Aside from 2014 ICPE conference and HotTopics workshop which commenced this year in Dublin, Ireland, SPEC Research Group (RG) also hosted its annual members meeting led by Chair Prof. Samuel Kounev where the DCU Modelling and Simulation RG was warmly invited. The broad goal of SPEC RG is to provide a link between industry and academia partners focusing on traditional system benchmark metrics such as performance, while also exploring security availability and elasticity. During the meeting all three working groups (Cloud, Intrusion Detection Systems (IDS) and newly formed Big Data) gave short overview of their vision, scope with achievement reports followed by the members presentations.
As one of the major outputs, Cloud RG presented “Cloud Usage Patterns – A Formalism for Description of Cloud Usage Scenarios” report (SPEC-RG-2013-001) compiled by collaboration between multiple partners in order to provide application domain independent common notation for usage of cloud resource provisioning services, see here.
Further Karlsruhe Institute of Technology (KIT) presented two ideas in terms of aiding Cloud benchmarking:
- Eclipse IDE based modelling platform for load intensity variations called LIMBO. This platform provides full UI supported functionality to create custom load intensity profiles, extract load intensity from captured traces and generate time-stamped log files as input parameters for benchmark tools such as JMeter, see here.
- Foundation work for resource elasticity benchmarking in cloud environments. The work focuses on precise definition of elasticity concepts and then introduces ways to measure the effectiveness of the solution. The idea was presented using CPU utilisation as an example, but can be also extended to suit other resources such as Network, Memory and I/O, see here.
To learn more about SPEC RG visit: http://research.spec.org