Horizonte needs Customer Care to see information to solve queries, that Finance approves charges and that analysts work without unnecessary personal data. To protect data requires to relate purpose, identity, scope and evidence.
Key concepts
Data Security Protects confidentiality, integrity y availability. To avoid misreading isn't enough if someone modifies bills without permission or if authorized users are unable to access them as they require them. The program starts with risks and data value for each process.
Authenticate confirms an identity. Authorisation determines what you can do in a context. Figure It protects the contents by means of keys. Audit preserves evidence of activity. A cipher connection can carry an excessively privileged consultation: mechanisms are complemented and should not be interacted with an explanation.
The minimum privilege grants what's necessary for an assignment. The minimization reduces the data treated for the purpose. An analyst may require amounts by segment and period without names, identity papers and addresses. An authorized account does not involve permission for any future use.
Access, Protection and Evidence
The classification relates sensitivities and context to controls. No public data have any permitted use. A combination of attributes can allow to identify people. The owner defines useful categories and revisions them and lays down treatment for copies, exports and derivatives.
Mask modifies or replaces sensitive values; submitting select a representative subset. To reduce rows does not automatically cover up their identities and to mask columns does not necessarily reduce their volume. A combination of both may be necessary in tests, while maintaining the necessary relationships and evaluating residual risk.
The segregation of functions prevents the concentration of incompatible decisions: those who manage permissions do not have to approve their own exceptional access. The review of accesses and the management of exceptions are part of the cycle. Each authorisation should have an appropriate risk-appropriate scope, reason and review mechanism.
Security at Oracle AI Data Platform
Configure Service Access with Oracle Cloud Infrastructure Identity and Access Management and assigns to Oracle AI Data Platform permissions about each object. For Horizonte's analyst, give an authorized product reading and check the inherited permissions. for engineering, assigns necessary actions about load processes. Use Audit Log to consult registered operations and to link them with identity, object and result.
To protect sensitive attributes while preparing, implement the government flow of sensitive data about Oracle AI Data Platform. It sets up rules for classification and treatment, introduces recommendations to steward and implements decisions with protected views or masked copies. Publish these results with permissions by role and regularly compare the configuration with the approved policy. That flow links the classification and minimum privilege of DAMA with engineering tasks and evidence of its compliance.
In Oracle Data Safe, records a copy of the database for testing and sets up Data Subsetting: select the initial table, filters and treatment of related tables. It combines politics with masking to reduce rows and protect values. Check the work report and check that customers, contracts and bills maintain the necessary terms and conditions for testing.
Explanated example
The Test Team at Horizonte requests a complete copies of customers and bills. It aims to check reconciliation and relations and to avoid identifying real people. The owner agrees to a representative set with regular cases and exceptions. The custodian retains client-contract-invoice relations and protects sensitive attributes.
At a supported base, Oracle Data Safe can help with subsetting and masking. The appropriate treatment is designed for other assets. Before to deliver, the computer checks that the set allows to test the stage and that does not put out unnecessary identifiers. The approval and checks are registered. call an environment "evidence" does not change the risk of content.
Common
- The figure equals the right authorisation. Check identities, roles and scope of reading or modification.
- A small sample's anonymous. It can preserve sensitive identifiers and combinations.
- A proposal generated by AI is an approved policy. The owner should validate scope, exceptions and application.
Decision exercise
A assistant from Customer Care has access to the workspace and needs to consult contracts. The computer propose to give permissions as an administrator because the connection already uses encryption. What decision reduces risk and preserves target?
Solution
Define necessary consultation, give the minimum permission about data and tools, review the effective identity of the connection and check the result. The encryption does not warrant general administration. There should be evidence of access and an appropriate review of use. The Winterspace permissions alone do not describe all of the authorisation.