220
SYLVIE POULIQUEN
The advantages of such a system are:
¾ One stop shopping for the users where they get the best available
data for ARGO in an unique format
¾ Data discovery and sub-setting tools are easy to implement as all the
data are in the same place
¾ A robust system, as the probability that both GDACs fail is very
small
¾ Easy to guaranty a quality of service in data delivery because GDAC
have the control of all the elements in-house
The disadvantages are:
¾ Data are moved around the network and must rely on the
"professionalism" of the DACs involved in the system to be sure
that GDACs have the best profiles available.
¾ Additional work at DAC level to convert their data from their home
format to the ARGO format. This may be hard to do for small
entities.
¾ Data format used for data exchange cannot evolve easily as it
requires coordination among all actors before implementation. Since
users, especially operational ones, do not like format changes it is
not such a big problem.
¾ If only one main server is set up than the system is fragile. Setting
up a mirroring system can over pass this problem with additional
synchronisation mechanisms.
5.1.2
Ocean US: Distributed data processing and distribution
The USA are developing an Integrated Ocean Observing System (IOOS)
ranging from global to regional to coastal. The purpose is to integrate
existing and planned observing systems that address both research and
operational needs. Considering the diversity of actors and of parameters
involved, this system must be a cooperative integration of independent
systems that will continue their missions independently while participating
in an integrated data system.
It is clear that in such a system the data processing is distributed and the
data stay on physically distributed repositories, some containing huge
amounts of data. The user connecting to the Ocean.US website will be able
to query for data without knowing where they physically reside.
The key elements of such a system are the metadata management, the
data discovery system and the data transport protocols.
SYLVIE POULIQUEN
The advantages of such a system are:
¾ One stop shopping for the users where they get the best available
data for ARGO in an unique format
¾ Data discovery and sub-setting tools are easy to implement as all the
data are in the same place
¾ A robust system, as the probability that both GDACs fail is very
small
¾ Easy to guaranty a quality of service in data delivery because GDAC
have the control of all the elements in-house
The disadvantages are:
¾ Data are moved around the network and must rely on the
"professionalism" of the DACs involved in the system to be sure
that GDACs have the best profiles available.
¾ Additional work at DAC level to convert their data from their home
format to the ARGO format. This may be hard to do for small
entities.
¾ Data format used for data exchange cannot evolve easily as it
requires coordination among all actors before implementation. Since
users, especially operational ones, do not like format changes it is
not such a big problem.
¾ If only one main server is set up than the system is fragile. Setting
up a mirroring system can over pass this problem with additional
synchronisation mechanisms.
5.1.2
Ocean US: Distributed data processing and distribution
The USA are developing an Integrated Ocean Observing System (IOOS)
ranging from global to regional to coastal. The purpose is to integrate
existing and planned observing systems that address both research and
operational needs. Considering the diversity of actors and of parameters
involved, this system must be a cooperative integration of independent
systems that will continue their missions independently while participating
in an integrated data system.
It is clear that in such a system the data processing is distributed and the
data stay on physically distributed repositories, some containing huge
amounts of data. The user connecting to the Ocean.US website will be able
to query for data without knowing where they physically reside.
The key elements of such a system are the metadata management, the
data discovery system and the data transport protocols.