Information Security & Firewall: coverage in Gravataí - RS

The security device as a single point of failure

In an operation with a single firewall, that device concentrates two incompatible roles: it is the security control and it is the mandatory path of the connection. When it stops, everything stops — and the very device meant to protect continuity is the one threatening it.

That does not force you to duplicate everything. It does force you to have a defined answer to the question "what do we do if this device dies on a Tuesday afternoon." The options are well known: a spare unit with the configuration already applied, a contingency alternative that sustains the essentials, or accepting the downtime with full awareness of how long it is.

What is unacceptable is the current default in many environments, which is having no answer at all — and discovering the answer on that Tuesday.

An expired signature is worse than a missing feature

A detail that goes unnoticed: a firewall with expired signatures does not announce that it is inspecting with stale information. It keeps showing protection as active, and it keeps blocking what it already knew about.

The effect is false security, which is the worst category. Checking the validity of signatures and licenses, and putting those dates under tracking, is one of the best-return checks in a mature environment — and it frequently reveals that renewals stopped longer ago than anyone imagined.

Accumulated rules: why nobody cleans up

An old firewall carries a rule base that grew by addition. Every new need produced an allowance; none was removed when it stopped being necessary, because removal is frightening and delivers no visible return.

The result is a long list with allowances nobody can explain, some of them probably broader than they should be, and the shared certainty that touching them might take something down.

The review is done by measurement, not by opinion. For a period, we track which rules actually receive traffic and which never do. Unused rules are disabled — not deleted — so rollback is immediate if something surfaces. After a cycle with no complaints, they go for good. It is a slow, risk-free method — the opposite of the cleanup blitz that breaks the operation.

A single link and the cost of restoration

The other risk recorded for the city is the absence of link redundancy. The decision about a second path is usually made by looking at the monthly price, and it should be made by looking at another number: how long restoration takes when something happens to the external infrastructure.

The second path does not need to replicate the primary — it sustains the essentials: ERP, dispatch, invoicing, telephony. And it is worth more if it runs on a different technology, because then it fails for different reasons; two links over the same physical infrastructure go down together.

With two paths, SD-WAN chooses between them continuously, measuring latency, loss, and jitter, and switches automatically when one degrades. That covers partial degradation, not just outright failure — and partial degradation is the most common scenario and the hardest to diagnose without measurement.

Segmenting to contain the legacy

There is an additional reason to segment in a mature plant: it is the only control available for equipment that cannot be updated. A machine running old software, a line-operation system, an asset whose warranty lapses if anyone touches it.

You do not protect those items with updates; you protect them with isolation. A dedicated segment, communication allowed only with what the device needs to reach, no browsing, and vendor access that is restricted, temporary, and logged — instead of the door left open since installation, which the initial survey almost always finds still active.

The sectors that drive demand in Gravataí

Industry, logistics, services, and retail form the city's profile. Industry concentrates the legacy equipment and the vendor-access question. Logistics depends on the link for dispatch and tracking. Services and retail weigh heaviest on payment processing and invoicing, where downtime stops the sale.

How we run the deployment

It starts by surveying what exists: the links, the edge device and its support and signature status, the current rule base, active remote access, and real traffic. Replacing the edge without that map is the fastest way to block something the operation depends on.

Then comes the design — segmentation, per-segment policy, legacy isolation, authenticated remote access — and the definition of the answer to the failure scenario of the edge device itself.

Activation is scheduled in an agreed window, because the swap interrupts the connection for a few minutes. From there, the inherited rules enter the measurement-based review cycle described above, and link behavior goes under continuous tracking — without it, nobody notices when the secondary path stops working, and it will only be called upon the day the primary goes down.