Motivating example and problem analysis - Reliability Assurance of Big Data in the Cloud - page 19

Motivating example
and problem analysis

3

Cloud data storage cost and data reliability are two of the major concerns for storing
big data in the Cloud. The ultimate goal of this topic is to find a solution for providing
cost-effective data storage while meeting the data reliability requirement throughout
the Cloud data life cycle. Such a goal could substantially benefit data-intensive appli-
cations from eliminating the excessive storage cost for data reliability. The research in
this topic is originally motivated by real-world data-intensive applications for pulsar
searching in astrophysics, which could process and generate hundreds of terabytes of
data. For storing such massive scientific data in the Cloud, several challenges need to
be tackled.
In this chapter, we introduce the motivating example of our research as the scenario
for problem analysis and point out the challenges that need to be tackled. In Sec-
tion 3.1 , the motivating example of the pulsar searching survey is described by show-
ing a pulsar searching application instance in detail. In Section 3.2 , challenges related
to the storing the pulsar searching data in the Cloud are analyzed with four research
issues determined in detail. Finally, in Section 3.3 , we summarize the works presented
in this chapter.

3.1 Motivating example

The initial idea of this research was motivated by the astrophysics pulsar searching
surveys conducted by the pulsar group at the Centre for Astrophysics and Supercom-
puting at Swinburne University of Technology, using the observation data from Parkes
radio telescope, NSW ( http://www.parkes.atnf.csiro.au/ ) , which is one of the most
famous radio telescopes in the world. The application for pulsar searching surveys is
a typical data and computation-intensive scientific application. It contains complex
and time-consuming tasks and needs to process hundreds of terabytes of complicated
scientific data, which is of typical big data characteristics.

3.1.1 The pulsar searching application process
Figure 3.1 shows the process of the pulsar searching application at the high level in the
form of a work flow. There are three major parts in the pulsar searching process, which
are raw data recording, data preparation, and pulsar seeking [11] :

1. Raw data recording:

In the Parkes radio telescope, there are 13 embedded beam receivers for receiving raw signal data
from the universe. Raw signal data are recorded at a rate of 1 GB per second by the Australia
Telescope National Facility (ATNF) Parkes Swinburne recorder ( http://astronomy.swin.edu.au/
pulsar/?topic=apsr ). Depending on different areas in the universe where the scientists want to