Network security keeps linked systems trustworthy by controlling who can connect, what traffic is allowed, and how paths are monitored.
It is the practice and the goal together, covering policy and design as well as the devices that enforce them.
How it works
Confidentiality, integrity, and availability of networked services are the outcomes, whether the threat is eavesdropping, tampering, or a simple outage of a path.
Architecture choices such as zones, strong remote access, and monitored gateways express those outcomes in the shape of the network.
A single sensor on a mirror port is only one instrument inside this practice, and endpoint software is another, not a rival definition.
The practice fails when nobody owns the rules for which teams may open a path, even if every box has a current signature file.
Concrete examples
- A company requires a device certificate before a laptop may join the internal network from home, and the rule is written in the access standard.
- Finance systems accept connections only from an application tier, which is an architectural rule rather than a one time alert.
- An operations review asks who approved a new path from a vendor network and finds the approval missing, then closes the path.
Protective measures
- Write which flows are allowed between zones and review exceptions before they become permanent tunnels.
- Require authenticated remote access instead of exposing administration ports to the whole internet.
- Monitor paths for change, not only for known bad signatures, so a new route is visible the week it appears.
- Give one team accountability for the network rules so local exceptions do not quietly erase the design.