HCloud, a Healthcare-Oriented Cloud System with Improved Efficiency in Biomedical Data Processing - Cloud Computing with e-Science Applications - page 166

evaluations based on the testing demonstrated the effectiveness and usabil-
ity of the system.
This remainder of the chapter is organized as follows: The application
scenarios, architecture, and key components of HCloud are described in
Section 8.2, which also provides the details of the data analysis services in
HCloud. Section 8.3 gives the details of the Map-Reduce paradigm immersed
in the platform, as well as the health care services that HCloud can provide.
Section 8.4 provides information on performance testing and evaluation;
a conclusion is drawn in Section 8.5.
8.2 The HCloud Platform
In recent years, researchers have made some useful attempts to implement an
efficient health care system with the power of cloud computing. For example,
Zhang et al. [10] proposed a cloud security model based on EHRs that belongs
to an MIS. Narayanan et al. [11] discussed access control to the health care
system by considering role task management. Chang et al. [12] proposed an
ecosystem approach to solve patient-centric health care and evidence-based
medicine. However, previous works mainly focused on the storage, access,
and management of private health information, which are quite primitive
applications regardless of the computational power of the cloud platforms.
It is expected that a cloud-based system not only stores the information but
also performs basic analysis of health status and provides useful advice or
warnings to patients, which is the purpose of our work.
8.2.1 Challenges to the Cloud Platform for Health Care
Cloud computing inherited the features of high-performance parallel com-
puting, distributed computing, and grid computing and further developed
these techniques to achieve location transparency to the end user and
improve user experiences. In addition, a general cloud platform must face
some challenges in health care service areas, as discussed next.
8.2.1.1 Heterogeneous Physiological Data Access
One challenging task for the health care cloud system is to handle the multi-
modal and nonstationary characteristics of special physiological signals,
such as those for HBP, electrocardiography (ECG), and photoplethysmogra-
phy (PPG). It is quite an inefficient job for a cloud system to store the numeric
small-size physiological signal data on the ordinary distributed file system.
Most of the distributed file systems are more suitable for large-size file stor-
age than for small-size storage because there are bottlenecks for small-size