Home
Forgot your password?
New user?
Oceans IE Phase II kickoff meeting
Data Turbine and OOSTethys toolkits integration
DataTurbine, as other real time distributed
system is asynchronous, while SOS is synchronous. In asynchronous
systems data lives on the wire, and other systems can publish or
subscribe data to the wire. In the case of DataTurbine the wire is
called buffered ring, publishers are called sources and subscribers are
called sinks. SOS is synchronous since there are operations that follow
request/response pattern, such as HTTP GET. For example describeSensor
returns an XML with a SensorML document.
There are advantages of
asynchronous real time distributed systems, such as speed of
transaction and scalability on peer-to-peer distributed systems.
However complex queries that required much more complex data models are
hard to achieve in asynchronous systems. For example, querying data
streams giving a geospatial extend and a specified procedure ( e.g. sensor ID).
The
following are the approaches to integrate asynchronous systems, such as
Data Turbine with synchronous systems, such as OOSTethys SOS JAVA
toolkit (hereafter SOS Server):
- Implementation of messaging
client in the back end of the SOS Server. In this case the SOS Server
instantiates a Data Turbine Sink for channels of interest. A harvester
stores the data in a data base so the SOS Server can respond to queries
that are not supported by Data Turbine, such as spatial extent or
filtering by property.
- Implementation of an observation offering
in an SOS server whose result can be access in asynchronous mechanism.
In this case the result in the observation doesn’t contains data, but
contains metadata of how to access DataTurbine ring buffer, such as
host string and channel name.
The ideal case would be to
construct an SOS server automatically from DataTurbine channels. Each
Channel has name, userInfo and Data. The userInfo can be used to set
metadata that can provide information to construct a getCapabilities,
sensorML or observationOffering metadata in the SOS Server.
Data Turbine model:
- ChannelMap has Channels
- Each Channel has
- name
- userInfo ( metadata ), Via the method: PutUserInfo method we could add SensorML
- The data to the channel could be added as follows:
- cmap.PutTime(secondsInDoubleUTC,0);
- cmap.PutDataAsString(chanIndex,"2009-06-10T18:28:41Z,36.69623,-122.39965,12.929,3.88978,33.100174"); This is the same as the data block in SweCommon
