Physical location awareness for enterprise IT assets

Abstract

In this paper we discuss techniques to automatically discover and track the physical location of networked devices in support of enterprise IT asset management. Recording the location of physical IT assets has long been the domain of manual change-tracking processes and wall-to-wall inventory activities. Meanwhile, increasing end-user device mobility, company growth and merger dynamics, adoption of outsourcing models, as well as transformation of formerly physical servers into virtual machines challenge the efficiency of manual location tracking. Dedicated sensor networks such as those used for RFID tags have been proposed as a replacement. In practice however, extensive infrastructure deployment costs seem to preclude industry adoption. Hence, we focus on location discovery techniques that leverage the ubiquitous wired and wireless enterprise networking infrastructure, to which an increasing percentage of IT assets is at least occasionally connected. Based on this assumption, we present a novel system architecture and algorithm for enterprise-network topology-centric IT asset location estimation based on adaptive learning techniques and spatial hierarchies. We use authentic and simulated data to show the functional and non-functional characteristics of this system.