1. Effective Resources Provisioning for Hosted
SaaS application in cloud computing
George Jimaga James
May 26, 2019
College of Software Engineering, Nankai University
7. Introduction
• Cloud computing is the big shift from the traditional way
business think of resources utilization.
• Computing in cloud provides two major advantage to the end
users and cloud operator
• Flexibility
• Cost efficiency
2/11
9. Motivation
Machine Learning everywhere
Machine Learning is present in so many segments of technology,
that we dont even realise it while using it
ML Services
1. Image recognition in facebook Moments
2. Video analysis in YouTube captions
3. Correlation analysis in movie
recommendations
etc
ML Techniques
... Regression
... Classification
... Reinforcement
learning
etc
4/11
10. Motivation
Cloud Computing platform
Cloud computing is the on-demand availability of computer
system resources, especially data storage and computing power,
without direct active management by the user.
5/11
11. Objectives/Aim
1. Design algorithm to minimize SLA violation by resources
provisioning and request rescheduling
2. Minimize the total cost
3. Improve CSL
6/11
13. Literature Review
• [Reig et al., 2010] contributed on minimizing the resources
consumption for serving requests and executing them before
it’s deadline with a prediction system. Their prediction system
enables the scheduling policies to discard the service of a
request if the available resources capability is not able to
complete the request before it’s deadline.
• [Fu and Vahdat, 2002] proposed an SLA-based dynamic
scheduling algorithm of distributed resources for streaming. It
evaluated various SLA-based scheduling heuristics on parallel
computing resources resources utilization such number of
CPU nodes and income as evaluation metrics.
8/11
14. Literature Review
• [Lee et al., 2010]investigated the profit driven service request
scheduling for for dependent tasks without user-driven
consideration.
• [Popovici and Wilkes, 2005] mainly considered QoS paramters
on the resources provider’s side such as price and offered load,
but didn’t focus on the user side.
• [Chaisiri et al., 2011] proposed optimisation of resource
provisioning cost in Cloud computing by applying stochastic
programming approach in multiple phases. They minimized
the cost by considering the uncertainty
9/11
15. Thesis outline
1. Introduction
2. Related Work
3. Methodology
4. Experimental setup
5. Results and Discussion
6. Future Work
7. Conclusion
10/11
16. References
Chaisiri, S., Lee, B.-S., and Niyato, D. (2011).
Optimization of resource provisioning cost in cloud
computing.
IEEE transactions on services Computing, 5(2):164–177.
Fu, Y. R. and Vahdat, A. (2002).
Sla-based distributed resource allocation for streaming
hosting systems.
Lee, Y. C., Wang, C., Zomaya, A. Y., and Zhou, B. B. (2010).
Profit-driven service request scheduling in clouds.
In 2010 10th IEEE/ACM International Conference on Cluster,
Cloud and Grid Computing, pages 15–24. IEEE.
Popovici, F. I. and Wilkes, J. (2005).
Profitable services in an uncertain world.
In Proceedings of the 2005 ACM/IEEE conference on
11/11