페이지 이력

...

CV data processing objects are described as a set of process controller (representing the computational functionality of registered execution resources) monitored and managed by a coordination service. The coordination service delegates all processing tasks sent to particular execution resources, coordinates multi-stage workflows and initiates execution. Data may need to be data staging onto be staged onto individual execution resources and results data persistence  persisted for future use; data channels can be established with resources via their process controllers. The following diagrams shows the staging and persistence of data.

...

Data processing requests generally originate from CV Presentation Objects#experiment_experiment laboratory which validate requests by invoking an  AAAI service . The  CV Presentation Objects#experiment_experiment laboratory will send a process request to a coordination service , which interprets the request and starts a processing workflow by invoking the required CV Component Objects#process_process controller. Data will be retrieved from the data store and passed to the execution platform, the  coordination service will request that a data transfer service to prepare a data transfer.

Data will be retrieved from the data store and passed to the execution platform, the  coordination service will request that a data transfer service to prepare a data transfer. The  data transfer service will then configure and deploy a CV Component Objects#data_data exporter which will handle the transfer of data between the storage and execution platforms, i.e. performing data staging. A data-flow is established between all required  CV Component Objects#data_data store _ controllers and  CV Component Objects#process_process controller via the  CV Component Objects#data_data exporter. After the data-flow is established, processing starts. Processing can include a host of activities such as summarising, mining, charting, mapping, amongst many others. The details are left open to allow the modelling of any processing procedure. The expected output of the processing activities is a derived data product, which in turn will need to be persisted into the RIs data stores.

Data Persistence

책갈피
data_persistence data_persistence

...

Data processing requests generally originate from CV Presentation Objects#experiment_experiment laboratory which validate requests by invoking an  AAAI service . The  CV Presentation Objects#experiment_experiment laboratory can present results and ask the user if the results need to be stored, alternatively the user may configure the service to automatically store the resulting data. In either case, after processing, the CV Presentation Objects#experiment_experiment laboratory will send a process request to the coordination service , which interprets the request and invokes the CV Component Objects#process_process controller which will get the result data ready for transfer.

The  data transfer service will then configure and deploy a CV Component Objects#data_data importer which will handle the transfer of data between the execution and storage platforms. A data-flow is established between CV Component Objects#process_controller and CV Component Objects#data_store_process controller and data store controller via the  CV Component Objects#data_data importer. After the data-flow is established, the data transfer starts. The persistence if data will trigger various curation activities including data storage, backup, updating of catalogues, requiring identifiers and updating records. These activities can occurs automatically or just as signals sent out to warn human users that an action is expected.