Technical Team/2018-12-4

December 4, 2018

Attendees

• Gary O'Neall
• James Neushul
• Alexios Zavras

SPDX Model Updates for SEVA fields

• Discussed the vulnerability relationship to SpdxItem, Package, and File
  • Most vulnerabilities will be associated with package
    • NIST NVD’s can only be associated with packages
  • There are some scanners that will associate vulnerabilities with files
  • Agreed that we should associate the vulnerability to Item so that it can apply to both Package and File
  • vulnerability information is optional (0 to many)

SEVA Vulnerability inclusion in SPDX

• Most of the information is based on the NIST NVD definitions
  • Modeled after the definition documentation
  • A schema for the NVD which includes all of the details (e.g. scoring) does not exist (or at least we could not find one)
  • Names do not match the NVD names to support distinguishing an official NVD schema if one to be found or produced
• Reviewed the SPDX-SECURITY schema at https://spdx-ccm.specchain.org/xsdccm/home
• There is also an SPDX-SEC-ISM schema which includes classification information (ISM)
  • We agreed that the ISM information would be valuable to commercial use, but can be a separate decision from the vulnerability information
• Discussed whether we should include all of the detail present in the SEVA document in the SPDX document or should we somehow reference an external document
  • There are a large number of fields to be reviewed
  • We are not vulnerability experts, and may not have the background to decide on a model and tools of our own
  • Agreed we should reference an external document
• We agreed that the vulnerability information should be included
• We discussed having a “proposal” or “working draft” similar to the W3C for the external document

Next Steps in Vulnerability inclusion in SPDX

• Gain a larger consensus on referencing an external document for vulnerabilities
• Draft the external document reference language
• Discuss / agree on the external reference content

Documentation for SPDX including SEVA

• James provided a document
• Produced from the XML document
• Extension to allow security classifications for each element (e.g. confidential)
  • Uses ISM schema
• Discussion on approach
  • Document in sections vs get experience with proposed sections prior to specifying
  • Document in sections first is the traditional approach for SPDX, may have “culture shock” switching to a more prototyping approach
  • Danger in specifying first – may create implementation challenges and may create specs that are not widely used
• Discussion on scope and prioritization
  • Large scope – many new sections and properties
  • Suggestion to prioritize by importance and ease of standardization
• Discussion on the modeling
  • Agreed modeling will be helpful
• Plan going forward:
  • Next week Mathew will review some of the new sections/areas being proposed and help prioritize which areas would be most useful
  • Next week’s call will focus on the use cases / prioritization
  • In 2 weeks, Gary will propose an update to the model which will include the higher priority areas/sections