DEV Community: PlantWired

InfluxDB vs TDengine: MIT single-node or AGPL with clustering and tag-count prices

PlantWired — Tue, 03 Nov 2026 19:11:46 +0000

InfluxDB and TDengine are both open-source time-series databases used for plant and sensor data, but the open-source deal and the pricing unit differ. InfluxDB 3 Core is permissively licensed (MIT or Apache-2.0) and single-node, with HA and clustering in the CPU-priced Enterprise edition. TDengine's open-source core is AGPL-3.0 and includes clustering; its commercial All-in-One adds industrial connectors and a data platform, priced by tag count and free up to 5,000 tags.

Every row below is taken from the vendors' own public pages.

Open-source license

InfluxDB 3 Core: MIT or Apache-2.0, at the user's choice. Core is a single node with no high availability; HA, read replicas, historical query capability and single-series indexing are Enterprise additions.
TDengine OSS: AGPL-3.0 (copyleft). The open-source edition is a full-featured database including clustering; Enterprise and Cloud are built on the same code base.

Commercial pricing (public pages, checked Oct/Nov 2026)

InfluxDB: Enterprise is custom-priced by CPU configuration (30-day trial). Cloud Dedicated is also custom-priced; Cloud Serverless is usage-based. There is also Amazon Timestream for InfluxDB.
TDengine All-in-One is priced by tag count: free to 5,000 tags (including small production use); 10,000 tags cost US$15,500/year or US$38,750 perpetual; high availability adds +30%. Cloud starts from US$970/month for 5,000 tags.

Plant-floor collection

InfluxDB collects via InfluxData's Telegraf agent (400+ plugins), including OPC UA and MQTT input plugins.
TDengine commercial plans include zero-code connectors for MQTT, Kafka, OPC and PI System, plus edge-to-cloud replication.

Querying and built-in processing

InfluxDB is built for fast recent-data queries (vendor docs cite under 10 ms for last-value queries) and ships an SQL editor in InfluxDB 3 Explorer, plus an embedded Python VM for plugins and triggers.
TDengine extends SQL for time series (interpolation, downsampling, time-weighted averages) and includes caching, stream processing, Kafka-like data subscription and the TDgpt AI agent.

Managed cloud and platforms

InfluxDB: Cloud Serverless, Cloud Dedicated, or Amazon Timestream for InfluxDB. Core runs on Linux, macOS and Docker.
TDengine Cloud runs on AWS, Azure or Google Cloud. Self-hosted packages cover Linux and Windows, plus Docker and Kubernetes.

Quick fit check

Pick InfluxDB if you want a permissive MIT/Apache license and a large Telegraf ecosystem, and you can live with a single node or pay for Enterprise HA — or you prefer usage-based cloud pricing or AWS-managed Timestream for InfluxDB.

Pick TDengine if you want clustering in the open-source edition and are comfortable with AGPL-3.0 obligations, or you want historian-style tag-count pricing with OPC, MQTT and PI System connectors in the box (free under 5,000 tags).

The full side-by-side table with every source link is on PlantWired: InfluxDB vs TDengine. Related: Canary Historian vs TDengine, Factry Historian vs InfluxDB, the TDengine profile, the InfluxDB profile, and Choosing a process historian.

Prices and licensing terms change, and our data comes from public vendor pages, so it surely has gaps. If you run either in production and something's wrong, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app/newsletter.

Canary Historian vs TDengine: two historians that publish prices by tag count

PlantWired — Mon, 02 Nov 2026 19:10:09 +0000

Canary and TDengine are two of the few historians that publish their prices, and both price by tag count. Canary is a classic process historian with its own lossless storage, store-and-forward collectors and unlimited Axiom and Excel clients. TDengine sells a time-series database plus an industrial data platform (TDengine All-in-One) that is free up to 5,000 tags, with an AGPL-3.0 open-source core.

What each one is

Canary: a process historian with its own storage, Axiom trending, Excel clients and store-and-forward collectors.
TDengine: a time-series database for Industrial IoT (TDengine TSDB) plus an industrial data management platform (TDengine IDMP), sold together as TDengine All-in-One.

Pricing (public pages, checked Nov 2026)

Canary prices by tag count, perpetual or annual subscription, with Enterprise licensing for unlimited tags. Example: 3,500 tags cost US$9,970 for the license plus US$1,795 CustomerCare perpetual (US$11,765), or US$3,980 per year as a subscription. There is no free tier on the pricing page.
TDengine prices by tag count, annual or perpetual, with high availability at +30% and over 500,000 tags on quote. It's free up to 5,000 tags, including small production use (optional support is US$3,375/year). Example: 10,000 tags cost US$15,500/year or US$38,750 perpetual; 100,000 tags cost US$46,000/year or US$115,000 perpetual.

Open source

Canary is commercial software under its own Software License Agreement.
TDengine's core database, clustering included, is open source as TDengine OSS under AGPL-3.0.

Data collection

Canary has collectors for OPC UA, OPC DA, MQTT Sparkplug B (or JSON), SQL and CSV, not separately licensed, and Store & Forward buffers data to disk during outages.
TDengine has zero-code connectors for MQTT, Kafka, OPC, PI System and more, plus edge-to-cloud replication, in all plans.

Analytics and clients

Canary includes unlimited Axiom and Excel clients; an ODBC connector is an upgrade.
TDengine includes SQL, stream processing, AI forecasting and anomaly detection, data modeling, dashboards, event management and an Excel add-in.

Managed cloud and deployment

Canary Private Cloud runs on US Azure, from US$12,320/year for 3,500 tags plus a US$1,500 one-time setup. Self-hosted, the historian server sits with collectors installed close to the data source.
TDengine Cloud runs on AWS, Azure or Google Cloud, from US$970/month (Starter, 5,000 tags). Self-hosted, it ships Linux and Windows packages, Docker or Kubernetes, for edge, VM, on-prem server or cloud.

Quick fit check

Pick Canary if you want a proven process historian with lossless storage, unlimited trending clients and support included in the price, or your plant runs OPC and Sparkplug B sources and you want unlicensed collectors with store-and-forward.

Pick TDengine if you're under 5,000 tags and want a free production start, want an open-source core you can inspect, or want SQL, stream processing, AI analytics and PI System or Kafka ingestion in one product.

The full side-by-side table with every source link is on PlantWired: Canary Historian vs TDengine. Related: InfluxDB vs TDengine, Factry Historian vs Canary Historian, the TDengine profile, and Choosing a process historian.

Prices and licensing terms change, and our data comes from public vendor pages, so it surely has gaps. If you run either in production and something's wrong, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

Ignition vs ICONICS GENESIS: one license per server or five license tiers

PlantWired — Sun, 01 Nov 2026 19:08:55 +0000

Inductive Automation's Ignition and ICONICS GENESIS (version 11, the successor to GENESIS64) are both full HMI/SCADA platforms that now market unlimited licensing. They get there differently: Ignition sells one perpetual license per server and publishes its prices, while GENESIS uses five license tiers and runs on a Windows, IIS and SQL Server stack.

Licensing model

Ignition: one license per server. Standard licenses are perpetual (one-time); Cloud Edition is usage-based.
GENESIS: five tiers, Pure, Plus, Professional, Prime and Platinum. All licenses on one network must be the same tier.

Tags and clients

Ignition includes unlimited tags, unlimited clients and unlimited concurrent Designers per server license.
GENESIS Pure is tag-limited (75, 150, 500 or 1,500 tags); Plus and above include unlimited GENESIS tags. Client entitlements are limited by hardware performance, not by the license.

Pricing (public pages, checked Oct 2026)

Ignition has a full public price list: the platform from US$1,200, the Perspective module at US$11,225, the Industrial Historian suite at US$3,500 and the Alarm Management suite at US$3,200.
GENESIS: we found no public price list on the official pages we reviewed.

Server OS and clients

Ignition servers run on Windows, macOS and Linux, with iOS and Android clients through Perspective (HTML5).
GENESIS runs on Windows 10/11 and Windows Server 2019, 2022 and 2025 (Desktop Experience) and needs IIS and SQL Server. Its HTML5 WebHMI runs in Edge, Firefox, Safari and Chrome, with 3D graphics in the browser.

MQTT and Sparkplug B

Ignition is built on open technologies including OPC UA and MQTT, with Cirrus Link MQTT modules on its price list.
GENESIS does MQTT publish/subscribe (with brokers like Mosquitto or HiveMQ), has a Sparkplug B subscriber and talks to Azure IoT Hub over AMQP; the publisher and JSON support are add-ons.

Historian and alarms

Ignition sells historian and alarm management as suites (prices above).
GENESIS includes its Data Historian and Alarm Server in every tier, with ISA 18.2 compliance features.

Free options

Ignition: a fully functional free trial and a free Maker Edition for non-commercial use.
GENESIS: no free edition listed; unlicensed systems run in a "Demo" license tier.

Quick fit check

Pick Ignition if you need Linux or macOS servers, want to avoid a Windows/IIS/SQL Server dependency, or want to price the project yourself from a public list before talking to sales.

Pick GENESIS if you run Mitsubishi Electric equipment (it counts Mitsubishi Electric FA tags separately and integrates natively), or you need BACnet, SNMP and Azure IoT Hub connectivity in the same Windows-based package.

The full side-by-side table with every source link is on PlantWired: Ignition vs ICONICS GENESIS64. Related: Ignition vs VTScada, Ignition vs Node-RED, MES vs ERP vs SCADA, and all comparisons.

Prices and licensing terms change, and our data comes from public vendor pages, so it surely has gaps. If you run either in production and something's wrong, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

Ignition vs VTScada: unlimited tags per server or an all-in-one SCADA priced by I/O

PlantWired — Sun, 01 Nov 2026 14:10:40 +0000

Inductive Automation's Ignition and Trihedral's VTScada are two SCADA platforms that both publish their prices and sell perpetual licenses. They count very differently, though: Ignition is licensed per server with unlimited tags and clients, while VTScada is licensed per installed PC by I/O tag count, with historian, alarms and reports built into one install.

What you license

Ignition: one license per server, then you add modules or suites as needed.
VTScada: one license per installed PC (Runtime or Development Runtime), priced by I/O tag count.

Tags and clients

Ignition includes unlimited tags and unlimited clients per server license.
VTScada sells tag bands from 1,000 up to 2,500,000 I/O tags, and the count includes all applications on that PC. Thin clients are licensed per concurrent user (single, 5-pack or unlimited options).

Pricing (public lists, checked Oct 2026)

Ignition has a full public price list, with the platform from US$1,200. Standard licenses are perpetual; Cloud Edition is usage-based. Modules are extra, for example Perspective (listed at US$11,225), plus Reporting, Historian and Alarm Notification suites.
VTScada publishes a US price list and bundles as a PDF; the enterprise list (100,000+ I/O) is on request. License prices are one-time, and support and upgrades are optional yearly plans.

What's in the base product

Ignition is a platform plus modules, built on SQL, Python, OPC UA and MQTT.
VTScada ships HMI, historian, ISA 18.2 alarm management and reports in one install. Alarm notifications are an option priced by tag count, and version control comes with Development Runtime.

Platform and connectivity

Ignition servers run on Windows, macOS and Linux.
VTScada runs on a Windows PC, laptop or panel PC per license, with 100+ drivers (Modbus, DNP3, Siemens S7, Rockwell CIP/ENIP and more), plus MQTT 3.1.1 with Sparkplug B and an OPC UA client since v12.

Free options

Ignition: a fully functional trial and a free, non-commercial Maker Edition.
VTScada: VTScadaLIGHT, free up to 50 I/O on up to 10 PCs, and a 90-day trial.

Quick fit check

Pick Ignition if you want unlimited tags and clients from one server and to grow by adding modules, or your servers run Linux or macOS, or you want a usage-billed cloud option.

Pick VTScada if you want historian, alarm management and reporting in one install priced by tag count, especially for water, power or oil & gas telemetry, where its RTU and radio drivers (DNP3, MDS, Motorola ACE) are built in.

The full side-by-side table with every source link is on PlantWired: Ignition vs VTScada. Related: Ignition vs ICONICS GENESIS64, Ignition vs Node-RED, MES vs ERP vs SCADA, and all comparisons.

Prices and licensing terms change, and our data comes from public vendor pages, so it surely has gaps. If you run either in production and something's wrong, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

HiveMQ vs Mosquitto: a tiny EPL broker or a clustered broker with published prices

PlantWired — Sat, 31 Oct 2026 19:08:16 +0000

HiveMQ and Eclipse Mosquitto both speak MQTT 5, 3.1.1 and 3.1, and both have a free open-source broker. They sit at opposite ends of the scale, though: Mosquitto is a small broker that runs almost anywhere, while HiveMQ is a clustered enterprise broker with published prices.

Open-source option

HiveMQ has a Community Edition: Java-based, Apache-2.0, with full MQTT 3.x and 5 support.
Mosquitto is open source as a whole, under EPL-2.0 or EDL-1.0 (SPDX: EPL-2.0 OR BSD-3-Clause).

Pricing (as of Oct 2026)

HiveMQ self-managed publishes monthly prices: a free Lab license (1 broker, 25 connections, 30 days, renewable at no charge), Launch at US$299/month (10,000 connections, 500 msg/s) and Run at US$499/month (10,000 connections, 1,000 msg/s). Enterprise is priced by connections and throughput.
HiveMQ Cloud Starter runs from US$0.34/hour plus US$0.80 per million messages (10,000 connections, 1 MB/s, 99.95% SLA).
Mosquitto is free. Cedalo sells a Pro Edition, hosted or on-prem instances and consulting, and mosquitto.org lists no prices. The open-source project has no managed tier of its own.

Clustering and footprint

HiveMQ uses dynamic, masterless clustering with no single point of failure, and cites vendor benchmarks of 200M concurrent connections and 1M messages/s. It's built for clustered server deployments, self-managed or in HiveMQ Cloud.
Mosquitto is lightweight and runs on anything from a low-power single-board computer to a full server. Bridges connect brokers together; high availability is a feature of Cedalo's Pro Edition.

Protocols and integrations

HiveMQ speaks MQTT over TCP, TLS and WebSockets. Kafka and Amazon Kinesis come with Starter; Google Pub/Sub, MongoDB, MySQL, Snowflake, data lake and bridge integrations are on Enterprise.
Mosquitto offers listeners for MQTT, MQTT over WebSockets or an HTTP API. Core MQTT features are open source; integrations, REST APIs and persistent queuing are in the Pro Edition.

Getting started

HiveMQ: the free Lab license for one self-managed broker, or the Community Edition.
Mosquitto: installers for Windows, Homebrew on macOS, and snap on Linux and Raspberry Pi, plus a public test broker at test.mosquitto.org.

Quick fit check

Pick HiveMQ if you need a clustered broker with no single point of failure and published prices you can budget for, or you want Kafka, Kinesis or database integrations and UNS governance from the broker vendor.

Pick Mosquitto if you want a small, permissively licensed broker for a pilot, an edge gateway or a single-board computer, and one broker or a few bridged brokers is enough (or you'll buy Cedalo's Pro Edition when you need HA).

The full side-by-side table with every source link is on PlantWired: HiveMQ vs Mosquitto. Related: HiveMQ vs EMQX, EMQX vs Mosquitto, MQTT Sparkplug B explained, and The Unified Namespace (UNS) stack, explained with real tools.

Our data comes from public project and vendor pages and surely has gaps. If you run either in production and something's wrong, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

HiveMQ vs EMQX: published broker prices or a free single-node BSL license

PlantWired — Fri, 30 Oct 2026 14:12:43 +0000

HiveMQ and EMQX are both clustered, enterprise-grade MQTT brokers, and both show up a lot as the backbone of a Unified Namespace. The real differences are in licensing, how pricing is published, and what sits around the MQTT core.

Open-source option

HiveMQ has a Community Edition: Java-based, Apache-2.0, with full MQTT 3.x and 5 support.
EMQX moved to a single edition under the Business Source License 1.1 in v5.9.0. Production use on a single node is free, and each version converts to Apache-2.0 after four years.

Pricing (as of Oct 2026)

HiveMQ self-managed publishes fixed monthly prices: a free Lab license (1 broker, 25 connections, 30 days, renewable at no charge), Launch at US$299/month (10,000 connections, 500 msg/s) and Run at US$499/month (10,000 connections, 1,000 msg/s). Enterprise is priced by connections and throughput.
EMQX self-managed Enterprise is an annual, custom-priced subscription, with free PoC trials on request.
Managed cloud: HiveMQ Starter runs from US$0.34/hour plus US$0.80 per million messages (10,000 connections, 1 MB/s, 99.95% SLA). EMQX Serverless starts at US$0/month for up to 1,000 connections (99.9% SLA), and Dedicated Flex starts at US$234/month (99.99% SLA).

Clustering and protocols

HiveMQ uses dynamic, masterless clustering with no single point of failure, and cites vendor benchmarks of 200M concurrent connections and 1M messages/s. It speaks MQTT 3.1, 3.1.1 and 5 over TCP, TLS and WebSockets.
EMQX has masterless clustering plus Cluster Linking, and the project cites 100M+ concurrent clients in one cluster. Beyond MQTT 5.0, 3.1.1 and 3.1 it adds MQTT over QUIC and gateways for MQTT-SN, CoAP and LwM2M.

Integrations and data governance

HiveMQ includes Kafka and Amazon Kinesis on Starter; Google Pub/Sub, MongoDB, MySQL, Snowflake, data lake and bridge integrations are on Enterprise, along with Unified Namespace creation and governance and in-stream transformation.
EMQX lists 50+ integrations (for example Kafka, PostgreSQL, MongoDB and InfluxDB), a SQL rule engine and Flow Designer, plus Smart Data Hub for schema registry, validation and message transformation.

Quick fit check

Consider HiveMQ if you want published fixed monthly prices for a self-managed broker or an Apache-2.0 community edition, or you're building a governed UNS and want that tooling from the broker vendor.

Consider EMQX if you want a free managed tier to start, need CoAP, LwM2M, MQTT-SN or MQTT over QUIC next to MQTT, or a single production node under BSL plus a built-in rule engine covers you.

These are our editorial reading of the sourced facts, not a recommendation. The full table with every source link is on PlantWired: HiveMQ vs EMQX. Related: EMQX vs Mosquitto, HiveMQ vs Mosquitto, MQTT Sparkplug B explained, and The Unified Namespace (UNS) stack, explained with real tools.

Our data comes from public vendor and project pages and surely has gaps. If you run either in production and something's off, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

EMQX vs Mosquitto: a tiny EPL broker or a clustered BSL broker with a rule engine

PlantWired — Thu, 29 Oct 2026 14:11:44 +0000

Eclipse Mosquitto and EMQX both speak MQTT 5.0, 3.1.1 and 3.1, so on paper they look interchangeable. In practice they're built for very different jobs, and since EMQX 5.9.0 the licenses differ more than many people realize.

License

Mosquitto is free open source under EPL-2.0 or EDL-1.0 (SPDX: EPL-2.0 OR BSD-3-Clause).
EMQX unified its Open Source and Enterprise editions in v5.9.0 under the Business Source License 1.1. Production use on a single node is free, and each version converts to Apache-2.0 after four years.

Clustering and scale

Mosquitto is lightweight and runs on anything from a low-power single-board computer to a full server. Bridges connect brokers together; high availability is a feature of Cedalo's commercial Pro Edition.
EMQX has masterless clustering plus Cluster Linking across regions, and the project cites 100M+ concurrent MQTT clients in a single cluster.

Protocols and integrations

Mosquitto offers listeners for MQTT, MQTT over WebSockets or an HTTP API. Core MQTT features are open source; integrations, REST APIs and persistent queuing are in the Pro Edition.
EMQX adds MQTT over QUIC and gateways for MQTT-SN, CoAP and LwM2M, plus a SQL rule engine, Flow Designer and 50+ integrations (for example Kafka, PostgreSQL, MongoDB and InfluxDB).

Paid options and getting started

Mosquitto has installers for Windows, Homebrew on macOS, and snap on Linux and Raspberry Pi, plus a public test broker at test.mosquitto.org. Cedalo sells the Pro Edition, hosted or on-prem instances and consulting; mosquitto.org lists no prices.
EMQX Serverless starts at US$0/month for up to 1,000 connections, Dedicated Flex from US$234/month, and BYOC and Enterprise are custom-priced (as of Oct 2026).

Quick fit check

Pick Mosquitto if you want a small, permissively licensed broker for a pilot, an edge gateway or a single-board computer, and bridging a few brokers is enough (or you'll buy the Pro Edition when you need HA).

Pick EMQX if you need clustering, a rule engine and built-in integrations to Kafka, databases or cloud services in the broker itself, and a single production node under BSL is enough, or you want a managed broker with a free tier.

The full side-by-side table with every source link is on PlantWired: EMQX vs Mosquitto. Related: HiveMQ vs EMQX, HiveMQ vs Mosquitto, MQTT Sparkplug B explained, and The Unified Namespace (UNS) stack, explained with real tools.

Our data comes from public project and vendor pages and surely has gaps. If you run either in production and something's wrong, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

Factry Historian vs InfluxDB: buy a historian or build one on a time-series database

PlantWired — Wed, 28 Oct 2026 14:03:57 +0000

This isn't a classic head-to-head. Factry Historian actually stores its time series in InfluxDB (v1.x and v2.x are the supported back ends), with configuration, metadata and events in PostgreSQL. So the real question is whether you buy a finished process historian or build your own around a general-purpose time-series database.

What you get

Factry Historian is a complete process historian: collectors, asset models, event processing and a Grafana plugin, sold as one product.
InfluxDB is a database for event and time-series data. Collection, modeling and dashboards are pieces you add yourself, for example with InfluxData's Telegraf agent.

How they price

Factry charges one flat fee per production site in Standard, Advanced or Enterprise tiers, with unlimited users, tags, collectors and integrations. Historian prices aren't published.
InfluxDB 3 Core is free and open source (MIT or Apache-2.0). Enterprise is custom-priced by CPU configuration with a 30-day trial, and Cloud Serverless is usage-based.

Scale and collection

Factry has no tag limits in any tier, and Enterprise adds multi-site support. Its collectors cover OPC UA, OPC DA, Modbus TCP and MQTT (JSON and Sparkplug B), buffering on disk until the historian acknowledges the data.
InfluxDB 3 Core is a single node with no high availability; Enterprise adds HA, multi-node clustering and read replicas. Industrial collection isn't built in, but Telegraf has input plugins for OPC UA, Modbus, MQTT and Siemens S7.

One detail worth checking: InfluxDB 3 isn't listed among Factry's supported back ends, although InfluxDB 3 Core does accept the 1.x and 2.x write APIs.

Context, deployment and outputs

Factry includes asset models in Standard, and asset prototypes, event processing and manual entry forms in Advanced. It runs on practically any Linux OS, on-premises or in the cloud, and outputs to Grafana, Seeq, Power BI, a REST API, MQTT forwarding and Parquet exports. Helpdesk support is in every tier, with 24/7 phone in Enterprise.
InfluxDB offers an embedded Python VM for plugins and triggers, but no asset model is described on the pages we checked. Core runs on Linux, macOS or Docker; there are also managed Cloud Dedicated, Cloud Serverless and Amazon Timestream for InfluxDB options. Core has community support only; Enterprise comes with a support SLA.

Quick fit check

Pick Factry if you want a ready historian with collectors, asset models, events and dashboards, a flat site fee with no tag counting, and no collection and modeling layer of your own to maintain.

Pick InfluxDB if you have a data or IT team that will assemble collection, modeling and dashboards around a database, and you want a free open-source start or usage-based managed cloud.

The full side-by-side table with every source link is on PlantWired: Factry Historian vs InfluxDB. Related: Canary Historian vs InfluxDB, AVEVA PI System vs InfluxDB, InfluxDB vs TimescaleDB, and the guide Choosing a process historian.

Our data comes from public vendor pages and surely has gaps. If you run either in production and something's wrong, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

Factry Historian vs AVEVA PI System: flat site fee vs Flex subscription

PlantWired — Tue, 27 Oct 2026 14:17:17 +0000

Factry Historian pitches itself as a replacement for legacy historians. AVEVA PI System (formerly OSIsoft PI) is the historian a lot of plants already run. Both buffer data through network outages, and neither publishes historian prices. The real differences are how you pay, what it runs on, and how much of the ecosystem you need.

How they price

Factry charges one flat fee per production site, with unlimited users, tags, collectors and integrations, in Standard, Advanced and Enterprise tiers. Prices are on request, and Factry offers a guided proof of concept.
AVEVA PI is sold as a subscription through AVEVA PI Data Infrastructure, paid with AVEVA Flex credits: per-server subscriptions or an aggregate tag subscription shared across servers. Prices are on request and contracts go through an AVEVA account rep.

Scale and collection

Factry sets no limits on tags or users; capacity depends on the infrastructure you run it on. Collectors cover OPC UA, OPC DA, MQTT (JSON and Sparkplug B), Modbus TCP, Siemens S7, SQL and REST, and they store and forward to disk until the historian acknowledges the data.
AVEVA PI says a single PI Server can manage up to 20 million tags. It has over 450 PI Interfaces and over 20 PI Connectors, plus AVEVA Adapters (OPC UA, Modbus TCP, BACnet, DNP3, MQTT including Sparkplug B). The PI Buffer Subsystem queues data to a local file and flushes it in order when the connection comes back.

Storage and deployment

Factry stores time series in InfluxDB (v1.x or v2.x) and configuration, metadata and events in PostgreSQL. It runs on practically any Linux OS, on-prem, in a corporate data center or in the cloud.
AVEVA PI uses PI Data Archive for time series, and PI Asset Framework (assets, analytics, event frames) needs Microsoft SQL Server. PI Data Archive 2024 runs in production on Windows Server 2022, 2019 or 2016, with edge, on-prem and CONNECT cloud components around it.

Visualization and outputs

Factry has an official Grafana datasource plugin, a Seeq connector, an Excel add-in, MQTT (Sparkplug B) output, Parquet export and a REST API.
AVEVA PI offers optional add-ons such as PI Vision (web displays), PI DataLink (Excel) and PI Integrator for Business Analytics, plus open APIs and an SDK.

Quick fit check

Pick Factry if you want a predictable per-site fee with no tag counting, your team is comfortable with Linux, InfluxDB, PostgreSQL and Grafana, and you want direct database access.

Pick AVEVA PI if you need a long-established enterprise historian with hundreds of interfaces, asset framework, event frames and notifications, and your IT standard is Windows Server and SQL Server.

The full side-by-side table with every source link is on PlantWired: Factry Historian vs AVEVA PI System. Related: Factry vs Canary Historian, Factry vs InfluxDB, Canary vs AVEVA PI, and the guide Choosing a process historian.

Our data comes from public vendor pages and surely has gaps. If you run either in production and something's wrong, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

Factry Historian vs Canary Historian: flat site fee vs tag-count pricing

PlantWired — Sun, 25 Oct 2026 14:08:51 +0000

Factry Historian and Canary Historian get shortlisted for the same job: a vendor-neutral process historian that collects plant data reliably and serves it back fast. Both buffer data on disk during network outages and both work with Ignition. Where they really differ is how you pay and where your data lives.

How they price

Factry charges one flat fee per production site, with unlimited users, tags, collectors and integrations, in Standard, Advanced and Enterprise tiers. Historian prices aren't published; Factry offers a guided proof of concept.
Canary licenses by tag count, perpetual or annual subscription, and publishes a price calculator. At 3,500 tags it shows US$9,970 perpetual plus US$1,795 CustomerCare, or US$3,980/year as a subscription (rechecked Nov 2, 2026). Enterprise licensing covers unlimited tags.

If you have many tags per site, the flat fee is easier to forecast. If you want to price it yourself before talking to sales, Canary's calculator wins.

Collection and buffering

Factry has collectors for OPC UA, OPC DA, MQTT (JSON and Sparkplug B), Modbus TCP, Siemens S7, SQL and REST, all with store and forward.
Canary has collectors for OPC UA, OPC DA, MQTT Sparkplug B 2.2/3.0 (or JSON), SQL and CSV, plus gRPC and a Web API for custom collectors. Collectors aren't licensed separately, and its Store & Forward service backfills automatically and can push to several historians.

Storage and deployment

Factry stores time series in InfluxDB (v1.x or v2.x) with PostgreSQL for configuration, metadata and events. It runs on practically any Linux OS, on-prem, in a corporate data center (hub and spoke) or in the cloud, and has an official open-source Grafana datasource.
Canary uses its own daily HDB files, losslessly compressed, and never down-samples or interpolates. It's self-hosted, or Canary Private Cloud on US Azure from US$12,320/year. The Axiom trending client is included and an ODBC connector is an upgrade for BI tools.

Quick fit check

Pick Factry if your team already runs InfluxDB, PostgreSQL, Linux and Grafana and you want a predictable per-site fee with no tag counting.

Pick Canary if you want a self-managed historian with its own lossless storage, unlicensed collectors and no separate database to administer.

The full side-by-side table with every source link is on PlantWired: Factry Historian vs Canary Historian. Related: Canary vs AVEVA PI, Canary vs InfluxDB, Canary vs Proficy Historian, and the guide Choosing a process historian.

Our data comes from public vendor pages and surely has gaps. If you run either in production and something's wrong, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

Choosing a SCADA/HMI platform: 10 side-by-side comparisons

PlantWired — Fri, 23 Oct 2026 14:06:42 +0000

Picking a SCADA/HMI platform usually starts with the PLCs you already own and ends with a licensing spreadsheet. Here are the comparisons we maintain on PlantWired for that choice. Each page lists vendor, category, pricing model and documented integrations, and the newer ones add sourced head-to-head tables.

Licensing is where they differ most

Ignition vs VTScada: both publish prices and sell perpetual licenses, but count differently. Ignition is licensed per server with unlimited tags and clients; VTScada is licensed per installed PC by I/O tag count, with historian, alarm management and reporting in one install. VTScadaLIGHT is free up to 50 I/O.
Ignition vs ICONICS GENESIS64: ICONICS GENESIS (version 11, successor to GENESIS64) uses five license tiers from tag-limited Pure to unlimited-tag Plus and above, and runs on Windows with IIS and SQL Server. Ignition servers run on Windows, macOS and Linux.
Ignition vs Node-RED: different jobs that often coexist. Node-RED is a free, Apache-2.0 flow tool for wiring devices and APIs at the edge; Ignition is a commercial SCADA/HMI platform with operator screens, alarming and history.

Staying inside your PLC vendor's ecosystem

Ignition vs SIMATIC WinCC: WinCC is Siemens' SCADA/HMI family, tightly integrated with TIA Portal and SIMATIC controllers.
Ignition vs FactoryTalk View: FactoryTalk View is Rockwell's HMI software (Machine and Site editions), designed around Logix controllers.
WinCC vs FactoryTalk View: the Siemens and Rockwell defaults side by side, useful for mixed plants.
Ignition vs AVEVA InTouch HMI: InTouch (formerly Wonderware InTouch) is a long-established HMI for operator screens and alarming.
Proficy iFIX vs WinCC: iFIX is GE Vernova's Proficy HMI/SCADA, common in process and utilities.

The next generation

WinCC vs WinCC Unified: WinCC Unified is the web-based successor generation, engineered in TIA Portal for panels and PC stations.
WinCC Unified vs FactoryTalk Optix: Siemens' and Rockwell's newer, browser-friendly HMI platforms.

Where SCADA fits

If you're still drawing the boundaries between SCADA, MES and ERP, start with MES vs ERP vs SCADA: who does what in a factory. All 65 comparisons live at plantwired.vercel.app/compare.

Our data comes from public vendor sources and surely has gaps. If you run one of these in production and something's wrong or missing, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.

MQTT brokers and UNS plumbing: 8 side-by-side comparisons

PlantWired — Thu, 22 Oct 2026 19:01:59 +0000

If you're building a Unified Namespace (UNS), the same shortlist keeps coming up: which MQTT broker, how to get Sparkplug B out of the edge, and whether you need a DataOps layer between the plant and IT. Here are the comparisons we maintain on PlantWired for exactly those choices. Each page lists the vendor, category, pricing model and the integrations we've documented, with sources.

The broker itself

HiveMQ vs EMQX: both offer open-source and enterprise tiers. HiveMQ is frequently used as the backbone of a UNS; EMQX is positioned as a high-throughput broker.
HiveMQ vs Eclipse Mosquitto: Mosquitto is the lightweight, free, open-source broker that often shows up first in a pilot.
EMQX vs Eclipse Mosquitto: two open-source options at very different scales.
Apache Kafka / Confluent vs HiveMQ: not the same job. MQTT moves device and plant data; Kafka is event streaming into enterprise data pipelines. Many stacks use both.

Sparkplug B at the edge

HiveMQ vs Cirrus Link MQTT Modules: Cirrus Link's modules add MQTT/Sparkplug B to Ignition for report-by-exception data movement, and they commonly publish to a broker like HiveMQ, so this is more "how they fit" than "pick one".
New to Sparkplug? Our explainer covers the topic namespace, Protobuf payloads and birth/death certificates: MQTT Sparkplug B explained.

Edge and DataOps

Kepware vs Litmus Edge: a classic connectivity server versus an edge platform that collects, normalizes and forwards machine data.
HighByte vs Litmus Edge: HighByte Intelligence Hub models and contextualizes OT data before it goes to cloud or IT systems.
HighByte vs Flow Software: two takes on the information layer above the plant floor.

Putting it together

The full picture, layer by layer (connectivity, edge, broker, contextualization, storage, apps), is in The Unified Namespace (UNS) stack, explained with real tools. All 65 comparisons live at plantwired.vercel.app/compare.

Our data comes from public vendor sources and surely has gaps. If you run one of these in production and something's wrong or missing, tell us in the comments and we'll fix it.

PlantWired is a free, vendor-neutral map of factory software. A short weekly brief on new integrations is at plantwired.vercel.app.