HACKATHON

Summary

The Hackathon kickoff session welcomed participants, particularly newcomers to the IETF and hackathon, emphasizing the event's goal to advance IETF work through practical coding and collaboration. Key administrative details were covered, including project organization, schedule adjustments due to Daylight Saving Time, presentation guidelines, and available technical resources for project development and networking. Participants were encouraged to engage, utilize provided resources, and prepare for result presentations and the Hack Demo Happy Hour.

Key Discussion Points

Hackathon Goals:
- Advance IETF work by focusing on code and collaboration.
- Attract new software developers to standards work.
- Emphasize "rough consensus and running code."
IETF Note Well and Licensing:
- Participants reminded to review the IETF Note Well.
- Code contributions to open-source projects are governed by those projects' licenses and rules, distinct from IETF contributions.
Project Organization and Team Formation:
- Approximately 40 projects listed on the wiki, with more expected.
- Physical board near the entrance for team locations.
- "Lost and found" resource for individuals seeking projects or projects seeking specific skills.
Schedule Overview:
- Kickoff followed by a "team schedule" page for remote participant coordination (e.g., Zoom/WebEx/Gather sessions).
- Lunch at 1:00 PM (later than usual, aligning with the week's schedule).
- Dinner provided on Saturday evening.
- Hackathon concludes Saturday evening at 9:00 PM.
- Daylight Saving Time (DST) Change: DST ends Saturday night/Sunday morning. Clocks shift back one hour, effectively delaying Sunday's schedule by an hour (e.g., doors may be closed if arriving at the previous "scheduled" time).
- Sunday Schedule: Lunch, followed by results presentations.
Results Presentations (Sunday):
- Recommended to stop coding around 1:30 PM to prepare.
- Optional, but highly encouraged; presentations should be 2-3 minutes.
- Focus on objectives, outcomes, lessons learned, and planned contributions to working groups.
- Presentations uploaded via the DataTracker to the specific "results presentation" session.
- Important: Upload presentations before the session starts; use the exact same filename for revisions to avoid duplicates.
- Presentations will be run in upload order via MeetEcho.
- Hackathon ends Sunday at 4:00 PM.
Collaboration Tools and Network:
- Gather: Virtual collaboration space with hackathon-specific rooms and tables available. Attendees should report any user space limitations.
- Hackathon Network: Special network setup available for projects with specific network characteristics. Remote access to the IETF network is possible.
- Personal APs: Guidelines provided to prevent interference with IETF Wi-Fi; contact organizers for assistance.
IETF GitHub Organization:
- Available for hosting project code, with a dedicated repo for IETF 124.
- Templates for project presentations are available for download from this GitHub repo.
Linking Implementations to Drafts:
- Ability to add links to project code (e.g., GitHub repo) as an "additional resource" on individual drafts in the DataTracker.
- For Working Group drafts, chairs have permissions to add these links.
Post-Hackathon Engagement:
- Hack Demo Happy Hour: Monday, 7:00 PM - 8:00 PM, for extended demos and deeper discussions. Sign-up requested by 1:00 PM Monday for logistical planning.
- Code Lounge: Regular shared space for continued work post-hackathon.
Sponsorship: Acknowledgment and thanks to Ericsson (Gold Sponsor), CNNIC, and ICANN (Bronze Sponsors).

Decisions and Action Items

Decision: Daylight Saving Time ends tonight (Saturday into Sunday), shifting all Sunday activities one hour later than their nominal schedule times (e.g., a 9 AM start will feel like 10 AM to attendees who don't adjust).
Action Item (Attendees): Newcomers are encouraged to seek assistance finding teams or projects.
Action Item (Attendees): Review the IETF Note Well and understand the licensing implications for code contributed to open-source projects.
Action Item (Attendees): Utilize the wiki, physical project board, and "lost and found" for project and team coordination.
Action Item (Remote Teams): Use the "team schedule" page to coordinate remote meeting times and platforms (e.g., Zoom, WebEx, Gather).
Action Item (Attendees): For results presentations, upload slides to the DataTracker's designated "results presentation" session before the session starts. Use the exact same filename for any revisions.
Action Item (Attendees): Report any issues with Gather user space to the organizers.
Action Item (Teams): Adhere to guidelines if setting up personal APs to avoid IETF Wi-Fi interference.
Action Item (Teams): Consider adding implementation links to individual drafts in the DataTracker (contact WG chairs for WG drafts).
Action Item (Attendees): Sign up for the Hack Demo Happy Hour by 1:00 PM on Monday if planning to demo.

Next Steps

Continued hackathon development work throughout Saturday.
Preparation and presentation of hackathon results on Sunday afternoon.
Engagement with the broader IETF community at the Hack Demo Happy Hour on Monday evening.
Ongoing code development in the Code Lounge.

Session Date/Time: 02 Nov 2025 19:00

HACKATHON

Summary

The HACKATHON session featured a series of rapid-fire presentations from various teams showcasing their projects, achievements, and future plans. Projects spanned a wide range of IETF-relevant technical areas, including network management, security, routing, congestion control, AI integration, quantum networking, and application-layer protocols. Many teams demonstrated working prototypes, interoperation successes, and identified key learnings to inform future draft development and community work.

Key Discussion Points

MCP for Network Management:
- Goal: Bridge AI with network management to achieve intent-based automation by encapsulating CLI commands as MCP tools, exposing YANG data via MCP, and enabling closed-loop automation.
- Achieved: Demonstrated a functional RAMCP server supporting both CLI and NetConf connections, allowing natural language commands to query and configure network devices (e.g., MTU changes).
- Next Steps: Enhance with full YANG integration for NetConf, MCP server discovery, and support for advanced functions like VRF, ECR, and ARDP.
Secure Software Provisioning (TIP and Verigium):
- Context: Explored practical implementation challenges for Trust Execution Environment Provisioning (TIP) based on RFC 9397.
- Achieved: Integrated Verigium (remote attestation verifier) and delivered an application embedded within a TIP message, implementing a generic attestation scheme.
- Learned: The TIP protocol requires a generic attestation scheme for key confirmation, and error handling for remote attestation failures needs clearer definition in the TIP specification.
- Next Steps: Deploy attester on DRT, add secure update support, improve error handling, and open-source the demo.
CMAP:
- Goal: Implement algorithms and models for CMAP, focusing on micro-segment ID (Microsid) support and "what if" scenarios for link failures.
- Achieved: Implemented Microsid support, a mechanism for link enable/disable, and an algorithm to compute packet paths. Initial path modeling using RFC 8345 was demonstrated in an operator's lab and virtual lab.
- Learned: RFC 8345 is suitable for basic path examples but insufficient for complex policies with multiple candidate paths or segment lists.
- Next Steps: Evaluate the model defined in the 'table and simapion' draft, identify gaps, implement candidate paths, and extend path computation with external relationship mechanisms.
YANG Provenance Signatures:
- Goal: Enhance YANG-based data (telemetry, configuration, control) with signatures to guarantee origin and integrity, as per the "Jan Provenance" draft.
- Achieved: Enhanced the library for YANG signatures, improved model augmentation procedures for better integration with YANG practices (schema-driven). Demonstrated integration with the Kafka schema registry for telemetry workflows.
- Next Steps: Identify further application areas, progress the draft, and maintain an updated reference implementation.
RADIUS Updates:
- Problem: Addressed critical signaling deficiencies in the 1993-era RADIUS protocol, including discarding well-formed packets and limited ID space.
- Achieved: Implemented a new packet type to indicate "received, but unable to process, send elsewhere" in FreeRADIUS and radsec proxy. Identified corner cases and issues in the current draft specification and implementations. Initiated discussions on adopting QUIC for improved security and transport.
- Next Steps: Resolve identified spec issues, fix implementation bugs, and continue discussions on QUIC adoption at the RadSec meeting.
SRv6 SFC with SROWF Functions:
- Objective: Develop a comprehensive management system for SRv6 Service Function Chaining (SFC) using a controller for SR policy computation and a manager for network function deployment and endpoint configuration.
- Achieved: Successfully implemented SRv6 SFC path computation, including loose source routing in Polaris PC and updating the BGPRLS subsegment API for GoBGP version 4. Demonstrated deployment of virtualized network functions to OpenStack Hypervisor and SRv6 configuration.
- Next Steps: Continued development and refinement of the controller and manager components.
Scone Interop:
- Achieved: Successfully flowed Scone packets over the internet for the first time, demonstrating interoperability between test networks in Germany and local deployments. A Facebook app client on an Android phone successfully used Scone with Facebook servers via an Ericsson-Skone enforcer for video flows. Progress was made on open-source project PRs.
- Next Steps: Prepare for the Scone meeting to push towards a Working Group Last Call for the protocol draft.
Multi-CDN:
- Problem: Mitigating risks associated with relying on a single CDN by dynamically selecting the best CDN for clients in real-time.
- Achieved: Developed an AI-based solution for content steering, integrated into Dash.js, utilizing CMCD/CMSD metrics, real-time throughput tracking, RTT/stall tracking, and forecasted CDN states.
- Learned: A purely local optimization approach for individual clients does not guarantee global efficiency and can lead to CDN overloading.
- Next Steps: Develop a multi-agent DRL solution to balance global Quality of Experience (QoE) trade-offs.
PQ and X.509:
- Goal: Conduct interoperability testing for Post-Quantum Cryptography (PQC) algorithms, signatures, and certificates, building on 11 previous hackathons.
- Achieved: Six different groups successfully implemented Composite Signatures (following a draft RFC with official OIDs). Interoperability testing for composite KEM (currently in Working Group Last Call) was also conducted. Ongoing draft maintenance, CMP test suite augmentation, and CMS test suite automation were performed.
- Learned: The latest composite signature draft is relatively easy to implement, with the primary challenge being integration with underlying crypto libraries.
- Next Steps: Continue monthly meetings, encourage new implementations, and further expand interop testing results.
Knowledge Graph for Network Optimization:
- Goal: Create a Proof of Concept (POC) by fusing the CMAP model with the NOIA incident model to demonstrate how existing IETF work can be integrated into a knowledge graph ontology.
- Achieved: Replaced the network data in NOIA with an RDFS version of the CMAP model, aligning ontologies to link data. Demonstrated querying CMAP data as a knowledge graph and associating trouble tickets with network components and responsible personnel.
- Next Steps: Build a workbench for further exploration and invite community collaboration.
Roca-SX in Parallel Mode:
- Goal: Implement the Roca-SX 256-bit symmetric key encryption algorithm in parallel mode using AVX-512 for enhanced performance.
- Achieved: Developed and evaluated Roca-SX2 (2-parallel) and Roca-SX4 (4-parallel) versions. Roca-SX4 achieved over 955 Gbps on an AMD CPU, outperforming high-AE in 4-parallel mode.
- Next Steps: Update the Roca-SX draft with details of the parallel mode implementation and upload source code to GitHub.
KIRA (Scalable Zero Attach Routing Protocol):
- Goal: Debug forwarding table inconsistencies and enhance multi-path support for KIRA, a Rust-implemented ID-based IPv6 routing protocol.
- Achieved: Made progress on identifying forwarding inconsistency issues, attributing connectivity problems in large topologies to timing issues in test environments rather than implementation flaws. A path collection structure was also finalized.
- Next Steps: Focus on achieving reliable full connectivity and developing dynamic features.
AI-It-N (AI Network):
- Goal: Build a generic AI network to facilitate various AI services (training, agentic operations) by connecting data, model, agent, and compute providers.
- Achieved: Developed a platform with client and provider interfaces. A toy example demonstrating distributed training with dataset selection using an Enwatt-based solution was built and shown to be functional.
- Next Steps: Develop communication protocols, ensure interoperability with other agentic protocols (ToA, MCP, ANP), address security and privacy, and add new services.
PacketScope (eBPF for Protocol Stack Defense):
- Problem: Network stacks are black boxes, hindering diagnosis and defense against modern attacks.
- Achieved: Deployed PacketScope, an eBPF-based system, to monitor live traffic and detect/block SYN flood attacks. The "guardian" module integrated eBPF for zero-copy/latency data collection with AI Modules (AIMs) for reasoning and auto-generating eBPF filtering rules.
- Learned: eBPF represents a shift to intelligent, real-time network defense, significantly reducing operational burden.
- Feedback for WG: Request for more flexible eBPF hooks for modern protocols like QUIC and HTTP/3.
- Next Steps: Extend tracing to application-layer protocols, build cross-host correlation, and enhance compatibility using BTF.
L4S Interop:
- Context: Continued interoperability testing for Low Latency, Low Loss, Scalable Throughput (L4S) congestion control and Accurate ECN (RFC 8888).
- Achieved: Chrome now supports L4S on both the receive (RFC 8888 feedback) and send (experimental L4S congestion controller) sides. Testing with Netflix and NVIDIA cloud gaming streams identified several bugs. Netflix further tested its rate control, and Nokia experimented with new growth algorithms.
- Next Steps: Continued testing in a shared workspace and further discussions in the ICCRG.
Quantum (Network Specifications):
- Goal: Develop a set of network specifications for quantum networks, initially targeting data center and multi-computer interconnects, with future interest in wider networks.
- Achieved: Focused on quantum network architecture, Q-node specifications, and EPPPS node types. Engaged in discussions regarding related technologies like SRv6, PQC, message brokers, and network design.
- Next Steps: Active participation in the QRG meeting, seeking community contributions, and promoting available quantum internet simulation and physical layer control software packages.
IPv6 Web Resource Checker:
- Problem: Difficulty in confirming IPv6-only readiness for web applications due to dynamic content and numerous resource dependencies.
- Achieved: Released a Selenium-based test service on GitHub with a publicly available demo. This tool crawls web pages, analyzes performance logs, and verifies if all resources are loaded over IPv6. Demonstrated that the IETF website is IPv6-only ready, while GitHub is not.
- Next Steps: Ongoing improvements to the open-source project.
DNS over CoAP:
- Goal: Integrate DNS over CoAP functionality into the Unbound DNS resolver.
- Achieved: Successfully integrated TLS PKI into DTRS (DNS over TLS for Restricted Servers) and made CoAP resource paths configurable. Identified a bug in libcoap related to piggybacked ACKs.
- Learned: Key challenge was mapping variables between Unbound and CoAP for TLS integration.
- Next Steps: Resolve the identified bug, merge the pull request, and explore further extensions for DNS over CoAP within Unbound.
Testing Congestion Control:
- Goal: Test FQ Codel (RFC 8290) against FQPI (an active TSVWG draft) and develop an AQM evaluation suite compliant with RFC 7982.
- Achieved: Conducted FQ Codel vs. FQPI tests on a live testbed, using RUL real-time response and load tests from Flent, with aggregated RTT and throughput graphs. Developed an AQM evaluation suite on Nest, compliant with RFC 9743.
- Learned: Identified open questions regarding recommended topologies and performance parameters for evaluating multi-path transfer protocols and testing congestion control/queue management in multi-path bottleneck scenarios.
- Next Steps: Present full results at the TSVWG and CCWG sessions, seeking feedback on evaluation topologies.
COSE PQC (CBOR Object Signing and Encryption Post-Quantum Crypto Extensions):
- Goal: Integrate Post-Quantum Cryptography (PQC) algorithms, specifically ML-DSA, into the decodey COSE library.
- Achieved: Successfully integrated ML-DSA into the library using liboqs for underlying PQC implementations.
- Learned: Implementation took longer than expected, and integrating with different crypto libraries was complex. AI code generation tools were not as helpful as anticipated.
- Next Steps: Publish polished code, implement other PQC algorithms (SLH-DSA), and collaborate with the PQ and X.509 team.
Validate Network Telemetry Message Implementations:
- Goal: Validate implementations of network telemetry message drafts in two data collection systems: NetCOES and PMACCT.
- Achieved: Collected packet captures and generated output for validation. Found specific errors in PMACCT related to anydata objects, timestamp formats, and lost namespaces. Improvements were integrated into yanglint for enhanced YANG structure validation.
- Next Steps: Address identified issues in YANG structures and telemetry messages, and ensure support for YANG features.
DKIM-2:
- Goal: Develop a replacement mechanism for the existing DKIM email signing protocol.
- Achieved: Significant improvements were made to early-stage drafts, including format changes based on implementation experience.
- Learned: Different versions of the evolving specification hindered interoperability testing during the hackathon. A key takeaway was the need for more precise specification definitions to enable successful interop.
- Next Steps: Continue interop testing throughout the week and hold a dedicated session for further discussion and refinement.
Dry Run DNSSEC:
- Goal: Implement a dry run DNSSEC feature within the Unbound DNS resolver.
- Achieved: A proof of concept was initiated, enabling local configuration of trust anchors to simulate dry run DNSSEC behavior.
- Next Steps: Further development is required to fully implement the feature.
DNS Miscellaneous Projects:
- Achieved: Work was done on implementing the Deleg protocol (a new delegation method), DNS transport signaling, and integrating libuv and OpenSSL's QUIC support into ISC BIND. Projects also included implementing poison-licious drafts for sharing DNS resolver caches in ISC BIND, extensive discussions on various internet drafts, and preliminary work on Post-Quantum DNSSEC testbeds and ML-DSA signature algorithm support.
Secure Hybrid Network Monitoring:
- Goal: Develop a system to monitor characteristics of hybrid networks (mixed cloud environments) with a focus on security aspects, such as geolocation, operator, link quality, and tunneling protocol properties.
- Achieved: Implemented core functionality for acquiring telemetry information from NOAA. Use cases were examined, confirming the system's usefulness, including a demonstration of real-time anomaly detection for unexpected BGP route switches.
- Next Steps: Implement more PCS functionalities and provide technical feedback to improve relevant internet drafts.
SCHC Management (Static Context Header Compression):
- Context: Adapting SCHC for various environments beyond LPWAN requires dynamic rule management for changing traffic.
- Solution: Integration of CoreConf (network management for constrained environments using YANG, SID, CoAP, CBOR) to manage SCHC rules.
- Achieved: A Proof of Concept (POC) demonstrated CoreConf integration in an SCHC endpoint, handling fetch, iPatch, and RPC operations to manipulate remote context. This enables efficient, traffic-adaptive compression and optimized rule management using a hierarchical tree structure.
- Next Steps: Further refinement of the CoreConf integration and rule management optimizations.
5G I2NSF (Integrated Security System for 5G Network with I2NSF):
- Goal: Create an edge-based I2NSF framework for 5G networks to overcome delays associated with cloud-based policy enforcement during UE handover.
- Achieved: Learned the design and implementation of I2NSF on a VIOG network with 3-5GC. Demonstrated the concept of placing Network Security Functions (NSF) inside the User Plane Function (UPF) for optimal, low-latency policy migration during UE handover.
- Next Steps: Convert YANG/SML data models to YAML, and design/implement the PILC protocol.
ILMP (Identifier Locator Network Protocol):
- Context: Continuing development of ILMP, an IPv6 extension for ID-locator based addressing, designed to avoid tunnels, proxies, and NATs while supporting existing applications.
- Achieved: Continued testing of multi-homing with TCP and UDP, demonstrating smooth performance with TCP cubic. EngineX and Firefox streaming DASH video over ILMP also showed successful operation. Wireshark support for ILNP was successfully integrated into the main distribution.
- Next Steps: Continue detailed testing and performance measurements, and update the FreeBSD code base.
YANG Data Model for Multi-Statements of SCIT (Software Component Identification Tag):
- Problem: The existing SCIT draft struggles to represent hierarchical information flow within complex software and computing systems.
- Achieved: Identified gaps in the SCIT draft, conducted "slide hacking" to clarify scope and requirements, and developed example use cases (with an OCP case study).
- Learned: The SCIT architecture and API are robust and ready for RFC publication. Future work should focus on strengthening the use case section for any new drafts.
- Next Steps: Strengthen the use case section to clearly articulate benefits before publishing a new draft.
VCon (Conversational Data Exchange):
- Goal: Implement and test specific aspects of the VCon specification, focusing on "transfer" and "SIP ID" objects.
- Achieved: Implemented the "transfer" interface in an open-source Python library to manage transfer dialog objects and their relationships (e.g., between consultative and target calls). Integrated this into NetCallCenter software for testing with live call data.
- Learned: Identified challenges in handling cases where certain call legs are not recorded, and how to indicate their existence without full metadata.
- Next Steps: Continue work on the "session ID" functionality.
WebAuthN Authentication (Web of Things Authentication):
- Goal: Experiment with and implement components of the WebAuthN architecture to better understand its application for agents accessing web resources with signed identification.
- Achieved: Developed a simple CLI-based implementation of the verifier component in Python, porting existing Cloudflare/TypeScript code. Demonstrated an agent signing HTTP requests using HTTP Business Signature and a verification server validating these signatures via a directory server.
- Learned: Gained insights into the new registry syntax and potential for applying experiences from constrained environments to directory services and rate limiting.
- Next Steps: Continue implementing and testing the technology.
Identity Crisis (Attested TLS / Attested Ad Hoc):
- Problem: The "identity crisis" (lack of guarantees about software/hardware running on a server alongside TLS authentication) is being actively exploited.
- Core Idea: Combine PKI with attestation evidence, so that if one mechanism fails, the other can still provide protection.
- Achieved: Discussed open questions from formal analysis and explored how this work could be utilized for Attested Ad Hoc (an adopted item in the LAKE WG).
- Learned: Attested Ad Hoc requires separate analysis from Attested TLS. Post-handshake attestation is a suitable choice for standardizing Attested TLS to preserve PKI while adding attestation. Traditional PKI hierarchies are not suitable for Attested Ad Hoc and Attested TLS; a one-way authentication model is more appropriate.
- Next Steps: Present the one-way authentication model to the TLS Working Group and UFMRG research group.
I2ICF (Interface to Network Computing):
- Goal: Extend the I2ICF framework to build an AI agent for autonomous vehicles capable of obstacle detection and automatic stopping using camera and LiDAR data.
- Achieved: Integrated LiDAR processing into the existing I2ICF framework's Edge Server. LiDAR data is calibrated with camera input to estimate distance, triggering stop functions if obstacles are too close. This established a real-time perception-to-stop system.
- Next Steps: Improve system speed and stability, test multi-robot collision avoidance, design intent/policy translators, and create YANG data models for moving object management.

Decisions and Action Items

RADIUS Updates: Address identified corner cases and issues in the current draft specification and implementations. Decide on next implementation steps regarding QUIC.
YANG Provenance Signatures: Identify other application scenarios for the signature enhancement library.
Roca-SX: Update the draft to include details of the parallel mode implementation.
Validate Network Telemetry Message Implementations: Fix identified errors related to anydata in YANG structures, timestamp formats, and lost namespaces.
DNS over CoAP: Fix the identified bug in libcoap and push for the PR to be merged into Unbound.
COSE PQC: The presenter will publish polished code and join the PQ-X.509 team to continue work on PQC integration.
Secure Software Provisioning: Support for secure updates (beyond initial installation) needs to be addressed.
Multi-CDN: A shift to a multi-agent DRL solution is planned to address global QoE tradeoffs.
PQ and X.509: Continue holding monthly meetings to progress implementations and expand testing.
Knowledge Graph for Network Optimization: Begin building a workbench for exploration and seek community engagement.
AI-It-N: Develop appropriate communication protocols and address security/privacy considerations.
PacketScope: The working group is encouraged to consider more flexible eBPF hooks for modern protocols.
DKIM-2: Emphasize the need for more exact specification definitions to facilitate interoperability.
YANG Data Model for Multi-Statements of SCIT: Strengthen the use case section in any future drafts to clearly articulate benefits.
Attested TLS / Attested Ad Hoc: Present the one-way authentication model in the TLS Working Group and UFMRG research group.

Next Steps

Hack Demo Happy Hour: Scheduled for Tuesday. Attendees are encouraged to sign up by 13:00 tomorrow (for table planning) to showcase projects in more detail.
Code Lounge/Shared Workspace: Available throughout the week for continued collaboration and code development.
Individual Project Meetings: Several projects announced specific side meetings or working group sessions later in the week for deeper dives (e.g., RadSec, Scone, KIRA, AI-It-N, ICCRG, QRG, TSVWG, CCWG, DKIM-2, Attested TLS/Ad Hoc).
Community Engagement: Project teams are actively seeking contributions and feedback from the IETF community.
Sponsor Recognition: Special thanks extended to Ericsson (Gold Sponsor), ICANN, and CNN (Bronze Sponsors) for their financial support, and to Meetecho for enhancing presentation management functionality.