
Advanced metering infrastructure is usually sold on what it removes from the workday. In a Water Finance and Management piece on AMI data use, the case is put plainly: "utility staff no longer have to walk meter routes," and the article treats the benefits as ample. Billing accuracy and earlier leak detection are the other items on the list. Those gains are worth having. They also come with a property that seldom appears in the business case, which is that a network built to report to a central system can be reached through that system.
Where the Read Path Accepts Commands
A meter read moves from the register through a communications layer to a head-end system, then into billing, operations screens, and often a customer portal. Each hop has an owner, a credential set, and a vendor relationship. Many deployments were designed around the first two hops, the reads themselves, and treat the rest as administrative plumbing. The design question that matters for resilience is narrower. Which of these hops can change state, meaning disconnect a service, alter a register configuration, or push a firmware update, and which only sends data upward?
The answer varies by system and by configuration, so it cannot be assumed from a vendor brochure. It has to be read from the network diagram, the contract, and the administrative console. Where those three sources disagree, the disagreement is itself the finding.
Why a Supply Cut Is Not Always Physical
In an interview with Peter Gleick published by Circle of Blue, reporter Brett Walton frames conflict risk around a sharp observation: "daily life is more dangerous than in conflicts past." The summary of the interview also lists a cut to the water supply among the ways that danger reaches civilians. The interview concerns war zones, and a utility in any country should not read it as a forecast of its own risk. It does sharpen a planning question. A physical supply cut is usually imagined as a valve closed by hand or a pump that stops. A networked meter fleet adds a second route, in which the disruption is administrative rather than mechanical.
A Practical Inventory
Start with the list of accounts and service roles that can write to the network, not only read from it. For each, record who holds the credential, how it is approved, and whether it is shared. Next, record vendor support access: whether the vendor can log in remotely, from which locations, under what approval, and how that session is logged. Then check whether the same credentials open both the read path and any control function. If they do, that is a design fact to write down before anyone calls it a limitation.
Fallback comes next. If the head-end is unavailable, can staff still obtain reads, bill accounts, and identify leaks, or does that work depend on the same vendor and the same network? Finally, check log retention. A record that rolls over in a few days will not help an investigation that starts a week later, and the utility should know that number before it needs it.
Write the Assumption Before the Event
An inventory is only useful if it is dated and owned. Keep it with the network diagram, the vendor contract, and the incident response plan, and review it when the vendor changes, when a new service role is added, or when the read path is extended to a new area. Note where a field is unknown rather than filling it with a guess. A gap listed honestly is more useful in a tabletop exercise than a confident answer nobody has tested.
This is a planning framework, not a security assessment of any particular system. Confirm the answers with the vendor and with qualified cybersecurity staff, and treat remote access as one line in a wider risk record, not as a reason to abandon the network. The benefits described in the AMI article remain the reason most utilities deploy the technology. The point here is that the same reach that delivers the reads also has to be mapped.