Managing Distributed Cloud Applications and Infrastructure by Theo Lynn & John G. Mooney & Jörg Domaschka & Keith A. Ellis

Managing Distributed Cloud Applications and Infrastructure by Theo Lynn & John G. Mooney & Jörg Domaschka & Keith A. Ellis

Author:Theo Lynn & John G. Mooney & Jörg Domaschka & Keith A. Ellis
Language: eng
Format: epub
ISBN: 9783030398637
Publisher: Springer International Publishing


By adopting different techniques, for example regression or machine learning, a workload modeller can construct multiple workload models. For flexibility in system integration, the modellers can be implemented using various technologies. The workload models are then wrapped and exported as a microservice using a REST API Gateway. On top of model services, a set of adapters are built to provide a unification layer. While workload models are constructed using historical workload data, they can be updated continuously at run-time by the workload modeller. Predictors in the platform make use of the adapters in order to access the available models and to make their predictions.

On top of this platform, robust and efficient approaches for autoscaling are constructed based on the results of workload modelling and prediction. Optimisers encapsulating optimisation algorithms and application-specific constraints make use of prediction for proactive optimisation. Note that an optimiser can call multiple predictors that access different models constructed using different techniques. This also implies that predictors are developed for every constructed model. Such implementation enables the capability of extension when a new optimiser (of a new application) or new model (using new techniques) is added to the system.



Download



Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.