IETF VELOCE process decouples YANG source code from documentation standards

The shift to a code-first development model addresses a critical bottleneck in streaming infrastructure standards where updating a single line of YANG code previously required re-publishing entire text-heavy RFCs. By adopting GitHub-centric workflows, the IETF is aligning its standards process with modern software engineering practices, potentially reducing the multi-year lag in network protocol updates. For the streaming ecosystem, this means faster iteration on the data models that govern content delivery networks and automated traffic engineering. Strategists should monitor the two-year experimental window to see if this process successfully transitions to a Best Current Practice, which would signal a permanent change in how streaming-related network standards are maintained.

The push to modernize YANG-based network management extends well beyond the IETF's VELOCE experiment. In early 2026, Google published new documentation aimed at optimizing websites for generative AI features in Search, signaling a broader industry shift toward machine-readable content structures that parallel the code-first philosophy VELOCE brings to network data models. The same principle of separating structured data from prose for machine consumption is driving investment across infrastructure vendors. Arrcus, a key participant in the IETF OPSAWG working group, has built its entire routing and switching platform around YANG-native configuration, making faster module iteration directly relevant to its product roadmap and to operators deploying programmable networks at scale.

On the business and governance side, the move toward automated standards development reflects pressure from cloud and streaming operators who need network models that keep pace with service deployment cycles. Akamai introduced AI Brand Presence in August 2026 to help organizations optimize content for AI search and agentic traffic, demonstrating how edge infrastructure providers are already adapting their platforms to machine-driven discovery. That same urgency applies to the data models governing CDN and traffic engineering: if YANG modules cannot be updated quickly, operators face a growing gap between what their infrastructure can do and what standardized management interfaces expose. The VELOCE process's two-year experimental window aligns with operator planning cycles, giving the IETF a concrete deadline to prove that CI/CD validation can replace the traditional RFC publication bottleneck.

From a technical standpoint, the entity analytics and identity resolution work happening in adjacent security stacks illustrates why structured, machine-readable definitions matter beyond network management alone. Elastic Security Labs detailed its Entity Store approach to resolving non-human identities including AI agents and service accounts, noting that autonomous agentic AI risks are among the fastest-growing concerns across modern attack surfaces. The same YANG data models that VELOCE aims to accelerate will increasingly need to represent autonomous agents and their permissions within network infrastructure. Meanwhile, shows how cloud providers are already operationalizing fine-grained, machine-readable access policies at scale, a pattern that YANG-based network automation must match to remain relevant in environments.