CloudPick: a framework for QoS-aware and ontology-based service deployment across clouds
Dastjerdi, AV and Garg, SK and Rana, OF and Buyya, R, CloudPick: a framework for QoS-aware and ontology-based service deployment across clouds, Software: Practice and Experience, 45, (2) pp. 197-231. ISSN 0038-0644 (2015) [Refereed Article]
The cloud computing paradigm allows on-demand access to computing and storage services over the Internet.
Multiple providers are offering a variety of software solutions in the form of virtual appliances and
computing units in the form of virtual machines with different pricing and QoS in the market. Thus, it is
important to exploit the benefit of hosting virtual appliances on multiple providers to not only reduce the
cost and provide better QoS but also achieve failure-resistant deployment. This paper presents a framework
called CloudPick to simplify cross-cloud deployment and particularly focuses on QoS modeling and
deployment optimization. For QoS modeling, cloud services have been automatically enriched with semantic
descriptions using our translator component to increase precision and recall in discovery and benefit from
descriptive QoS from multiple domains. In addition, an optimization approach for deploying networks of
appliances is required to guarantee minimum cost, low latency, and high reliability.We propose and compare
two different deployment optimization approaches: genetic-based and forward-checking-based backtracking.
They take into account QoS criteria such as reliability, data communication cost, and latency between
multiple clouds to select the most appropriate combination of virtual machines and appliances. We evaluate
our approach using a real case study and different request types. Experimental results suggest that both
algorithms reach near-optimal solution. Further, we investigate the effects of factors such as latency, reliability
requirements, and data communication between appliances on the performance of the algorithms and
placement of appliances across multiple clouds. The results show the efficiency of optimization algorithms
depends on the data transfer rate between appliances.