Objectives

To understand how an ICT application and its underlying infrastructure performs at any given point in time is crucial for any advanced system for autonomous management the application and resources. Therefore, the objective is to develop methods and software for the on-line creation of knowledge relevant for resource management optimization. Relevant knowledge includes primarily identifying anomalies in performance and functionality as well as the identification of what are the root causes the problems. This may, for example, include identifying performance bottlenecks occurring as a consequence of load variations or other applications compete for the same resources; misbehaving software components due to incompatible updates; or unexpected behaviour due to intrusion or malicious software. Secondary, it involves devising corrective actions to resolve the issues. The knowledge will be derived from vast amounts of basic monitoring data, extending on state of the art data analytics and machine learning techniques. Projected future needs of Elastisys customers is will contribute to the detailed project specification.

Expected results

  • Methods and software for on-line detection of performance and functionality anomalies.
  • Methods and software for identifying resource bottlenecks and anomaly root causes.
  • Methods and software for automatically resolving performance bottlenecks and anomaly root causes for a few common such issues.
  • An integrated system for the above functionality, demonstrated in real-world use cases.
  • Software for visualizing the system status including anomalies, bottlenecks, and root causes as well as support for dynamic adjustment of relevant policies and thresholds.

 

Location

Elastisys is a spin-off company from the successful cloud resource management research at Umeå University, Sweden. Elastisys provides products and services for scalable and responsive IT systems, with an emphasis on auto-scaling with multi-cloud capabilities. Elastisys products and services extend on decades of internationally leading research in distributed systems, high performance computing, and autonomous management of virtualized resources.

Planned secondments

14 months at Université de Rennes 1 (France); 7 months at Las Naves (Valencia, Spain).

Conditions

Please write your application in English, as it will be examined by multiple people in different European countries. Please do not apply to more than two positions within the FogGuru project.

Apply Online

Fields with (*) are compulsory.