Closing the Gap between Cloud Providers and Scientific Users - Cloud Computing with e-Science Applications - page 138

of this chapter. Additional work needs to be done to cover potential flaws
and support scientific applications that are more complex. Further testing is
required to adapt the solution to changing requirements and diverse research
groups' needs.
Future plans for the e-Clouds project include the implementation of new
features to favor collaboration among researchers and results validation
[37]. This way, the platform can become part of scientific day-to-day work.
In addition, new applications with different technical requirements will be
tested, including large-scale and long-lasting executions. In this respect,
there is some pending development regarding reliability and error handling.
Although the RM is capable of handling a minimum degree of parallelism,
several improvements in both the front and back end need to be done to sup-
port the execution of highly parallel applications (using a message passing
interface or graphics processing units) with effective resource management.
Together with this, additional work is required to support application work-
flows transparently. Some already existing alternatives are being considered
to support these requirements.
Finally, further optimization of resource scheduling is required to apply
data-mining techniques to estimate execution time and cost and take advan-
tage of the residual time of clusters and VMs. Although there is an impor-
tant challenge in proposing a general solution, some opportunistic ideas are
applicable to the e-Clouds scenario.
References
1. Rehr, J. J., F. D. Vila, J. P. Gardner, L. Svec, and M. Prange. Scientific Computing
in the Cloud. IEEE , 12, no. 3 (2010): 34-43.
2. Jackson, K, et al. Performance analysis of high performance computing appli-
cations on the Amazon Web Services Cloud. 2010 IEEE Second International
Conference on Cloud Computing Technology and Science (CloudCom), 2010: 159-168.
3. Wang, Lei, Jianfeng Zhan, Weisong Shi, and Yi Liang. In cloud, can scientific
communities benefit from the economies of scale? IEEE Transactions on Parallel
and Distributed Systems 23, no. 2 (2012): 296-303.
4. Anderson, David. BOINC: a system for public-resource computing and storage.
In Fifth IEEE/ACM International Workshop on Grid, Pittsburgh: ACM, 2004, 4-10.
5. Andrade, Nazareno, Walfredo Cirne, Francisco Brasileiro, and Paulo Roisenberg.
OurGrid: an approach to easily assemble grids with equitable resource sharing.
Lecture Notes in Computer Science 2862/2003 (2003): 61-86.
6. Goldchleger, Andrei, Fabio Kon, Alfredo Goldman, Marcelo Finger, and Germano
Capistrano Bezerra. InteGrade: object-oriented grid middleware leveraging
the idle computing power of desktop machines. Concurrency and Computation:
Practice and Experience 16, no. 5 (2004): 449-459.