Learning Service Semantics for Self-organization in Distributed Environments: Concepts and Research Directions

Abstract A key challenge in performing analytics in distributed environments is to automatically compose services to dynamically match operational tasks to information requirements, accounting for impact, in a many-to-many temporally and spatially complicated and complex situations. In dynamic and agile environments, such as coalition environments, the state of the network and resources cannot be completely known in advance nor controlled due to the evolving nature of the network and constraints that may preclude partners from accessing complete state information about different parts of the system. In addition, there may be requests made to the system that have not been made before, requiring services to be created on the fly. Motivated by these observations, in this paper, we present a critical analysis of gaps in the state-of-the-art and our vision to address those through novel theoretical contributions. We envision that such formalized and theorized fundamentals will enable service elements to automatically configure themselves to perform analytic tasks based on user specified goals by taking account of context—be it system or user context.
Authors
  • Graham Bent (IBM UK)
  • Geeth de Mel (IBM UK)
  • Raghu Ganti (IBM US)
  • Tom La Porta (PSU)
  • Gavin Pearson (Dstl)
  • Tien Pham (ARL)
  • Sebastian Stein (Southampton)
  • Leandros Tassiulas (Yale)
  • Ian Taylor (Cardiff)
Date Oct-2018
Venue IEEE Military Communications Conference (MILCOM) 2018