EdgeLake 2.0: Building Momentum at the Edge
Author: Eric Aquaronne, EdgeLake
EdgeLake, the open source distributed data infrastructure project hosted at LF Edge, has shipped version 2.0.2606, and it arrives alongside a wave of real-world deployments, growing community participation, and deeper integration across the LF Edge ecosystem. Together, they mark a project moving from proof-of-concept technology toward production use at scale.
What’s New in v2.0.2606
The journey to 2.0 ran through v1.4.2510 and the v1.4.2603 midpoint release, each adding pieces of what’s now a considerably more modular, more connected platform. Highlights from this release include:
- 10 new features, including Modbus ingestion, a UNS (Unified Namespace) with dynamic mode support, and cluster diagnostics
- Roughly 25 fixes and improvements spanning networking, mapping, policy handling, and startup/error handling
- 100+ commits since the last annual review
Architecturally, EdgeLake continues to lean into a “modular by design” philosophy. Southbound connectivity now spans MQTT, Kafka, OPC-UA, Modbus, REST, and gRPC, bringing in data from cameras, sensors, and gateways. On the northbound side, REST now supports POST in addition to GET, joining gRPC, Kafka, and MQTT as outbound integration options. Components build independently and can be selected per node at build time, and worker threads have been generalized beyond file processing to handle broader workloads.
Looking ahead, the team is laying groundwork for zero-touch provisioning: blockchain-based node identity and deployment script selection are both in progress, aimed at making fleet-wide rollout increasingly automated.
It’s also worth noting how the EdgeLake family of work products breaks down: EdgeLake itself is the open source core (connectivity, ingestion, query, and node identity), AnyLog EDF (EdgeData Fabric) is a commercial superset adding high availability, security hardening, strengthened for production deployments, and Anylog EDM (Edge Data Manager) is a GUI for network management and topology, usable with both AnyLog and EdgeLake deployments.
Real Deployments, Real Verticals
Perhaps the most compelling sign of progress is where EdgeLake is actually running. Completed proof-of-concept and production case studies now span:
- Smart City infrastructure
- Smart Electric Cruise Ships
- Smart Buildings
- Oil & Gas, with a Houston-based industrial IoT and software development specialist
The project has also integrated with Filecoin and Akave for decentralized video processing and storage. Notably, EdgeLake positions itself as the data infrastructure layer beneath AI workloads rather than an AI/model-serving layer itself. Several additional PoCs in other industry segments are underway but not yet public, pending user agreement to share details.
At least three companies have moved EdgeLake into production and are willing to be named as adopters, a meaningful signal for a project that only entered LF Edge’s Growth Stage at the end of 2025.
Community and Ecosystem Momentum
TSC (Technical Steering Committee) meetings are now drawing 40+ registered participants, with 10+ typically joining live calls. Meeting recordings and cleaned videos are published to the wiki and YouTube after every session. The project maintains active technical ties across LF Edge and sibling projects:
- LF Networking: 5G blueprint work
- LF Energy: substation data integration
- LF AGL: automotive systems
- LF Edge / Open Horizon: full stack integration
On the orchestration front, EdgeLake is designed to slot into existing fleet management tooling rather than compete with it. IBM Open Horizon, Zededa, Barbara, and Siemens are all active conversations for fleet-wide deployment, monitoring, and lifecycle management.
The EDM management GUI has its own momentum story worth telling on its own: it started as a course project at UC Santa Cruz in Winter/Spring 2025, was hardened through the rest of that year, and picked up topology views, UNS-based search, and a query builder in the 2025-2026 cycle. Native zero-touch configuration, managing node services directly from the GUI, is next on its roadmap.
What’s Next
The 2026-27 roadmap centers on more connectors, improved high availability, deeper integration across the stack, a standardized UNS, and better network-wide insight tooling, with particular focus on the EDM tool as the primary interface users interact with. The project’s broader goal is scaling from small pilot deployments toward much larger fleets (on the order of 10,000 agents) while making the full deployment lifecycle (deploy, configure, update) as turnkey as possible through pre-built kits like VM images and “demo in a box” packaging.
EdgeLake remains focused on adoption over net-new feature sprawl for now: the priority is making an already capable platform easier to pick up, easier to run at scale, and easier to plug into the orchestration tools teams already use.
Want to see EdgeLake in action? Check out the published case studies on the LF Edge site, or join an upcoming EdgeLake TSC call. Recordings are posted to the project wiki and YouTube after every session.
