A self-adaptation framework to address evolution and change in service-based cloud environments
As deployments of composite service-based cloud applications gradually evolve and mature, they are also expected to grow in size and complexity. Appropriate mechanisms and tools for the development of enforcement mechanisms through run-time monitoring are therefore essential to prevent a cloud application infrastructure from quickly dissolving into a non-reliable environment.
To address issues of evolution and change in cloud application platforms, we are investigating development of a self-adaptation framework, which would increase the overall degree of self-management and autonomicity at the Platform-as-a-Service (PaaS) level. Existing mechanisms mainly deal with self-management at the infrastructure level, whereas adaptation at the PaaS level (e.g., modifying the actual structure/behavior of an application at run-time) is currently beyond their capabilities.
A promising direction for our research is the MAPE-K reference model recommended by IBM. According to this generic conceptual model, an adaptation cycle is a continuous process consisting of 4 steps: Monitoring, Analysis, Planning and Execution. The model also includes the Knowledge base – a source of self-reflective knowledge, which is shared and accessed throughout the whole adaptation cycle.
Applying the MAPE-K model to the domain of self-adaptation in cloud application platforms helped us to identify potential research and implementation challenges, associated with the monitoring and analysis stages, as well as the functional properties of the future adaptation framework.
Based on the requirement of re-using as much of the existing work as possible, we propose to utilise the Semantic Web stack as an enabling technology. Accordingly, we intend to use (i) Web Ontology Language (OWL) ontologies to capture the self-reflective knowledge of the system (e.g. its internal structure, relations between components, etc.); (ii) Resource Description Framework (RDF) as a common format for representing data coming from sensors; (iii) RDF data streams to support continuously flowing data from sensors; (iv) continuous SPARQL-based query languages to encode critical condition patterns, and support situation assessment and detection of critical conditions; (v) OWL ontologies and Semantic Web Rules Language (SWRL) to support reasoning about possible adaptation actions. The Semantic Web technologies in the context of monitoring of data streams generated by sensors have already been employed by the Semantic Sensor Network research community. There are several OWL ontologies, developed by this community, which may be applicable to our domain, and therefore seem to be relevant.
W3C Semantic Sensor Network Incubator Group
Main research interests:
- Information and Knowledge Management
- Cloud Computing
- Autonomic Computing
- Service-Oriented Computing and
- Software Engineering
Dautov, R., Paraskakis, I., Stannett, M.: Utilising Stream Reasoning Techniques to Create a Self-Adaptation Framework for Cloud Environments. In Proceedings of the 6th IEEE/ACM International Conference on Utility and Cloud Computing (UCC 2013), Dec 9-12, Dresden, Germany, 2013.
Dautov, R., Stannett, M., Paraskakis, I.: On the Role of Stream Reasoning in Run-time Monitoring and Analysis in Autonomic Systems. In Proceedings of the 8th South East European Doctoral Student Conference (DSC 2013), Sept 16-17, Thessaloniki, Greece, 2013, pp. 247–258, PDF
Dautov R., Paraskakis I.: A vision for monitoring cloud application platforms as sensor networks. In Proceedings of the 2013 ACM Cloud and Autonomic Computing Conference (CAC ’13), August 05 – 09, Miami, FL, USA — ACM New York, NY, USA 2013, Art. 25, LINK
Dautov R., Paraskakis I., Kourtesis D., Stannett M.: Addressing Self-Management in Cloud Platforms: a Semantic Sensor Web Approach. In Proceedings of the International Workshop on Hot Topics in Cloud Services (HotTopiCS 2013), April 20, Prague, Czech Republic, 2013, pp. 11-18, LINK
Dautov R.: An ontology-driven approach to self-management in cloud application platforms (Poster). In 9th Summer School on Ontology Engineering and the Semantic Web, July 8-14, Cercedilla, Spain, July 2012, PDF
Dautov R., Paraskakis I., Kourtesis D.: An ontology-driven approach to self-management in cloud application platforms. In Proceedings of the 7th South East European Doctoral Student Conference (DSC 2012), Sep 24-25, Thessaloniki, Greece, 2012, pp. 539–550, PDF
Dautov R.: Literature review on Ontologies. South-East European Research Centre. Thessaloniki, Greece, 2012.
Dautov R.: Literature review on Stream Reasoning. South-East European Research Centre. Thessaloniki, Greece, 2012.
Dautov R.: Literature review on Autonomic Computing. South-East European Research Centre. Thessaloniki, Greece, 2012.
Hamadache K., Zerva P., Polyviou A., Simko V., Dautov R,. Gonidis F., Paez Anaya I.: Cost in the Cloud Rationalisation and Research Trails. In Proceedings of the 2nd International Conference on Advanced Cloud and Big Data (CBD 2014), November 20-22, Huangshan, Anhui, China, 2014 (To appear)