Web-based applications are becoming increasingly popular due to less demand of client-side resources
and easier maintenance than desktop counterparts. On the other hand, larger attack surfaces
and developers’ lack of security proficiency or awareness leave Web applications particularly vulnerable
to security attacks. One existing approach to preventing security attacks is to compose
a redundant system using functionally similar but internally different variants, which will likely
respond to the same attack in different ways. However, most diversity-by-design approaches are
rarely used in practice due to the implied cost in development and maintenance, significant false
alarm rate is also another limitation. In this work, we employ opportunistic diversity inherent
to Web applications and their database backends to prevent injection attacks. We first conduct a
case study of common vulnerabilities to confirm the effectiveness of opportunistic diversity for
preventing potential attacks. We then devise a multi-stage approach to examine database queries,
their effect on the database, query results, and user-end results. Next, we combine the results
obtained from different stages using a learning-based approach to further improve the detection
accuracy. Finally, we evaluate our approach using a real world Web application.