Information Systems Security: 6th International Conference, ICISS 2010, Gandhinagar, India, December 17-19, 2010

2.1 Web Application Vulnerabilities Many web application vulnerabilities havebeenwell documented andthemi- gation methods havealso beenintroduced [1]. The most common cause ofthose vulnerabilities isthe insu?cient input validation. Any data originated from o- side of the program code, forexample input data provided by user through a web form, shouldalwaysbeconsidered malicious andmustbesanitized before use.SQLInjection, Remote code execution orCross-site Scriptingarethe very common vulnerabilities ofthattype [3]. Below isabrief introduction toSQL- jection vulnerability though the security testingmethodpresented in thispaper is not limited toit. SQLinjectionvulnerabilityallowsanattackertoillegallymanipulatedatabase byinjectingmalicious SQL codes into the values of input parameters of http requests sentto the victim web site. 1: Fig.1. An example of a program written in PHP which contains SQL Injection v- nerability Figure 1 showsaprogram that uses the database query function mysql query togetuserinformationcorrespondingtothe userspeci?edby the GETinput- rameterusername andthen printtheresultto the clientbrowser.Anormalhttp request with the input parameter username looks like http://example. com/ index.php?username=bob . The dynamically created database query at line2 is SELECT * FROM users WHERE username=‘bob’ AND usertype=‘user’ . Thisprogram is vulnerabletoSQLInjection attacks because mysql query uses the input value of username without sanitizingmalicious codes. A malicious code can be a stringthatcontains SQL symbols ork- words.Ifan attacker sendarequest with SQL code (‘alice’-‘) - jected http://example.com/index.php?username=alice’- , the query becomes SELECT* FROM users WHERE username=‘alice’–‘ AND usertype='user’ .