Noda scales real-time building data with InfluxDB Cloud

Overview

Making commercial buildings smarter with real-time data

Commercial buildings generate enormous amounts of operational data, but much of it has traditionally been difficult to access and use. Property managers may rely on monthly utility bills, engineers often collect readings manually, and legacy building management systems can leave data trapped in separate systems with little historical context.

Noda brings that information together. Its building management system connects utility meters, submeters, building equipment, and environmental sensors to provide real-time data for energy monitoring, utility budgeting, tenant billing, equipment performance, and building operations.

Challenge

Keeping up with a growing stream of building data

As more buildings, meters, and sensors came online, Noda had to ingest and serve an increasingly large volume of time series data. More than 800 buildings were sending minute-by-minute measurements from energy and water meters, equipment, temperature and humidity sensors, indoor air quality devices, and other connected systems. Customers expected that data to be available on demand and at the level of detail needed to understand exactly how their buildings were operating.

Noda initially used PostgreSQL for its time series data. But storing the data was only part of the problem. Customers needed to view the same measurements at different resolutions, so the engineering team built a separate Scala application to continuously aggregate incoming data into minute, 15-minute, hourly, and daily views.

As the workload grew, PostgreSQL could no longer keep up with the load, while the custom aggregation layer became increasingly difficult to maintain. The architecture was adding complexity at exactly the same time Noda needed to bring more devices and data sources onto the platform.

Noda needed a database built for time series workloads that could handle continuous ingestion, efficiently aggregate data for different use cases, and scale as the platform added more devices and data sources.

ENTER INFLUXDB

Moving from PostgreSQL to a purpose-built time series database

Noda migrated its production time series workload to InfluxDB Cloud.

Rather than move the entire platform at once, the team ran PostgreSQL and InfluxDB in parallel for several months. Every incoming data point was written to both databases so engineers could compare the results, validate the new system in production, and gradually enable InfluxDB customer by customer. Once the migration was complete, Noda retired both the PostgreSQL time series instance and the custom Scala aggregation application.

InfluxDB also changed how Noda handled data aggregation. Continuous queries could process the incoming minute-level data into the 15-minute, hourly, and daily views its applications needed, moving work that had previously required a separate application into the time series database itself.

Noda runs its production environment on InfluxDB Cloud, removing the burden of operating and scaling the database infrastructure internally.

result

Turning building data into faster decisions

With InfluxDB Cloud, Noda could scale both the amount and variety of data it collected while continuing to give customers granular, on-demand access to building performance data.

That data powers applications for real-time energy monitoring, utility budgeting, tenant billing, equipment analysis, dashboards, and predictive analytics. Customers can work with minute-level measurements when they need detailed operational visibility or use aggregated views for applications such as utility billing and portfolio analysis.

The same time series foundation also supports Noda’s predictive analytics. The platform combines historical building data from InfluxDB with factors such as weather and occupancy to model expected energy use, helping customers see when a building is performing outside its expected range. Those predictions are then stored back in InfluxDB alongside the underlying operational data.

Just as importantly, the move reduced the amount of infrastructure Noda’s developers had to build and maintain themselves. Instead of spending engineering resources on the time series database and custom aggregation infrastructure, the team could focus on customer-facing capabilities such as predictive analytics, mobile applications, integrations, and new sources of building data.

what’s next

Extending real-time data into predictive operations

Noda continues to expand the types of building data flowing through its platform, incorporating sources such as temperature, humidity, indoor air quality, occupancy, and weather alongside utility and equipment data. Those additional signals give teams more context for predictive analytics and create opportunities to identify inefficiencies earlier and optimize building performance more proactively.

As the platform grows, Noda is focused on choosing technologies that can scale with the business while allowing its engineering team to stay focused on the customer problems it is trying to solve.