Skip to content

autodocs_module_performance_modelling

Tony Woode edited this page Apr 8, 2015 · 1 revision

Module performance_modelling Owner: tim@mirriad Work package: WPx.y Description: this uses recorded performance data to generate and refine parametric performance models

Method call: database.get_facts() Implemented in: database Called by: performance_modelling Description: get generic facts, based on pattern matching Parameters: Name: f Type: fact_query Description: a query pattern for a simple fact, eg. 'available VMs = ?'

Returns: Name: f Type: A list of facts Description: a simple logical or numerical assertion in RDF-like notation

Responsible parties: jazz@imperial, neil@mediachannel, tim@mirriad Work packages: WPx.y

Method call: performance_predictor.query_performance_prediction() Implemented in: performance_predictor Called by: performance_modelling Description: this estimates the predicted performance of an action, based on the facts available to the inference engine Parameters: Name: a Type: action_spec Description: a specification of an action to be performed

Returns: Name: p Type: performance_prediction Description: a set of numerical predictions, likely accompanied by metadata

Responsible parties: james@mirriad, jazz@imperial, tim@mirriad Work packages: WPx.y

Method call: database.create_performance_rule() Implemented in: database Called by: performance_modelling Description: creates a performance rule in database Parameters: Name: p Type: performance_rule Description: a logical rule modeling system performance, either specified manually, or learned from recorded data

Returns: Name: u Type: uri Description: a Uniform Resource Identifier -- commonly used as unique identifiers for things like CRUDL

Responsible parties: james@mirriad, neil@mediachannel, sam@smoke, tim@mirriad Work packages: WPx.y

Clone this wiki locally