, the main works of this dissertation are summarized asfollows:1. Introduces fundamental knowledge and main methodology of vulnerabilitymeasurement of file format and makes brief retrospect to the current development of fileformat-based vulnerability test, which lay the foundation of following measurement.2．Analyses the file formats of Microsoft office and build a database of test-usedocuments based on the analytic results, which includes the field attributes anddefinition type of file formats.3．Integrates the advantages of traditional vulnerability measurement and makescorresponding modification to come up with a file format-based fuzz measurementmodel and then its engineering realization.The platform of vulnerability detection based on file format generated from this study is verified to have such features as a high-accuracy position, a completeattainment of aberrant information. Moreover, it could reproduce the fault without anydistortion, which provides full and accurate information for the analysis and verificationof vulnerability as well as the design of fixing plan.