Identity and Access Management in Distributed Enterprise Applications: A Role-Based and Attribute-Based Access Control Framework

Main Article Content

Shubh Prabhat

Abstract

Enterprise applications are typically deployed as distributed services in multiple locations‚ including co-located and partner data centers․ All distributed services require a mechanism to enforce access control for authenticated users․ RBAC is a simpler access control mechanism; however‚ if there are clearance and other location-based and time-based access restrictions‚ it may require many roles․ ABAC is a more efficient access control mechanism for implementing clearance and other location-based and time-based restrictions but is a less convenient mechanism for managing access policies because it lacks other organizational features of access policies․ This paper proposes an IAM architecture that integrates RBAC and ABAC features by adding attributes to RBAC role permissions․ A two-step evaluation process that includes opportunities for caching and no attribute queries for environment attributes provides efficient access enforcement․ The framework includes duplicate policy enforcement systems and features in deployed applications and event notifications to quickly remove user access․ A Python prototype implementation and simulation of an organization with 5‚000 employees‚ 20‚000 objects‚ and 200‚000 session accesses showed that a combined role-based and attribute-based access control policy requires 274 policy elements‚ while a pure role-based policy requires 6‚912 roles․ A pure attribute-based policy would provide 22․2% more access than necessary․ The combined policy allows more accurate access decisions‚ requiring one policy rule check per transaction‚ compared with 30․4 policy rule checks under a pure attribute-based policy․ Attribute screening was reduced from 2․00 checks per transaction with no caching to 1․23 checks per transaction using a decision cache of 4‚096 entries․ System transaction volume increased almost linearly with the number of decision point replicas‚ to 16 replicas․ Changes to the employee workforce require 33% fewer administrative tasks than changes under a role-based system․

Article Details

Section
Articles