Composite Pattern for Autonomic Switching of Service Back-Ends between the Fog and the Cloud
Dionysis Athanasopoulos · 2021
Given that cloud machines are usually remotely located from the devices of the end-users of the front-end of mobile apps, the end-users can face delays. The Fog has been introduced to augment mobile apps with machines for data analytics that are close/at the network edge. However, edge machines are resource constrained and hence, the execution of heavy data-analytics on edge machines is not always feasible. Thus, light versions of data-analytics algorithms should be deployed on edge machines. But, how can software engineers develop mobile apps that autonomically switch between the Fog and the Cloud? To answer this, we found the composite pattern of the Autonomic Integrator that extends the back-end of mobile apps to use alternative data-analytics algorithms. The pattern first includes the definition of the conceptual model of an extensible back-end that integrates back-end instances deployed on the Fog and the Cloud. Secondly, the pattern includes the conceptual model of an autonomic component. The autonomic component decides at runtime the switching of the front-end to a back-end instance that has the lowest response-time. Finally, the pattern covers the integration between the extensible back-end and the autonomic component.