Scientific Methodology and Work Packages
The project is organized into six work packages, as described below.
Work Package 1
This work package will ensure the overall coordination and day-to-day management of the project, including the overall project plan. It will also handle all the financial, legal and contractual matters.
Work Package 2
This work package will focus on developing observation and communication methods that are essential for an efficient operation of PANACEA's autonomic control of cloud resources. The major outcomes of this work package will be the detailed definition of:
- A Pervasive Monitor, which includes observation methods and computing elements that will continuously measure and infer with a proper level of resolution, the operating state (e.g., component load) and anomalous events that could adversely affect the delivery of services (e.g., malfunction or inappropriate behaviours).
- An Autonomic Communication Overlay, which will provide a robust, self-adaptable topology to be exploited by both the Pervasive Monitor and by the autonomic manager to optimize federated cloud applications.
Work Package 3
The focus of this work package is the development of innovative solutions for autonomic management at a node level (server, software, database, and storage devices), which will be based on a set of advanced Machine Learning (ML) techniques. ML framework of WP3 will be focusing on reducing the complexity of the models by removing the number of irrelevant parameters from the training data set while preserving the accuracy of the prediction. A goal of WP3 is the prediction of time to crash of applications and triggering a proactive rejuvenation of user computational extensive applications, running on Virtual Machines (VMs).
The ML framework will allow the prediction of the time to crash in web servers. As of July 2011, Apache is running over 235 millions Internet Servers. As well, implementing proactive servers rejuvenation for preventing crashes and increasing their availability. The aim of WP3 is to correlate external observations, as a response time, with a web server performance by internally observed operating system data. This will allow pro-actively actions to be taken before server's response time is reaching a given threshold (maximum value).
Work Package 4
This work package will identify and test operational scenarios for PANACEA mainly driven by a representative use case. The work package will carry on work as a logical continuation of the conceptual efforts that will be developed in WP2 and WP3 and will focus on designing, implementing and evaluating applications that will use the methods and algorithms developed in those work packages. The WP will provide evaluation methodologies and will apply them on the provided components as well as on the virtualization and management middleware framework that will be used.
Work Package 5
This work package is devoted to the exploitation of the results produced by the project. Its major objective is to ensure that the technical results of the project are guided by, suitable for, and ultimately taken up by, the target users, participant organisations and community researchers.
Work Package 6
This work package is devoted to the dissemination of the results produced by the project. Its major objective is to ensure that PANACEA is known to its stakeholders and that these stakeholders interact with the project during its lifeteime, providing input to scientific and sustainability development, and following its lifetime in the exploitation phase.