Quality Management in Service-based Systems and Cloud Applications
This project seeks to provide a system for quality management in service oriented architectures (QoS). The underlying problems to tackle in this environment are aligned to the next major topics: 1) modelling of the domain-knowledge in QoS: SLAs, metrics, dimensions, etc. and 2) tracking applications activity and making decisions according to user alerts in real time (queries) and prediction and prevention of future events (analytics).
Research Aims and Questions
The main aim of the project is to tackle the problem of managing cloud application platforms, more specifically the use of QoS metrics as indicator to handle cloud services. In this context, it is necessary to provide an infrastructure that can deal with heterogeneous data streams in real-time processing the information available from applications to deliver reactive analysis or prediction services in order to optimize the cost, reliability, scalability, etc. As a consequence a layer for unifying information and data in an intelligent fashion is required (semantic technologies) and Big Data techniques for processing large data streams in real-time are also necessary. According to this description, the specific objectives of the project can be divided into four different levels:
- Domain-knowledge. This abstract level must define the concepts and relationships to represent QoS features such as metrics, dimensions or SLAs. Ontologies and rules can be used to formalize domain-knowledge
- QoS management system. This level takes the domain-knowledge as parameter and designs services and algorithms to be deployed in the execution environment. Services go from a simple alert service to complex analytical algorithms.
- Execution environment. It is a real-time platform in charge of providing the infrastructure to execute the algorithms over large and diverse data streams for both exploitation real-time (queries) and analytics. It must provide continuous computation.
- Data Streams. This level is comprised of the data sources from which data is gathered.
In this context, this approach tries to give an answer to the next emerging questions:
- Which are the concepts and relationships to take into account in QoS management?
- Which services must be provided to exploit domain knowledge and which algorithms are necessary to afford those services?
- How can we deal with the processing of data streams (Big Data) in real-time?
- How can we make decisions in real-time according to user alerts (queries)?
- How can we exploit the historical information gathered from different datasources and feedback the alert system?
- An early prototype of a real-time platform for dealing with data streams and execute simple rules is now available (documentation and source code). The current effort is now focused in levels 3 and 4.
Alvarez-Rodríguez, J. M., Labra-Gayo, J., Ordoñez de Pablos, P.: Leveraging Semantics to Represent and Compute Quantitative Indexes: The RDFIndex Approach. In Metadata and Semantics Research. Proceedings of the 7th Research Conference, MTSR 2013, November 19-22, Thessaloniki, Greece, Springer 2013, pp.175-187, LINK
Labra-Gayo, J. E., Heuring, J., Alvarez-Rodríguez, J. M.: Inductive representations of RDF Graphs. In Science of Computer Programming. (In Press) Available online since Jan. 21 2014 LINK
Kourtesis, D., Alvarez-Rodríguez, J. M., Paraskakis, I.: Semantic-based QoS management in cloud systems: Current status and future challenges. Future Generation Computer Systems Vol. 32 March, 2014. pp. 307-323, LINK
Labra-Gayo, J. E., Heuring, J., & Alvarez-Rodríguez, J. M.: Inductive Triple Graphs: A purely functional approach to represent RDF. In Graph Structures for Knowledge Representation and Reasoning. Third International Workshop, GKR 2013, Beijing, China, August 3, 2013. Revised Selected Papers, Vol. 8232, Springer 2014, pp. 92-110, LINK
Labra-Gayo, J. E., Heuring, J. T., & Alvarez-Rodríguez, J. M. : Inductive Triple Graphs: A purely functional approach to represent RDF. Technical Report (No. UU-CS-2013-009), Department of Information and Computing Sciences, Utrecht University, 2013, PDF
Labra-Gayo, J. E., & Alvarez-Rodríguez, J. M.: Validating statistical index data represented in RDF using SPARQL queries. In Proceedings of the RDF Validation Workshop, Practical Assurances for Quality RDF Data, Sep 10-12, Cambridge, MA, USA, 2013, PDF