Independent Data Layer
Noda's Independent Data Layer is a common structure for each building's equipment, points and relationships, so engineers and AI agents work from the same unified data across a portfolio.
A Noda LaunchPad reads the building management system at each site, and a Noda Hub collects meter readings over its own wireless network. Utility bills and interval data arrive alongside, with no new controls hardware.
Noda's data modeling agents read BMS graphics and raw point lists, build out the equipment and assign each point to it. Machine learning narrows each point to its most likely matches, and the agents apply engineering judgment to the rest.
Once trend data starts flowing, the agents check how each point behaves and flag flatlined sensors and readings that disagree with related points. A Noda data modeler signs off on the finished model.
Every chiller, air handler and terminal unit carries a standard name and a defined link to the equipment upstream and downstream, aligned with Project Haystack, Brick and Google's Digital Buildings Ontology.
Equipment, points, trend history, Insights and Projects are all reachable through Noda's GraphQL API, ready for a team's own analytics and reporting tools. Noda's MCP server opens the same data to AI tools.
query { equipment(name: "AHU-3") { oapCode points { name oapCode currentValue } } }
{ "oapCode": "AHU", "points": [ { "name": "SA_TMP", "oapCode": "DAT", "currentValue": 55.2 }, { "name": "RA_TMP", "oapCode": "RAT", "currentValue": 73.8 }, … ] }
Noda's agents model buildings in hours, work that used to take weeks by hand.
Point names, units and object types arrive over the live connection. BMS graphics, submittals, drawings and nameplate photographs come from the site.
Deployment consists of a Noda LaunchPad, a secure on‑premises gateway, at each site, connected to the building management systems already in place and to controllers and IoT devices on BACnet/IP, Modbus or MQTT. Meters connect through a separate Noda gateway. Noda does not require replacing a BMS or any other building system.
Noda connects over open protocols and aligns each point with the building industry's standard data models.
Each terminal unit links to its air handler and each air handler to its plant, so a fault can be traced through connected equipment.
Noda reads controllers over BACnet/IP, Modbus and MQTT, the open protocols most building systems already use.
Pulse and Modbus smart meters connect through a Noda Hub, which sends readings over its own cellular modem.
Bills and interval data from each building's utility accounts attach to the same model as the equipment.
Outside air conditions for each site put equipment behavior and energy use in context.
Existing Haystack, Brick or Digital Buildings Ontology tagging gives Noda's model a head start.
Point names, equipment tags and controls vendors vary from one building to the next, which makes building data hard to compare or analyze. Noda's Independent Data Layer gives each building a consistent shape: every piece of equipment has a predictable name, a defined place in its system and its own points.
Trend data, meter readings and utility bills hang off that shape, giving engineers, Noda's AI agents, Command Center and the Noda Assistant the same context in every building.
No. Noda connects to the building management system already installed through a Noda LaunchPad, a secure on-premises gateway at each site. Meters connect through a separate Noda Hub.
No. Data leaves the building over an outbound-only connection, cellular by default, so standard data collection needs no access to the building's IT network. A LaunchPad can sit on the customer's own network instead where the customer prefers.
Noda works with building management systems including Tridium Niagara, Johnson Controls Metasys, Siemens Desigo and Honeywell EBI, and with controllers and IoT devices that use open protocols.
No. Cleaning up building data is part of every Noda deployment. Noda's data modeling agents sort each point into its equipment and map how the equipment connects, then test each point against live trend data. Matches the agents are less sure of get a second check against historical readings, and a Noda data modeler runs automated quality control tests before approving the model.
Hours, for the agents' part of the work. Modeling a building by hand has typically taken a data modeler weeks. A Noda data modeler reviews each model before handover, and the first quantified Projects typically follow within 30 days of live data being available in Noda.
Yes. Noda's data model aligns Project Haystack, Brick Schema and Google's Digital Buildings Ontology, which lets a point tagged under any of the three land in the same place in Noda.
Yes. Noda's GraphQL API exposes the data layer to other software, and Noda's MCP server connects AI tools to the same data.
Yes. Noda's MCP server connects AI assistants and agents that support the Model Context Protocol to the Independent Data Layer. Each connection gets scoped access to building data, including equipment, points and trend history, organized by the same standard model Noda's own AI agents work from.
Yes. Noda is ISO 27001 certified and built on AWS infrastructure hosted in the United States. Noda uses end-to-end encryption (TLS 1.3 in transit, AWS KMS at rest), zero-trust access controls, SSO and MFA, and mutual TLS for edge-to-cloud communication. The LaunchPad uses outbound-only connections, so no incoming network ports need to open.