Above the Clouds: A Berkeley View of Cloud Computing

Michael Armbrust, Armando Fox, Rean Griffith, Anthony D. Joseph, Randy H. Katz, Andrew Konwinski, Gunho Lee, David A. Patterson, Ariel Rabkin, Ion Stoica, Matei Zaharia. Above the Clouds: A Berkeley View of Cloud Computing.

The authors identified three properties that gives Cloud Computing its appeal: short-term usage (which allows for scaling up/down resources based on demand), no up-front cost, and (an illusion of) infinite capacity on-demand. Due to the inherent elasticity and flexibility that cloud computing provides, any application that needs a model of computation, storage and/or communication fits perfectly in this paradigm.

A key decision for any application provider is to whether invest in the cloud on a pay-per-use basis or incur a one-time cost of setting up the whole datacenter. This paper highlighted the reasoning and rationale that should be put in before making those decisions in terms of long term economical benefits and being risk averse in case of load spikes.

Finally, the paper highlighted some key obstacles to the growth of cloud computing, and paired each of them with opportunities that may result:

1. Availability of the Service: Prevention against DDoS attacks, large scale failures.
2. Data Lock-In: Losing customer data due to dependence on a variety of service providers
3. Data Confidentiality: Making sure that the data is secure.
4. Data Transfer Bottlenecks: Tranferring user data to the datacenter economically.
5. Performance Unpredictability: Achieving disk I/O performance isolation even though the resources are shared.
6. Scalable Storage: Making sure that the store achieves scalability, data durability and high availability even as the size of data increases.
7. Distributed Debugging: Identifying bugs in Distributed Systems.
8. Scaling Quickly
9. Reputation Fate Sharing: Identifying adversaries, spammers.
10. Software Licensing: Solving challenges that result due to proprietary software licensing issues.