Hogna: A Platform for Self-Adaptive Applications in Cloud Environments
by Cornel Barna, Hamoun Ghanbari, Marin Litoiu, and Mark Shtern.
Deploying and managing autonomic applications in cloud is a time consuming operation, that require many components to work together. The management will need to extract metrics from the deployed system, analyze them and the make a decision for changes that need to be implemented.
Usually, a researcher's work is focused in only one component (investigating different strategies for adaptation, evaluating the impact of various metrics, developing new methods to analyze data to uncover problems, etc.), while the rest must just work, without the researcher having to spend too much time on them.
Furthermore, the researcher should be able to easily compare the results of his idea with the results of other solutions for the same problem (in a meaningful way).
Comparing multiple solutions
A researcher that wants test his idea related to autonomic systems, can also compare it with other alternative solutions. Hogna offers the possibility to replace the component that implements researcher's idea with another component (for example: compare an adaptation strategy with another one). The two datasets obtained can be easily ploted and compared.
Test a solution multiple clouds
The validity of the researcher's idea can be tested on multiple clouds with Hogna.
Hogna offers a performance model, OPERA, to assist the researcher developing his management strategy.
Hogna includes a sample application that can be automatically deployed in Amazon EC2 cloud and managed. There are two sample management strategies:
- one uses simple threshold rules to trigger add/remove virtual machines.
- the second uses the model to decide how many VMs should be added/removed.
The sample application includes gnuplot scripts to visualize the data and a workload generator. A special preconfigured AMI has been made publicly available, and can be used freely.
The manual include in the archive contains the details on how to use and expand the functionalities of Hogna.
Download Hogna and the documentation of how to install and use it here:
- C. Barna, H. Ghanbari, M. Litoiu, and M. Stern, "Hogna: Platform for Self-Adaptive Applications in Cloud Environments", in Proc. of SEAMS'15, 2015, IEEE.
- M. Shtern, B. Simmons, M. Smit, and M. Litoiu, "An architecture for overlaying private clouds on public providers," in Proceedings of the 8th International Conference on Network and Service Management, ser. CNSM'12. Laxenburg, Austria, Austria: International Federation for Information Processing, 2013, pp. 371–377.
- M. Shtern, B. Simmons, M. Smit, and M. Litoiu, "Navigating the clouds with a MAP", in 2013 IFIP/IEEE International Symposium on Integrated Network Management (IM'13), May 2013, pp. 464–470.
- C. Barna, M. Shtern, M. Smit, V. Tzerpos, and M. Litoiu, "Mitigating dos attacks using performance model-driven adaptive algorithms", ACM Trans. Auton. Adapt. Syst., vol. 9, no. 1, pp. 3:1–3:26, Mar. 2014.
- C. Barna, M. Shtern, M. Smit, H. Ghanbari, and M. Litoiu, "Modeldriven elasticity and dos attack mitigation in cloud environments" in 11th International Conference on Autonomic Computing (ICAC'14). Philadelphia, PA: USENIX Association, Jun 2014, pp. 13–24.