Speaker
C. Leggett
(LAWRENCE BERKELEY NATIONAL LABORATORY)
Description
It is essential to provide users transparent access to time varying
data, such as detector misalignments, calibration parameters and the
like. This data should be automatically updated, without user
intervention, whenever it changes. Furthermore, the user should be
able to be notified whenever a particular datum is updated, so as to
perform actions such as re-caching of compound results, or performing
computationally intensive task only when necessary. The user should
only have to select a particular calibration scheme or time interval,
without having to worry about explicitly updating data on an event by
event basis. In order to minimize database activity, it is important
that the system only manage the parameters that are actively used in
a particular job, making updates only on demand. For certain
situations however, such as testbeam environments, pre-caching of data
is essential, so the system must also be able to pre-load all relevant
data at the start of a run, and avoid further updates to the data.
In this talk we present the scheme for managing time varying data and
their associated intervals of validity, as used in the Athena framework
on ATLAS, which features automatic updating of conditions data
occurring invisibly to the user; automatic and explicit registration
of objects of interest; callback function hierarchies; and abstract
conditions database interfaces.
Primary author
C. Leggett
(LAWRENCE BERKELEY NATIONAL LABORATORY)
