Defining Emergent Software Using Continuous Self-Assembly, Perception, and Learning

Abstract

Architectural self-organisation, in which different configurations of software modules are dynamically assembled based on the current context, has been shown to be an effective way for software to self-optimise over time. Current approaches to this rely heavily on human-led definitions: models, policies and processes to control how self-organisation works. We present the case for a paradigm shift to fully emergent computer software which places the burden of understanding entirely into the hands of software itself. These systems are autonomously assembled at runtime from discovered constituent parts and their internal health and external deployment environment continually monitored. An online, unsupervised learning system then uses runtime adaptation to continuously explore alternative system assemblies and locate optimal solutions. Based on our experience over the last three years, we define the problem space of emergent software and present a working case study of an emergent web server as a concrete example of the paradigm. Our results demonstrate two main aspects of the problem space for this case study: that different assemblies of behaviour are optimal in different deployment environment conditions; and that these assemblies can be autonomously learned from generalised perception data while the system is online.

Autonomously Improving Systems in Industry: A Systematic Literature Review

“…1 the distribution between the different types on learning approaches used in the papers are shown. Most of the ML papers apply an online learning approach [31][32][33][34][35][36][37][38][39], a few use offline [40][41][42], or a hybrid [43][44][45][46] which is a combination of both online and offline approaches, but the industrial deployed systems tends to have a lower level of autonomy or specifically focuses on a few numbers of challenges. A/B/n testing papers [5,[47][48][49][50] simple adaption is done online in deployed systems.…”

Workflow Variability for Autonomic IoT Systems

“…The related work is classified into two categories concerning workflow variability: solution space variability and Models@Runtime. We omit approaches using variability at the planning-level (e.g., [12]) as they do not propose any model constructs for supporting workflow variability, but they are built on top of existing component models with reconfiguration capabilities (e.g., Fractal [13]).…”

Towards emergent microservices for client-tailored design

“…Emergent Software Systems, as described in [4,8], use continuous self-assembly over a pool of small software building blocks to derive systems that are autonomously composed as a factor of the operating environment and human-provided high-level goals. Starting from no initial knowledge, the ideal composition of behaviour for each range of observed deployment environment conditions is automonously learned using real-time reinforcement learning.…”