DoS with LIKE query

I was assessing an application. The application was properly sanitizing all the characters which have special meaning for SQL Injection attack. So SQL Injection was not possible in the application. But then again I came across few search modules in the application where it was taking input of part numbers to proceed.
I entered single quote(') and the application was perfectly filtering it returning "Parts can't be found". Then out of curiosity I entered '%' character and observed the response. Now the application stuck into the loop of the search continuously searching.The two things I deduced from it:
1. The application was using LIKE query to search matching terms.
2. This can be used to perform a DoS by overloading the database.

The % and _ qualifier is often overlooked by developers to filter as its not so devastating as other characters. They are used for matching 0 or more characters and single character respectively.

$searhterm = mysql_real_escape_string(“%anything”); // still %anything
mysql_query(“SELECT * FROM messages WHERE subject LIKE ‘{$searchterm}%’”);

The intention of the query above is to search the contents matching user specified $searchterm.
In normal conditions the query will execute fast. But when entered a term with a leading % quantifier the query takes too long to perform as it can't find the index. It progressively goes slower as amount of data in table grows.
Same is the case with _ (underscore).
Although these are valid inputs, we need to filter them out. in PHP there is a function which actually excludes the terms specified.
Use addcslashes() for escaping the above characters:

$searchterm = addcslashes(mysql_real_escape_string(“%anything_”), “%_”); // $searchtearm == \%something\_
mysql_query(“SELECT * FROM messages WHERE subject LIKE ‘{$sub}%’”);

Here, the input is processed by the database’s prescribed escape function and is then filtered
through addcslashes() to escape all occurrences of % and _.

In case of my application it can be determined by the fact that entering % one or more times was causing the whole application to be non responsive for a longer time



Popular posts from this blog

SQL Injection in search field

File Upload through Null Byte Injection