IVY

Summary

The IVY Working Group met at IETF 126 to discuss the progress of its core data models, passive network inventory, and entitlement inventory, as well as a report on the recent IVY Hackathon and a new proposal for cryptographic asset discovery.

Key milestones include the core inventory model being near the final publication stage, and draft-ietf-ivy-network-inventory-topology having passed Working Group Last Call (WGLC). Both draft-ietf-ivy-network-inventory-software and draft-ietf-ivy-network-inventory-location are maturing quickly, with early YANG Doctor reviews being requested. The group also discussed the scoping of passive inventory and a proposal to introduce Cryptographic Asset Discovery and Inventory (CADI) into the Operations and Management (OPS) area.

Key Discussion Points

1. Working Group Status & Admin

Slides: 00-admin slides
Presenter: Qiufang Ma & Daniele Ceccarelli
The core model has been returned to the authors for final edits to address comments from Mahesh Jethanandani, with plans to submit to the IESG shortly.
draft-ietf-ivy-network-inventory-topology successfully passed WGLC. Olga Havel has prepared the shepherd write-up, which will be submitted once the authors publish the final version addressing recent feedback.
A formal liaison relationship with the OIF has been established through CCAMP, with Italo Busi appointed as the liaison manager.

2. Network Inventory Software Extensions

Slides: draft-ietf-ivy-network-inventory-software
Presenter: Bo Wu
Document: draft-ietf-ivy-network-inventory-software
Updates:
- Revisions 03 and 04 addressed Adrian Farrel's thorough review.
- Added a new "Operational Considerations" section to clarify the read-only reporting approach (consistent with the base model).
- Refined descriptions for "installed" and "activated" software states.
- Cleaned up YANG examples, removed RFC 2119 boilerplate, and aligned the security section with RFC 9907.
Discussion: Fatai Liang noted that the Broadband Forum is actively utilizing this model for its mock data models and supports moving the document to WGLC. Daniele Ceccarelli advised all authors to update their drafts' security sections in accordance with RFC 9907 to facilitate smoother reviews.

3. Network Inventory Location

Slides: draft-ietf-ivy-network-inventory-location
Presenter: Bo Wu
Document: draft-ietf-ivy-network-inventory-location
Updates:
- Revision 06 addressed feedback from Brad Peters regarding rack security classifications.
- Introduced rack security identities: standard, baseline, medium, and high.
- Aligned the model with a read-only, controller-centric design pattern.
Discussion:
- Qiufang Ma suggested renaming the parameter rack-class to something more explicitly security-oriented (e.g., security-level or security-class). Bo Wu agreed to refine this.
- Daniele Ceccarelli questioned whether device-level security classifications might be needed for implementations that do not support the location model.
- Brad Peters (via chat) noted that these categories are aligned with national government security standards (e.g., Australian standards up to 2030). Bo Wu agreed to add informative references to these standards in the draft text.

4. Entitlement Inventory

Slides: A YANG Module for Entitlement Inventory
Presenter: Diego Lopez
Document: draft-ietf-ivy-entitlement-inventory
Updates:
- Changed maximum and current values to unsigned types to avoid negative values.
- Replaced underscores with hyphens in YANG names to align with standard design patterns.
- Added explicit operational warnings to avoid mutual or circular references (e.g., Entitlement A requiring B, while B requires A).
- Updated the security considerations section in line with RFC 9907.
Discussion: Qiufang Ma and Diego Lopez clarified that the circular reference warnings apply to relational dependencies between entitlements, not the capability mapping itself. Daniele Ceccarelli inquired about the status of capability modeling; Diego Lopez indicated they plan to align this work with other efforts, such as post-quantum capabilities.

5. Passive Network Inventory

Slides: YANG Data Model for Passive Network Inventory
Presenter: Fatai Liang
Updates: Revision 05 added a consensus-based definition of passive inventory (infrastructure carrying signals without requiring power).
Open Issues:
- Modeling Approach: The authors proposed modeling passive devices as a new network element type (Option 2) rather than a completely separate inventory object (Option 1) or a pure topology object (Option 3). This supports mixed active/passive elements under one interface.
- Link-to-Cable Mapping: Determining whether an explicit mapping is needed between topology links and physical cables.
- Device Categorization: Handling complex passive components like cable joint boxes, fiber closures, MPO cables, and internal fiber routing.
Discussion:
- Mahesh Jethanandani emphasized the urgent need to establish strict boundaries for what constitutes "passive inventory" to prevent scope creep (e.g., drawing lines around racks, waveguides, and power elements).
- Daniele Ceccarelli suggested defining a concrete list of supported passive elements on the mailing list instead of trying to find a perfect, all-encompassing logical definition.
- Roberto Manzotti noted that modeling passive components (e.g., splitters, combiners, AWGs) inside active network elements is highly valuable and was experimented with during the Hackathon.
- Mohammad Boroon and Nigel Davis discussed how domain-specific extensions, like microwave waveguides, fit into passive inventory. Daniele Ceccarelli reiterated that IVY is contribution-driven and can expand to transport-adjacent spaces if there is active interest and expertise.

6. IVY Hackathon Report

Slides: IVY hackathon
Presenter: Nigel Davis
Key Achievements:
- Successfully demonstrated multi-vendor interoperability between Cisco and Ciena controllers (Cisco HCO and Ciena Navigator) using both TAPI and IVY models.
- Proved the feasibility of translating and propagating equipment views from TAPI into the IVY structure.
Key Findings & Challenges:
- Functional Roles: Cisco pointed out that explicit functional labels (e.g., ROADM, OTN switch) are missing in both TAPI and IVY models. Nigel Davis and Italo Busi noted this was an intentional design choice to prevent issues with mutable, multi-functional hardware. They suggested using capability modeling to dynamically infer device functions instead.
- Hierarchy Mapping: Mapping TAPI's holder-embedded hierarchy to IVY's model was complex but ultimately successful.
- Operational Feedback: Operator groups (specifically TIP OPT / MUST) strongly supported the alignment and migration path from TAPI to IVY.

7. Cryptographic Asset Discovery and Inventory (CADI)

Slides: draft-liu-cadi
Presenter: Fatai Liang
Overview: CADI defines tools and guidelines to discover and inventory cryptographic assets (algorithms, certificates, keys, protocols) to facilitate Post-Quantum Cryptography (PQC) migration. Discovery can be active (self-declaration via CI/CD) or passive (configuration extraction, handshake simulation).
Discussion:
- Mahesh Jethanandani questioned whether this work falls within IVY’s L0-L3 hardware/software scope, suggesting sec-dispatch as a better starting point.
- Kent Watsen, Julien Meuric, and Diego Lopez strongly supported the work being handled within the OPS area (and potentially IVY). They noted that CADI is fundamentally an asset inventory and database task, not a protocol design task, making it operational in nature.
- Kent Watsen pointed out that the current IVY definition of "network element" excludes endpoints (clients and servers), which would likely need to be included for a comprehensive cryptographic inventory, potentially requiring a charter update.
- Roberto Manzotti supported the effort, noting the regulatory requirements to track security assets like secure boot at the component level.
- Chairs noted there is clear interest in the OPS area and will look into incorporating capability modeling and cryptographic inventory as part of a future charter discussion.

Decisions and Action Items

draft-ietf-ivy-network-inventory-software:
- Decision: Deemed stable and ready for next steps.
- Action Item: Chairs to request an early YANG Doctors review.
draft-ietf-ivy-network-inventory-location:
- Decision: Deemed stable pending minor naming updates.
- Action Item: Authors to address the naming of rack-class (potentially to security-class or security-level) and add informative references to national security standards. Chairs to request an early YANG Doctors review post-update.
draft-ietf-ivy-entitlement-inventory:
- Decision: Deemed ready for early review.
- Action Item: Chairs to request an early YANG Doctors review.
Passive Inventory Boundary:
- Action Item: Authors and working group to move the passive inventory scoping discussion to the mailing list to establish a clear boundary on what elements are in-scope.

Next Steps

Hackathon follow-up: Nigel Davis, Roberto Manzotti, and the Ciena team will meet to analyze the data collected during the Hackathon, review ITU-T G.7711/TAPI parameters, and bring proposals back to IVY for missing properties.
Charter Discussion: Chairs to initiate discussions on the mailing list regarding potential rechartering to explicitly cover capability modeling and cryptographic inventory (CADI).

Automatic IETF Minutes

IVY