How a remote site gets its IP address quietly decides whether you can reach it, how you poll it, and how you secure it. On cellular links especially, the default is a moving, hidden address that complicates everything. This guide explains the difference between static and dynamic IP addressing and why it matters for reaching remote SCADA sites.
Static vs Dynamic IP in one line: A static IP is an address permanently assigned to a device, so it never changes. A dynamic IP is assigned automatically and temporarily - by DHCP or a cellular carrier - and can change over time. For remote SCADA, static addressing makes a site easy to reach and poll, while dynamic addressing (the cellular default) requires other techniques to locate the site.
A static IP is configured to stay fixed - either set manually on the device or reserved for it - so the device always answers at the same address. A dynamic IP is handed out automatically from a pool and leased for a time; when the lease expires or the device reconnects, it may receive a different address. On a local network, DHCP does this assignment; on a cellular link, the carrier's core network does it when the modem attaches.
The practical consequence is predictability. A device with a static address is always in the same place on the network, which makes it easy to poll, firewall, and document. A device with a dynamic address is a moving target - you cannot assume it will be at the same address tomorrow, which complicates any scheme that needs to initiate a connection to it. That single difference drives most of the design decisions around remote reachability.
Most standard cellular data plans give a device a dynamic, private-range IP behind carrier-grade NAT. Two things follow. First, the address changes, so you cannot bookmark it. Second, and more important, the device is not directly reachable from the internet at all - carrier NAT hides it, so nothing outside can initiate a connection inward. A head end cannot simply poll a well site by dialling its IP, because that IP is private, changing, and unreachable.
There are ways to get a fixed, reachable address when you truly need one. A carrier can provision a static public IP on a SIM (usually a paid option), or a private APN can give each SIM a predictable private address inside a network you control. Dynamic DNS is a partial fix that maps a changing address to a stable hostname, but it does nothing about carrier NAT hiding the device. None of these is free or automatic, which is why most remote designs avoid needing inbound reachability at all.
The cleanest answer to dynamic, NAT-hidden addressing is to stop needing to reach into the site. If the remote device initiates the connection outward - dialling the head end and holding open a channel - then its own address being dynamic and private no longer matters, because the head end never has to find or connect to it. The site can sit behind carrier NAT on a changing IP and still deliver data reliably, and it exposes no inbound ports, which is far more secure.
This outbound-first model is why modern cloud SCADA suits cellular so well. A cloud SCADA such as Merobix reads field devices where the connection is established outward from the site over protocols like MQTT, or over an outbound VPN carrying Modbus or DNP3, so a dynamic carrier IP behind NAT is a non-issue. Static public IPs become an expense you can often skip: instead of paying to make every site reachable, you let each site reach out. When inbound access genuinely is required, a static IP or private APN is the tool, but for most telemetry the outbound approach removes the problem entirely.
A static IP is fixed - the device always uses the same address. A dynamic IP is assigned automatically and temporarily by DHCP or a cellular carrier and can change over time. Static is predictable and easy to reach; dynamic is the default on most connections and can move.
Most cellular plans give a dynamic, private IP behind carrier-grade NAT, so the address changes and the device is not directly reachable from the internet - carrier NAT blocks inbound connections. To poll inward you would need a static public IP or a private APN; most designs instead have the site connect outward.
Often no. If the field device connects outward to the head end - over MQTT or an outbound VPN - its dynamic, NAT-hidden address does not matter, and you avoid the cost and exposure of a static public IP. A static IP or private APN is only needed when you specifically must reach inward to the site.
This page references the protocol specifications published by the organizations below. Editions, product capabilities, and documentation change over time - confirm current requirements and specifications directly with the source.
Last reviewed: July 27, 2026. Merobix is not affiliated with, endorsed by, or sponsored by these organizations; their names are used only to identify the standards and products discussed.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.