Walk through the control room of any mid-sized warehouse or manufacturing plant and you will see the same thing: a row of monitors, each logged into a different vendor platform. One screen shows HVAC status. Another shows lighting control. A third tracks access events. A fourth has the chiller temperature. Before you even get to the fifth screen, the cognitive cost of assembling a complete picture of what is happening right now is already high — and you have not looked at the floor yet.
We built Meshkindle specifically because of this situation. I spent four years managing facilities operations at a Boston-area life sciences campus, and I can tell you from experience: the silo problem is not an accident. It is the product of rational decisions made by dozens of different vendors, each optimizing for their own product category. Understanding why silos form is the first step toward getting out of them.
How Protocol Fragmentation Creates Silos
Every major building system category has its own dominant communication protocol — and those protocols were not designed to talk to each other.
BACnet (Building Automation and Control Networks) is the standard language of HVAC, chillers, air handlers, and most commercial BAS equipment. It was standardized by ASHRAE in 1995 and remains the dominant protocol in commercial building automation. Modbus, which predates BACnet by nearly two decades, is deeply embedded in industrial equipment: pumps, compressors, PLCs, and power meters. Zigbee and Z-Wave dominate wireless sensors and lighting controls. Wiegand — an analog signaling standard from the 1970s — is still used by a significant portion of access control hardware in the field.
None of these protocols were designed to interoperate. They use different physical layers, different data models, different addressing schemes, and different polling cycles. A BACnet controller polling an AHU every 30 seconds operates in a fundamentally different world from a Zigbee temperature sensor pushing readings every 60 seconds over 802.15.4 mesh radio. Getting those two data streams into a single coherent feed requires either a custom integration project or a platform built specifically to bridge them.
Most facility operators have neither. What they have is seven different vendor dashboards.
The Vendor Incentive Problem
Fragmentation persists not just because of technical incompatibility but because of vendor incentives. A BAS vendor has a strong business reason to keep your HVAC data inside its platform: once you are monitoring through their dashboard, you are a candidate for their analytics add-ons, their service contracts, and their next-generation controller upgrades. Open integration means that data flows to wherever it is most useful, which may not be the vendor's proprietary cloud.
In our experience talking with facility managers across the Northeast, the integration problem is rarely acknowledged in vendor proposals. You get a detailed spec sheet for the BAS controller, a well-designed dashboard demo, and a reference list. What you do not get is a frank conversation about what happens when the building adds a new lighting system from a different vendor six months after installation.
The result is a facility where each system category is a closed loop. HVAC data lives in the BAS. Lighting data lives in the lighting management system. Occupancy data lives in the access control platform. Nobody is aggregating them. And the facility manager — who needs to understand all of them simultaneously to optimize energy consumption, detect anomalies, and respond to incidents — is manually cross-referencing four to seven dashboards multiple times per day.
What Silos Actually Cost
The cost of data silos is not abstract. There are three concrete categories where fragmentation creates quantifiable losses.
Detection lag. When temperature, humidity, and access events live in separate systems with no cross-correlation, anomaly detection depends on a human noticing something unusual in one of those feeds and then manually checking the others. The median time to detect an environmental anomaly in a facility without unified monitoring is 4.2 hours. In a server room, a cold-chain warehouse, or a pharmaceutical storage area, 4.2 hours of undetected thermal excursion is a significant liability.
Energy waste. HVAC accounts for roughly 40% of total energy consumption in commercial buildings. Optimizing HVAC requires knowing occupancy — but occupancy data lives in access control or occupancy sensors, not in the BAS. Without cross-system correlation, facility teams cannot automatically reduce HVAC output in zones where occupancy sensors show zero presence. HVAC energy waste in unmonitored buildings typically runs 25–40% above the theoretical optimum. That gap is essentially the cost of siloed data.
Maintenance inefficiency. Predictive maintenance requires correlating multiple sensor streams. A compressor failure often shows up first as a subtle change in vibration signature, then as a temperature trend deviation, before any fault code appears in the BAS. If vibration sensors report to one platform and the BAS lives in another, that early-warning signal never gets correlated. You find out the compressor is failing when it stops, not three weeks earlier when the data was already there.
Why the "Just Integrate" Advice Falls Short
Every facilities consultant will tell you the answer is integration. They are not wrong — they are just underselling how hard it is.
BACnet-to-cloud integration requires a BACnet/IP gateway, a cloud ingestion layer, a data normalization process (because BACnet object types do not map cleanly to cloud data models), and an ongoing maintenance commitment as firmware updates on either end break the integration. For a single building system, a custom integration project typically runs 3–6 months and $40,000–$120,000 in professional services. Multiply that by five or six systems per facility and it is no longer a cost-effective path for a mid-market operator.
There is also the ongoing maintenance tax. Custom integrations break. When the HVAC vendor pushes a BAS firmware update, the BACnet object tree can change. When the cloud platform updates its API, the ingestion connector needs to be updated. Facilities teams do not have the engineering bandwidth to maintain custom integration code on top of everything else they manage.
The Mesh Bridge Approach
The architectural alternative is a protocol-agnostic bridge layer that runs at the facility level rather than in the cloud. This is what we designed Meshkindle around: gateway nodes that speak BACnet/IP, Modbus TCP, Zigbee 3.0, and Z-Wave Plus natively, installed at anchor points throughout the facility, forming a self-organizing mesh network that collects telemetry from every device category and delivers a unified stream to a single dashboard.
The key difference from a custom integration project is the physical layer. A gateway node sitting on the facility floor, connected to a Modbus RS-485 bus on one port and a BACnet/IP network on another, does not need to go through the internet to correlate data from those two sources. Correlation happens at the node. The unified telemetry stream that reaches the cloud has already been normalized and cross-referenced. That architecture eliminates the cloud round-trip latency for local correlation and makes the unified feed available on the facility dashboard within seconds of a sensor update.
From a deployment perspective, the mesh approach also means no single point of failure. If one gateway node goes offline, adjacent nodes reroute telemetry automatically. The self-healing routing logic runs every 15 seconds — so a node failure triggers a reroute within about 12 seconds, and data delivery resumes without operator intervention.
What Changes When the Silos Are Gone
The operational change is less abstract than it sounds. When HVAC data, occupancy data, environmental sensor data, and access control events are all flowing into a single floor-map dashboard, a few things become possible that were not before.
First: real-time anomaly correlation. A spike in zone-level temperature combined with an access event showing a door prop can be flagged as a potential cold-chain breach in under 90 seconds, before any damage occurs. That same correlation in a siloed environment requires a human to notice the temperature spike in the BAS, then separately check the access control log, then make the connection.
Second: automated energy reduction. When occupancy sensors show a zone has been empty for 45 minutes, an automated rule can reduce HVAC output to that zone without manual intervention. That optimization is only possible when occupancy and HVAC data are in the same system with a rules engine that can act on their combination.
Third: coherent maintenance history. Every sensor reading, every alert, every fault code is time-stamped and stored in a single audit trail. When a compressor eventually fails, the maintenance team can pull the full history of vibration, temperature, and runtime metrics for that asset going back to installation — without hunting through five different platform exports and trying to align their timestamps.
The silos in your facility did not form because of bad decisions. They formed because every vendor made reasonable choices for their own product category. Getting out of them requires an architecture that treats the integration layer as a first-class problem — not an afterthought. That is where we spend most of our time.
