Diogo Domingues Regateiro

Business logics of relational databases applications are an important source of security violations, namely in respect to access control. The situation is particularly critical when access control policies are many and complex. In these cases, programmers of business logics can hardly master the established access control policies. Now we consider situations where business logics are built with tools such as JDBC and ODBC. These tools convey two sources of security threats: 1) the use of unauthorized Create, Read, Update and Delete (CRUD) expressions and also 2) the modification of data previously retrieved by Select statements. To overcome this security gap when Role-based access control policies are used, we propose an extension to the basic model in order to control the two sources of security threats. Finally, we present a software architectural model from which distributed and typed RBAC mechanisms are automatically built, this way relieving programmers from mastering any security schema. We demonstrate empirical evidence of the effectiveness of our proposal from a use case based on Java and JDBC.

In database applications, access control is aimed at supervising users’ requests to access sensitive data. Users’ requests are mainly formalized by Create, Read, Update and Delete (CRUD) expressions. The supervision process can be formalized at a high level, such as based on the RBAC model, but in the end the relevant aspect is the data being accessed through each CRUD expression. In critical database applications access control can be enforced not on a CRUD by CRUD basis but enforced at the level of sequences of CRUD expressions (workflow). This situation can occur whenever established security policies are based on strict procedures that define step by step the actions (sequences of CRUD expressions) to be followed. Current RBAC models do not support this type of security policies. To overcome this security gap, we leverage previous researches to propose an extension to the RBAC model to control for each role which sequences of CRUD expressions are authorized. We demonstrate empirical evidence of the effectiveness of our proposal from a use case based on Java and JDBC. Our use case is based on typed security layers built from a software architectural model and also from metadata based on the proposed RBAC model extension.

In database applications, access control security layers are mostly developed from tools provided by vendors of database management systems and deployed in the same servers containing the data to be protected. This solution conveys  several drawbacks. Among them we emphasize: 1) if policies are complex, their enforcement can lead to performance decay of database servers; 2) when modifications in the established policies implies modifications in the business logic (usually deployed at the client-side), there is no other possibility than modify the business logic in advance and, finally, 3) malicious
users can issue CRUD expressions systematically against the DBMS expecting to identify any security gap. In order to overcome these drawbacks, in this paper we propose an access control stack characterized by: most of the mechanisms are deployed at the client-side; whenever security policies evolve, the security mechanisms are automatically updated at runtime and, finally, client-side applications do not handle CRUD expressions directly. We also present an implementation of the proposed stack to prove its feasibility. This paper presents a new approach to enforce access control in database applications, this way expecting to contribute positively to the state of the art in the field.

Most of the security threats in relational database applications have their source in client-side systems when they issue requests formalized by Create, Read, Update and Delete (CRUD) expressions. If tools such as ODBC and JDBC are used to develop business logics, then there is another source of threats. In some situations the content of data sets retrieved by Select expressions can be modified and then committed into the host databases. These tools are agnostic regarding not only database schemas but also regarding the established access control policies. This situation can hardly be mastered by programmers of business logics in database applications with many and complex access control policies. To overcome this gap, we extend the basic Role-Based Access policy to support and supervise the two sources of security threats. This extension is then used to design the correspondent RBAC model. Finally, we present a software architectural model from which static RBAC mechanisms are automatically built, this way relieving programmers from mastering any schema. We demonstrate empirical evidence of the effectiveness

Nowadays, database application use tools like Java Database Connectivity, Hibernate or ADO.NET to access data stored in databases. These tools are designed to bring together the relational database and object-oriented programming paradigms, forsaking applied access control policies. Hence, the application developers must master the established policies as a means to develop software that is conformant with the established access control policies. Furthermore, there are situations where these policies can evolve dynamically. In these cases it becomes hard to adjust the access control mechanisms. This challenge has led to the development of an extension to the role based access control (RBAC) model where permissions are defined as a sequence of create, read, update and delete (CRUD) expressions that can be executed and the interfaces to access them. From these permissions it's possible to generate security artefacts on the client side, i.e. in a distributed manner, which allows the clients to access the stored data while satisfying the security policies defined. On top of this model extension, a security layer has also been created in order to make the access control secure and obligatory. For the RBAC model extension this work leverages a previous work that created a dynamic access control architecture for relational applications, here referred to as DACA (Dynamic Access Control Architecture). DACA uses business logic information and the defined access control policies to build dynamically the security artefacts for the applications. In situations where the access control policies can evolve dynamically, the security artefacts are adjusted automatically. This base work, however, defines as permissions CRUD expressions, which can be executed in any order, and needs an adequate security layer to authenticate users and protect the system form intruders. Hence, this work aims to create a new architecture, called “S-DRACA” (Secure, Dynamic and Distributed Role-based Access Control Architecture), which extends the work done with DACA so that it is capable of enforcing sequences of CRUD expressions that the applications can execute if the sequences are associated with their roles and the development of a security layer to make it secure. We discuss as well the performance of this system and its applicability to other environments outside of relational databases.