Faberwork helped build an intelligent software to analyze the HVAC data.
We show how pinpointing inefficiencies can conserve energy, cut expenses, and be sustainable.
This is the second of two discussions about energy management using Snowflake. It looks at the topic from a technical and infrastructure perspective, addressing specific data management issues.
Client
Faberwork’s client is a provider of innovative Energy Management Systems (EMS) optimizing energy usage and reducing costs in smart buildings. These systems collect and analyse data from individual sensors including lights, heating and cooling systems, and individual appliances. This data is integrated with existing building assets and manages the facilities to achieve their best performance.
Challenge
Faberwork was challenged with producing an intelligent system for managing the vast amounts of Interval Data (i.e., time series data). The environmental data was generated by 18,000+ IoT devices, each sending updates at brief defined intervals. The velocity of the data alone was a challenge. The Faberwork system needed to handle the high-speed influx of data in real time without performance degradation.
The data inputs had highly variable formats and structures. Since they were fed from a wide array of IoT devices, all this data needed to be brought into a unified system. Data bottlenecks were a constant problem because of infrastructure limitations, causing delays and inaccuracies in data processing.
Data sets were scattered across different systems, preventing a cohesive view. However, the overall system needed to offer real-time insights that allowed operators to extract actionable insights. All of this needed to be done at a minimum cost. The Faberwork solution needed to be cost-effective and scale with growing data volumes.
Solution
Faberwork turned to an integration of Snowflake and Confluent Kafka to handle the challenge. Each offered unique advantages.
Snowflake Data Cloud allowed elastic scalability, real-time processing, and semi-structured data support, as well as cost-effective storage and decoupled design.
Faberwork chose to work with Confluent Kafka to ease the data transmission issues. Kafka offers message-oriented middleware, distributed architecture, reliable data transmission, and buffering and batching.
Sequence Diagram explaining the flow of the Time Series Data:
Results
Faberwork’s solution provided the client with an architecture that was highly scalable but robust enough to tolerate spikes in demand. It leveraged Snowflake and Confluent Kafka to overcome the immediate challenges of managing time series data. But it also set the stage for future advancements and success.
Faberwork’s solution delivered the following benefits:
- Streamlined Data Workflows.
- Enhanced Operational Efficiency across the organization.
- Enabled Proactive Issue Resolution with real-time identification and resolution of potential issues, minimizing downtime.
- Provided Predictive Maintenance by optimizing maintenance schedules and reducing equipment failures and associated costs.
- Cost Savings from Growth and Data-driven Innovation
“Systems integration in large-scale energy management systems is critical. Snowflake has been a great partner in our work.”
—Ramsingh Palsaniya, Senior Technical Lead, Faberwork
