| Paper | Title | Page |
|---|---|---|
| TOPA01 | Data Management at JET with a Look Forward to ITER | 74 |
|
||
| Since the first JET pulse in 1983, the raw data collected per ~40s of plasma discharge (pulse) has roughly followed a Moore's Law-like doubling every 2 years. Today we collect up to ~10GB per pulse, and the total data collected over ~70,000 pulses amounts to ~35TB. Enhancements to JET should result in ~60GB per pulse being collected by 2010. An ongoing challenge is to maintain the pulse repetition rate, data access times, and data security. The mass data store provides storage, archiving, and also the data access methods. JET, like most fusion experiments, provides an MDSplus (http://www.mdsplus.org) access layer on top of its own client-server access. Although ITER will also be a pulsed experiment, the discharge will be ~300-5000s in duration. Data storage and analysis must hence be performed exclusively in real time. The ITER conceptual design proposes a continuous timeline for access to all project data. The JET mass data store will be described together with the planned upgrades required to cater for the increases in data at the end of 2009. The functional requirements for the ITER mass storage system will be described based on the current status of the ITER conceptual design. | ||
|
Slides | |
| TOPA02 | SDA Time Intervals | 79 |
|
||
|
SDA (Sequenced Data Acquisition) Time Intervals is a hierarchical logging system for describing complex large-scale repeated processes. SDA has been used extensively at Fermilab* for fine tuning during the Tevatron Collider Run II. SDA Time Intervals is a new system born during discussions between CERN and FNAL about routinely recording relevant data for the LHC. Its main advantages are extremly low maintenance and good integration with traditional "flat" dataloggers. The Time Intervals (TI) system records the time of key events during a process and relates these events to the data that the traditional datalogger archives. From the point of view of the application program, any number of datalogging systems can be refactored into human-understandable time intervals.
* SDA-based diagnostic and analysis tools for Collider Run II. T.B. Bolshakov, P. Lebrun, S. Panacek, V. Papadimitriou, J. Slaughter, A. Xiao. Proceedings of PAC 05, Knoxville, Tennessee, May 2005. |
||
|
|
Slides | |
| TOPA03 | The IRMIS Universal Component-Type Model | 82 |
|
||
| The IRMIS toolkit provides a relational description of the accelerator/facility hardware and how it is assembled. To create this relational model, the APS site infrastructure was successively partitioned until a set of familiar, "unit-replaceable" components was reached. These items were grouped into a set of component types, each characterized by the type's function, form factor, etc. No accelerator "role" was assigned to the components, resulting in a universal set of component types applicable to any laboratory or facility. This paper discusses the development of the universal component-type model. Extension of the component types to include port definitions and signal-handling capabilities will be discussed. This signal-handling aspect provides the primary mechanism for relating control system software to accelerator hardware. The schema is being extended to include references to the device support for EPICS-supported component types. This suggests a new approach to EPICS database configuration in which the user, after selecting a particular hardware component, is provided with links to the support software to be used in building the EPICS application. | ||
|
|
Slides |