Skip to content

Terminology

Nate Weisz edited this page Oct 5, 2015 · 2 revisions

The following lists key herd terms and their descriptions.

Term Description
Namespace The Namespace provides a organization of Business Object Data within the herd system. The Namespace will "own" the data and CRUD services for the Namespaces will be provided in the future.
Business Object Data A representation of actual data and its associated files that gets registered for a specific trade date. Business object data is associated with a business object format and indirectly a business object definition. It also consists of a list of storage units and their associated list of storage files. Each time business object data is registered, a new version is implicitly created.
Business Object Definition The name and associated metadata definition information of a business object (e.g. NEW_ORDER, ORDER_ROUTE, FINAL_ORDER etc.) that will ultimately have registered data associated with it. Business object definitions are associated with a data provider and an application.
Business Object Format Information that defines the format of the business object data for a specific business object definition. This consists of a usage, file type, schema, etc. A schema adds structure to the data so it can be properly interpreted down to each field. Different versions of a business object format are tracked using an incrementing business object format version number.
Data Provider An entity that provides business object data (e.g. EXCHANGE_A, EXCHANGE_B, LICENSED_TRADER_A). A data provider is associated with a business object definition. This is reference data.
File Type A type of file that business object data can be stored in (e.g. TXT, ORC, SEQ). This is reference data.
Storage An instance of storage that is associated with a storage platform (e.g. a specific instance of HDFS along with its associated access credentials, etc.).
Storage File A single fully qualified file path that points to a portion of business object data.
Storage Platform A platform name that has the ability to store data (e.g. S3, HDFS, GP, NZ). This is reference data.
Storage Unit A grouping of storage files associated with a specific storage.
Partition A partition in herd would be like a row in a database table (and the object would be the table). The partition values would represent primary keys for different rows and the data files would be the data for the column.
Expected Partition Value In order to facilitate queries about data availability, the herd system has the ability to define multiple sets of "Expected Partition Values" which allow queries to be run to ask what data is registered against the expected partitions within a range. For example, consider a trade calendar which defines expected business days - then queries can be made to determine what data files are registered between Date A and Date B taking into account expected business days.
Usage The usage of business object data (e.g. RAW, PRC) that is associated with a business object format.
Job A job within herd is synonymous with an Activiti workflow.

</ac:rich-text-body></ac:structured-macro>

Clone this wiki locally