One of the challenges of
information systems is the accessibility of products residing in
different remote servers and compose them to produce more complex
information products. Problems in achieving this goal are related to
legacy of
different information systems that includes formats, applications,
schemas, vocabulary,
technologies.
Furthermore, the products in electronic form are becoming fairly
complex, consisting of files (ASCII, XML, HTML, Word, etc.),
interactive animations (e.g.
Macromedia Flash), sound, video, images or embedded programs (e.g. Java
applets).
Four tiers can be seen as
a base
of a distributed system:
|
The product tiers,
formed by the servers containing the data; these could be in different
formats, such as netCDF, HDF, ASCII, etc. |
|
The product access and inventory tier, the access being
provided by servers such as HTTP, and the inventory by catalog services.
|
|
The product catalog tier, allowing the selection of
data. |
|
The product presentation tier, that is
constituted by portals and must allow
different services, such as visualization. |
In the case of harmonized components, the architecture to access
products in a distributed system can be easily defined. But in general
the architectures of data management systems residing in the different
places have a high heterogeneity. A ‘transformation
service’ is important in a system allowing access to data
residing also in other data management systems, in order to transform
queries and data between different logical scheme.
Different approaches can be adopted to access data residing in
different servers, solutions are linked to the existence of (inter
alia) common data policy, schema, vocabularies. The approach selected
to built the information system that will be presented in this paper,
is based on the idea to have a portal of portals, in which the
heterogeneity is maintained. The main portal will contain all the
information on metadata of products residing in the other portals, but
property and download services will remain in the ‘product
owner portal’. Vocabulary can be different (within certain
limits) in the different portals, but the main will contain a
trans-coding table.
Schema can be also different, it is only necessary to have access to
particular final products. However, it is necessary to have a minimum
common content in metadata, in order to allow queries by users. Format
of metadata can be different, since it is transformed in central
system. The disadvantage of a such approach is on the fact that the
inclusion of new portals requires the implementation/adaptation of the
‘transformation’ software. The advantage is on the
fact that the ownership of products remains in the original sites, and
this does not create conflicts. Furthermore, the existing data
management systems can be maintained. In operational sense, the
‘federation’ of different portals with their
heterogeneity is effective since it assures the continuous update of
portals content by local data managers, does not delay the response to
queries, and assure the maximum profit both to data holders and users.
In this paper it is demonstrated that most of the problems to access
products residing in different remote servers can be superseded in an
effective way, using free software (PHP, MySQL, etc.) actually used in
many information systems. The example is provided by the so called
Mediterranean Operational Network (MOON) – Volunteer
Observing Ship (VOS) Portal.
In September 1999, a Ship Of Opportunity Program was launched in the
Mediterranean with the aim to provide data to the Mediterranean
Forecasting System, i.e. the preoperational model forecasting the 3D
circulation, temperature and salinity of the Mediterranean Sea. For
this reason it was developed a portal for the access of data by users.
During the last decade, the scope was enlarged in order to reach the
need of a wider community of users: operational systems, climatic
variability, research, public companies, students and the public.
Furthermore, the portal has been implemented to access also products
residing in remote servers. Actually access to products deriving from
in situ and satellite data is offered to users.
The portal is providing the services defined in the INSPIRE directive:
discovery, view, downloading, transformation.
The implementation of the portal
has been based on
specific rules on
mission, quality and business:
|
Mission:
definition
of the general objectives; thematic information to be managed by the
network; definition, implementation and maintenance of services; |
|
Quality: quality
assessment and QC procedures for data, products and services;
|
|
Business:
specific data policy; pricing (if needed); ownership; use rules. |
To facilitate the discovery service, a friendly user interface based on
Google maps has been developed. This allow also to zoom on particular
areas, maintaining the necessary geographical information.
In order to organize any information system, it is necessary inter alia
to have additional definitions. In our case it is important to define
‘products’ and ‘services’:
- product
is anything that can be offered to a client and that might satisfy a
want or
need (reference)
- service
refers to a discretely defined set of technical functionalities