Biomedical Engineering Reference
In-Depth Information
miological observatory to build epidemiological studies. Contrary to cancer
screening associations, personal information on the patient (especially name,
surname, and social security number) is not relevant for epidemiological struc-
tures to produce statistics; on the other hand, exhaustive and disambiguation
on the cancer information for each patient are required in order to produce
reliable statistics on cancer incidence in a region.
The proposed grid architecture is built upon a central server hosting secu-
rity features and core grid services:
• The AMGA (ARDA Metadata Grid Application) [12] server, which
provides a way to access and store metadata. Especially when dealing
with medical images, the use of metadata is mandatory therefore AMGA
is able to glue the DICOM servers with the grid middleware [13]. AMGA
is a very attractive software to fulfi ll the strong security requirements and
access right management of medical data in a grid infrastructure.
• Pandora GateWay is a set of software designed as a service-oriented
architecture (SOA) developed by the maatG company. Pandora GateWay
is used to address medical data accessibility, exchange, and processing
while guaranteeing a high level of security for sensitive data. The main
added value of GateWay, compared to a classic SOA platform, is the
high-level security. The GateWay authentication service is based on
several security checkpoints required for login. The access point is a two-
factor authentication based on user certifi cate and pin code followed by
an authentication process using a Virtual Organization Membership
Service (VOMS) grid proxy [14]. VOMS is an authorization manager
which implements a public key infrastructure (PKI)-based authentication
with certifi cates delivered by trusted certifi cation authorities (CAs):
GridFtp server for data transfer [15]
Logical fi le catalog (LFC) server, for data management, included in
the gLite middleware
Then, in each medical structure which takes part in the sentinel network, a
grid node is deployed with two interfaces:
• T h e fi rst one is linked to the medical database inside the private medical
structure network. This connection is an automatic export system
[Structured Query Language (SQL) query
standardized output]. It is
less intrusive, offers enlarged customization for integration and standard-
ization of data, and does not overload the medical database during
working hours.
• The second interface is linked to the sentinel network (Internet), enabling
external users to query the grid data server, according to the local security
and authentication policy fi xed by VOMS.
+
For a better readability of cytopathological data, data sheet standardization
is used to simplify the integration of medical sheets in the database without
interfering with other data.
Search WWH ::
Custom Search