1 / 15

Privacy Enhanced Storage

Privacy Enhanced Storage. Periodontitis Pilot Study. WP6.3 Periodontitis Pilot Study (ACTA) Development of web-based tool for export and import from/to periodontitis data warehouse Goal is sharing of periodontitis research data Exported data should be anonymous

trella
Download Presentation

Privacy Enhanced Storage

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Privacy Enhanced Storage

  2. Periodontitis Pilot Study WP6.3 Periodontitis Pilot Study (ACTA) • Development of web-based tool for export and import from/to periodontitis data warehouse • Goal is sharing of periodontitis research data • Exported data should be anonymous • Export as such can trivially be achieved • Another less trivial issue persists: local use of data for research purposes (direct access to database, use of a local analysis tool)e.g. students should not see nominative data • Case study in collaboration with ACTA:How to create a fully protected data warehouse within the periodontitis pilot • Fully protected; solutions should not be restricted to simple access control

  3. “Cryptographically Enforced” Access Control Encrypted Storage • Reduces overall vulnerability of the systemi.e. database leaks do not infringe privacy • Facilitates export to research database WP6.3 is an ideal test bed for proof-of-concept • Privacy wise non critical (dental records)Complete implementation thus not crucial • Relatively small scaleContained environment, relaxed performance constraints, … • Technological Barriers: • Deployment of related Custodix solutions has mainly been targeted against common web application frameworks (WP6.3 includes a thick-client application) • Delphi technology is not targeted by Custodix solutions (.net / J2EE)

  4. Case admin. information Example of Web Based System Record Nr. No Risk Record Nr. Identities of: • Doctor • Patient Case admin. information Encryption Case data Non-identifying case data No Risk • No identities of patient or doctor required for further processing • Can be completely cleared of identifying information(also “indirect identifying information”) Encrypted • Can only be viewed (decrypted) by a limited number of authorised peopleThe Database does not contain any identifying information(only illegible encrypted data)

  5. Case admin. information System Overview • Custodix Software • Software provides on-the-fly encryption and decryption of sensitive data • No unencrypted sensitive data leaves the data collection environment Database Custodix Software case-ID • Local Data Managers... • Each have a personal Keypair (Private/Public Key on Smart Cards) • form an information sharing pool • are the only ones who can view and edit all information in the database Non-identifying case data • Key Management Service • Supports key management • Does NOT contain unencrypted secrets (i.e. no security risk)

  6. Smart Card Login Functionality Integration in the Periodontitis Project Key Management Server Database Privacy Enabling Daemon Existing DELPHIApplication ADO ? Callbacks interwoven calls

  7. Integration Approach Main goal: Transparency for the application developer: • Limit impact on application model • Key management is handled outside the application by an independent key management service Identified solutions are based on proxy design pattern: • Proxy Database Driver (SQL Wrapper) (Data Access Layer) • Enhanced User Interface Controls (Presentation Layer)

  8. proprietary EncryptionDecryption jdbc interface proprietary proprietary SQL Wrapper SQL Wrapper (Data Access Layer) • What? • Direct processing on SQL queries and responses • Gives a high level of transparency for the application developers(Note that issues with searches on encrypted fields still persist) • Problem • The SQL standard does not describe a communication protocol. Each database has proprietary solution for communication. • Handling complex queries (e.g. joins) is not straightforward Proprietary Solution • Not feasible for this project Wrapper Driver (Stub - Bridge) • “generic” approach • Still technology bound(e.g. jdbc driver) Application

  9. Enhanced UI Control Enhanced User Interface Control (Presentation Layer) • What? • Actual UI control is wrapped inside a wrapping UI control. • Event-driven, i.e. wrapping UI control performs encryption/decryption on the fly when data update events are received. • Problems • Weak data typing • All data validation checks must be performed in presentation layer • Must not break common data access abstraction strategies • Data binding user controls (e.g. ADO) • Object-relational mapping tools (e.g. Hibenate)

  10. Searches on Encrypted Data • i.e. the ability to look up patient records based on partial identifying information • Encrypted data is opaque (unstructured and unordered) and is thus not trivially indexable • Monitoring and analysing queries can lead to partial or complete re-identification • Possible approaches: • (encrypted) indexing service hosted by TTP • Local caching of search indexes on the client (data collection) • Client retrieves encrypted indexing structure from server • Builds local indexing table • Drawbacks: • querying the database becomes more complex as it involves more client-side processing • Synchronisation issues in data sharing environments

  11. Searches on Encrypted Data Periodontitis Database is of limited size (~700 patients) • TTP solution involves too much operational overhead for the problem at hand • Searching on partial data is not main concern for this project

  12. Data Sharing • Multiple users need to be able to share encrypted data • Must be able to accommodate dynamically changing user groups (roles) • Must be efficient both in time and space • Trivial solution: encrypt data for each group member • Very storage inefficient • Very computation intensive • Does not enforce group membership in a cryptographic way (relies on access control measures) • Approach based on three level encryption key infrastructure (see next slide), involving: • Users • Roles • Confidentiality Units

  13. User X  { Role, Role } RBAC – Enforced through Encryption Content encrypted with the same secret Confidentiality Unit (CU) • Smallest unit of sharable data Roles • Determine “Access Control” • {Roles} have access to a CU Users • Can act in a number of {Roles} CU CU CU CU CU Note: • Suggested: Use one CU (coupled CUID) for each patient record • But not required ()

  14. Notes on Security This solution is secure because: • No identifiable information leaves the Client computer in unencrypted form • All identifying information is encrypted • Only a strict group of users can decrypt that information • Personal keys can be stored in (tamperproof) Smart Cards(increasing security even further) • Database Access Control failure does NOT compromise privacy of patients (identifiers are encrypted) • The Key Management Service is NOT a risk factor in the protection scheme (It only stores Public Keys and key management metadata in encrypted form)

  15. Thank you for your attention! Custodix NV Verlorenbroodstr. 120 B-9820 Merelbeke Belgium http://www.custodix.com/ or info@custodix.com

More Related