---
title: "How ControlCom Turns 300+ Million Monthly Facility Data Points Into Instant Answers With Tiger Data"
published: 2026-08-06T14:40:28.000-04:00
updated: 2026-08-06T15:09:44.000-04:00
excerpt: "Industrial IoT on TimescaleDB: ControlCom streams 1K-10K points/sec from thousands of assets, powering instant dashboards and AI that caught $160K billing errors."
tags: Dev Q&A, IoT
authors: Andrew Stebbins
---

> **TimescaleDB is now Tiger Data.**

_ControlCom Connect streams telemetry from thousands of industrial and facility assets into TimescaleDB, powering instant dashboards and an AI assistant, the same platform that caught $160,000 in hidden utility billing errors at one healthcare site and flagged a tier 1 hospital's backup generators left in manual before a power failure could put patients at risk._

_A six-person team building a vendor-neutral industrial IoT platform streams telemetry from thousands of facility assets, from PLCs and meters to switchgear and generators, into TimescaleDB. The result: dashboards that re-bucket 300+ million data points a month on the fly, and an AI assistant that answers questions against live and historical readings in real time. That speed shows up across a growing list of catches, among them a healthcare deployment where automated monitoring caught $160,000 in utility billing errors that had gone unnoticed for years, and a tier 1 hospital where the same monitoring flagged backup generators left in manual mode before a routine service call could turn into a life-safety incident._

_This is an installment of our Community Member Spotlight series, in which we invite our community members to share their work, spotlight their success, and inspire others with new ways to use_ [_TimescaleDB_](https://www.tigerdata.com/timescaledb)_._

_Today we hear from Catalin Negru, founder and CEO of_ [_ControlCom Technologies_](https://www.controlcomtech.com/)_, and the team behind_ [_ControlCom Connect_](https://www.controlcomtech.com/platform/overview)_, an industrial IoT platform that unifies live equipment data across a facility portfolio. Catalin shares how a six-person team of software and electrical engineers evaluated three time-series options before settling on TimescaleDB, and why staying inside Postgres is what lets the platform's AI assistant and living asset graph stay fast as customer data volumes climb._

## About ControlCom Connect

The people who keep critical facilities running (healthcare, water and wastewater, prime power, data centers, manufacturing, oil and gas, cold chain) don't get a unified view of the equipment they're responsible for. A SCADA system, a handful of PLCs, a fleet of meters, a generator, and a compressor each speak their own protocol and report into their own app. When a power event hits, the operator's job becomes walking the floor panel by panel, guessing what dropped, what restarted on its own, and what is still down.

ControlCom Connect connects to whatever a facility already owns: SCADA systems, smart devices, controllers, meters, PLCs, HMIs, generators, switchgear, compressors, batteries, cranes, and tractors, plus software sources like APIs, databases, and scripts, over MQTT, Sparkplug B, OPC UA, Modbus, BACnet, Ethernet/IP, HTTPS, and webhooks. Every connected asset is wired into a living graph of the facility, so when an outage hits, the platform traces it downstream in seconds and tells the operator exactly which assets were affected, instead of sending someone to check panels by hand. Live metering and equipment status stream into drag-and-drop dashboards with multi-site comparison, smart notifications route to the right person over SMS, email, Slack, or mobile push with on-call coverage and escalation, and an AI assistant grounded in the organization's own assets and readings answers questions in plain language.

At one healthcare deployment, automated utility monitoring caught more than $160,000 in billing errors that had gone unnoticed until ControlCom's platform started watching the meter data. At a tier 1 hospital, the same kind of continuous monitoring flagged backup generators left in manual mode after a maintenance visit, before a power outage could turn a routine service call into a life-safety incident.

ControlCom Technologies is a six-person team of software and electrical engineers, founded and led by Catalin Negru.

## The Challenge

ControlCom Connect's entire platform runs on high-frequency telemetry streaming in from thousands of points across every connected asset, at a rate of 1,000 to 10,000 data points per second, and that volume only grows as customers add sites. Two parts of the product make query speed non-negotiable: the AI assistant has to answer natural-language questions against live and historical readings and return an answer while someone is waiting on it, and the living graph has to resolve relationships and pull time-series data for many points at once, on demand, the moment an outage hits. A user asking their facility a question, or watching the graph resolve an outage, can't sit and stare at a loading indicator.

> PostgreSQL was already home for the rest of ControlCom's stack, so the team needed a way to handle time-series at scale without leaving it.  _- Catalin Negru, Founder & CEO, ControlCom Technologies_

## Why Tiger Data: Architecture-First From Day One

Catalin found TimescaleDB by googling time-series databases, after the InfluxDB and AWS proofs-of-concept fell short. Because ControlCom Connect was already built on PostgreSQL, TimescaleDB (built on PostgreSQL) was an easy choice. The team kept the database they trusted, their existing tooling, and plain SQL, while gaining hypertables, compression, and fast time-series queries on top. There was no migration to a foreign database, no new query language for the engineering team to learn, and no second connection pool to manage alongside the relational data that already ran the platform.

> It is just Postgres, so we kept the database we trusted, our existing tooling, and plain SQL, while gaining hypertables, compression, and fast time-series queries on top.  _- Catalin Negru, Founder & CEO, ControlCom Technologies_

Compression mattered as much as query speed. ControlCom runs a hybrid deployment: a local TimescaleDB on every Edge Server for on-site resilience, and a central TimescaleDB on [Tiger Cloud](https://www.tigerdata.com/cloud) that holds the portfolio-wide record. Keeping cost in check on the cloud side, while ingesting from thousands of points across every customer's fleet, was a factor in the decision from day one.

## The ControlCom Connect Stack

ControlCom Connect's pipeline starts at the device: a PLC, controller, meter, or other piece of equipment. Each edge site runs an Edge Server, which also runs a local TimescaleDB that buffers and stores telemetry on site up to a configurable window, so a facility keeps monitoring itself even if it loses its connection upstream. The Edge Server publishes to an MQTT broker, which feeds into Kafka.

Kafka is where the real-time work happens. As readings stream through, ControlCom processes alarms and parses for anomalies the instant they arrive, writes the telemetry into the central TimescaleDB instance on Tiger Cloud, and pushes the update to Redis and out to every connected client in real time. The central TimescaleDB instance is the portfolio-wide source of truth: the AI assistant queries it for natural-language answers, the living asset graph queries it to trace an outage downstream, and the dashboards query it to render live and historical views across every site a customer operates.

![ControlCom Connect's data flow: facility equipment publishes through an Edge Server with a local TimescaleDB for on-site resilience, into an MQTT broker and Kafka. Kafka drives alarm and anomaly processing and writes to a central TimescaleDB on Tiger Cloud, which powers live dashboards, the AI assistant, and the living asset graph.](https://storage.ghost.io/c/6b/cb/6bcb39cf-9421-4bd1-9c9d-fa7b6755ba0e/content/images/2026/08/ControlCom-architecture-diagram.png)

__ControlCom Connect's data flow: facility equipment publishes through an Edge Server with a local TimescaleDB for on-site resilience, into an MQTT broker and Kafka. Kafka drives alarm and anomaly processing and writes to a central TimescaleDB on Tiger Cloud, which powers live dashboards, the AI assistant, and the living asset graph.__

## Results: What ControlCom has seen

### 300+ Million data points a month, re-bucketed instantly

ControlCom routinely pulls from 300+ million data points a month. Users are used to being amazed: flip a setting on a graph, widen or narrow the bucket size, push the start or end date out, and the chart updates instantly. Re-bucketing and re-ranging across that volume on the fly is the kind of workload that would crawl on a general-purpose setup. On TimescaleDB, data is served in real time, so exploring months of high-frequency data feels as snappy as scrolling a single day.

> Users are amazed when they flip a setting on a graph, increasing or decreasing the bucket size, or changing the start and end time, and the chart updates instantly.  _- Catalin Negru, Founder & CEO, ControlCom Technologies_

### An AI assistant that doesn't feel like a demo

In the industrial sector, the gap shows up the moment something goes wrong. Operators run facilities from a wall of dashboards, sometimes 80 of them, watching for any of 1,500 possible alarms. That works when conditions are calm. But when 20 alarms fire at once, the challenge stops being a lack of data and becomes finding the one signal that explains what happened, fast enough to act on it. That's the gap ControlCom's AI assistant is built to close.

ControlCom's AI assistant answers natural-language questions against both live and historical readings, which means every question triggers a real time-series query: an aggregation, a recent window, a multi-point comparison, run on the fly while the user waits for an answer. With TimescaleDB behind it, those queries come back fast enough that the conversation feels instant instead of leaving the user staring at a loading indicator, which is what makes an AI assistant grounded in live facility data actually usable rather than a demo.

### $160,000 in hidden utility billing errors

The clearest proof of what fast time-series queries make possible showed up outside the dashboard entirely. One healthcare deployment's automated utility monitoring surfaced more than $160,000 in billing errors that had gone unnoticed. Nobody was going to find that by eyeballing a meter. It took a platform that could watch every reading, all the time, and flag the anomaly the moment it appeared.

> One healthcare deployment, for example, surfaced over $160K in utility billing errors that would otherwise have gone unnoticed.  _- Catalin Negru, Founder & CEO, ControlCom Technologies_

### Beyond the balance sheet: Preventing a Tier 1 hospital blackout

Not every win shows up on a balance sheet. At one Tier 1 hospital, a maintenance crew working on the facility's backup generators left them in manual mode after finishing the job and leaving the site. ControlCom's platform caught the anomaly and escalated a notification to the facilities team that the system was out of spec. That catch depends on the same thing that makes the billing win possible: TimescaleDB lets ControlCom continuously track every asset's status across a portfolio, so a state change like a generator sitting in manual doesn't go unnoticed until someone happens to check. Had a power outage hit while the generators sat in manual, the hospital could have lost backup power entirely, with patients on operating tables and millions of dollars in potential damage on the line.

> This could have caused millions of dollars in damage if a power outage occurred. You have real people in operating rooms on the operating tables.  _- Catalin Negru, Founder & CEO, ControlCom Technologies_

## Looking ahead

ControlCom is pushing further into the real-time side of the platform: deeper stream processing and analytics, with anomaly detection that flags issues the moment a reading arrives instead of on the next dashboard refresh. The AI assistant and the living asset graph are next in line for the same treatment, since both depend on time-series queries staying fast as the underlying data grows.

That data volume is only going in one direction. As ControlCom adds more sites and more equipment for existing customers, and onboards new ones, the ingestion rate climbs across the board. The architecture decision that compounds as it scales isn't any single query. It's that the Edge Server's local TimescaleDB and the central TimescaleDB on Tiger Cloud are the same technology end to end: one set of hypertables, compression, and SQL, from the edge to the portfolio-wide record. No split time-series stack to reconcile as fleets grow, and no separate system to keep affordable as history piles up.