Login
The UniverSelf project is proud to announce the release of its second in a series of technical reports on:
Deliverable D3.3 - Identification of suitable classes of methods for learning and operations.
The report can be accessed under the Dissemination/Technical Reports section, or directly at the following link http://www.univerself-project.eu/technical-reports
Summary:
As per the Description of Work, deliverables D3.2/D3.3 identifies learning methods that are suitable to given problems. Furthermore, it presents cooperative network/service operations and introduces distributed cross-layer and cross-domain methods including observation, correlation, decision-making, and control.
More concretely, this document examines different problems that have been selected in the various use-cases of the project (see deliverable D4.1), and that are 1/ exemplary of important features one can expect from self-management methods, and 2/ of practical importance to network operators, as sources of OPEX/CAPEX reductions, 3/ instructive about the design of a Unified Management Framework (UMF). Two other main deliverable series are attached to WP3 on network empowerment. Deliverable D3.1 focuses on all aspects of parameter optimization in networks, whereas deliverable D3.4 addresses the cooperation strategies between different management functions and/or modules. The idea is that modern networks are intrinsically distributed systems, with diverse technologies, vendors, managed domains and objectives. Their management is also distributed, both for scalability issues, and to account for different managed domains and various objectives of the management functions. Therefore one must address issues like the cooperation to achieve a global objective, compatibility of local objectives, trustability, etc. In contrast, deliverable D3.2 and deliverable D3.3 address all the remaining aspects of network empowerment, which covers the design of algorithms for all parts of control loops: observation, knowledge extraction, decision, action, as well as stability issues. The selected problems cover fault diagnosis, healing, information dissemination, optimal content delivery, early detection of congestion, vulnerability detection and management, stability issues of control loops, and more.
The document is structured according to use-cases, in which several specific problems are extracted. These problems have different degrees of ambition and of maturity, and this version of the deliverable tries to capture the work achieved so far to address them. There will be future releases of this work, at the second and third year of the project, where the algorithmic solutions will be given with more details. The work described here will support both the demonstrations and validations scheduled in the life cycle of the use-cases, and will serve as a basis to design and challenge the concepts of UMF.
This document (D3.3) is the public part of the technical report D3.2 which is restricted to internal use to the UniverSelf project.
