Closing the Gap between Cloud Providers and Scientific Users - Cloud Computing with e-Science Applications - page 116

6.5.4 Monetization ................................................................................... 133
6.5.4.1 Storage Cost ..................................................................... 134
6.5.4.2 Computational Cost ........................................................ 134
6.5.4.3 Additional Costs .............................................................. 135
6.6 Results ......................................................................................................... 135
6.7 Conclusions and Future Work ................................................................. 137
References ............................................................................................................. 138
Summary
Cloud computing emerges as an alternative to traditional grid/cluster
approaches. Particularly, software-as-a-service model can be an option to
address the computational needs of small- and medium-size research groups,
with little or no knowledge and resources to deal with the complexities of
technology. Although there are still many problems to be solved and a long
way to go before the solution is optimal, the e-Clouds project manages to
hide the configuration required by public infrastructure-as-a-service (Iaas)
providers by delivering ready-to-use scientific applications that take advan-
tage of the cloud world.
6.1 Introduction
Everyday scientific work requires growing computational capacity to pro-
vide reliable and in-time results. The traditional approach to address these
needs includes the acquisition, configuration, and maintenance of a large
number of dedicated servers, introducing some constraints primarily asso-
ciated with the elevated costs and complex information technology (IT)
management. These high-performance platform requirements are a barrier
to entry for small- and medium-size research groups.
Public cloud infrastructures present themselves as an alternative to tradi-
tional cluster and grid solutions [1]. Cloud providers offer a large set of infra-
structure and application services to resemble the flexibility of private data
centers, with the benefit of a pay-per-use model. This allows users to run a
wide variety of applications, including enterprise, social, and mobile ones.
The question is then: How to adapt this model for scientific requirements?
As we show, almost everything required by a scientific application is avail-
able in the cloud. The main challenge is then that, despite the low prices and
flexible set of resources, the complex deployment and execution procedures
are an obstacle for researchers to adopt the technology.