International audienceContinuously adjusting the horizontal scaling ofapplications hosted by data centers appears as a good candidateto automatic control approaches allocating resources in closedloopgiven their current workload. Despite several attempts,real applications of these techniques in cloud computing infrastructuresface some difficulties. Some of them essentially turnback to the core concepts of automatic control: controllability,inertia of the controlled system, gain and stability. In thispaper, considering our recent work to build a managementframework dedicated to automatic resource allocation in virtualizedapplications, we attempt to identify from experiments thesources of instabilities in the controlled systems. As examples,we analyze two types of policies: threshold-based and reinforcementlearning techniques to dynamically scale resources. Theexperiments show that both approaches are tricky and thattrying to implement a controller without looking at the waythe controlled system reacts to actions, both in time and inamplitude, is doomed to fail. We discuss both lessons learnedfrom the experiments in terms of simple yet key points to buildgood resource management policies, and longer term issueson which we are currently working to manage contracts andreinforcement learning efficiently in cloud controllers
The problem addressed in this article is image indexing and retrieval according to the color. Indeed we propose a classification based on the dominant color(s) of the images. The process consists in two steps: first, assigning a colorimetric profile to the image in HLS space (Hue, Lightness, Saturation) and then, handling the query for the retrieval. To achieve the first step, the definition of hue is done using a fuzzy representation that takes into account the nonuniformity of color distribution. Then, lightness and saturation are represented through linguistic qualifiers also defined in a fuzzy way. Finally, the profile is built through fuzzy functions representing the membership degree of the image to different classes. Thus, the query for image retrieval is a pair (hue, qualifier). The second step looks for a match between the query and the profiles. In order to improve the software and to make it more flexible, the user can re-define the fuzzy representation of Hue, Lightness and Saturation, according to his own perception.
Abstract-Computing has reached the time of distributed applications everywhere. Service-oriented architectures are more and more used to organize such complex and highly dynamic applications into business processes calling services discovered in registries at load-time. In this context, Quality of Service (QoS) and agility in business processes become key issues. Instead of binding business processes to services at load-time, this paper proposes to monitor sets of candidate services for their current QoS and to choose among them at call-time. This new form of late-binding paves the way to more agile and robust applications called adaptive business processes. Besides the conceptual background and implementation of this late-binding in an industrial-strength web service platform, this paper presents the LCP-net formalism introduced to provide programmers with a mean to express qualitatively their preferences among the different QoS properties of services, hence tackling the multicriteria decision making arising from the run-time choice among candidate services given several unrelated QoS properties.
This contribution addresses the problem of expressing preferences among nonfunctional properties in a Web Service architecture. In such a context, semantic annotations are needed and added on service declaration and business process in order to select the best available service. These conditional and unconditional preferences are managed using Conditional Preference-Networks (CP-Nets). But in several cases, uncertainty related to the preferences has to be taken into account to achieve a better satisfaction rate. We propose the use of fuzzy linguistic information inside the whole process when it will be necessary.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.