Panacea
PANACEA has been a EU FP7 project to study and develop solutions for a proactive autonomic management system of cloud resources, based on a set of advanced machine-learning techniques and virtualization. By predicting anomalies (time to failure of cloud applications and DDoS attacks) before they occur, PANACEA was meant to enforce the following autonomic properties:
- self-healing against anomalies by recovering from multiple node and link failures, and using proactive rejuvenation of applications and servers for preventing crashes and increasing the availability, predicting the threshold violation of response time of servers,
- self-configuring by efficiently mapping user’s requirements onto distributed clouds and configuring on-the-fly in the presence of anomalies,
- self-optimizing using proactive migration of virtual machines from one cloud resource to another, maintaining the quality of service of end-to-end flows,
- self-protecting using proactive reconfiguration of overlay networks to protect against DDoS attacks.
In this project, some members of Lockless S.r.l. have worked with the International Research Institute for Autonomic Network Computing (IRIANC). In particular, they have designed a framework which allowed to monitor in real time the state of Virtual Machines hosted on a large number of Cloud Resources. By means of prediction, load balancing, and traffic monitoring, the framework has been used to instradate clients’ requests towards the less-bloated nodes, thus reducing the response time and increasing the responsiveness of (distributed) web-based applications.
The members of this project have been:
- Centre National de la Recherce Scientifique (LAAS-CNRS), Tolouse, France
- Imperial College, London, UK
- IRIANC, Germany
- ATOS Research, Barcelona, Spain
- IBM Haifa, Israel
- QoS Design, Tolouse, France
- Universidad Computense de Madrid, Madrid, Spain.