Bitfount Glossary

Term	Definition
Pod	“Processor of data” - A Pod is the component of the Bitfount network which allows for models or queries to run on remote data. Pods are co-located with data, check that users are authorised to perform a given operation, and then execute any approved computation.
Pod policies	A policy is a rule applied to a Pod to dictate who has access, what actions can be taken on a given Pod, and if and how to apply privacy-preserving techniques to training or query results. Bitfount enables access and usage policies with additional policy permissions layered on top.
Pod authorisation	The process of Pod authorisation enables a Pod creator to provide access to other Bitfount users. In authorising a Pod, the Pod creator dictates who has access, what actions they can perform, and how they can perform them.
Pod roles	A role is a collection of permissions granted to a user in relation to their access of a given Pod. You can think of the role as dictating what a user can do once granted access to a Pod.
Pod permissions	Permissions refer to the set of actions a user can take once given access to a Pod and how those actions can be performed. Permissions include which types of tasks can be performed and what privacy-preserving techniques should be applied.
Data Custodian	The Data Custodian is the user or organisation who stores the relevant dataset for collaboration on their servers or local machines. Typically, Data Custodians grant analysis permissions (without granting direct access to raw data) to Data Scientists for two basic use cases: 1. External: A Data Custodian wishes to collaborate with an external party (or several) without transferring data. 2. Internal: A Data Custodian wishes to apply privacy-preserving or usage-based policies to data within their organisation without giving internal data scientists direct access to raw data.
Data Scientist (modeller)	The Data Scientist is the user who wishes to access datasets to perform a given task.
Task	Task is a catchall term used at Bitfount to describe the activity a user wishes to perform on a dataset associated with a given Pod.
Task Element	A task element is an input required to execute a given task as defined by the activity a user wishes to perform. Protocols, algorithms, and models are all considered task elements.
Protocol	A protocol defines how communication is orchestrated between Pods and the Data Scientist.
Algorithm	An algorithm defines what computation is done on each Pod at each iteration of a protocol.
Model	A model is a parameter to Model-based algorithms. For this class of algorithm, the model is the model architecture used for doing the analysis.
Transformation	A transformation is a parameter used to manipulate datasets prior to task execution to ensure their format matches the required format for the task.
Secure Aggregation	A privacy-preserving technique, based on secure multi-party computation, which allows summations to be computed without any of the individual values being known to the modeller
Card	The name for the visual element in the Bitfount Hub in which a Pod or model’s metadata is organised.