bastion

Network Security Mechanisms

Build this skill

Decide which firewall rules, IDS or IPS, VPN and segmentation controls a small organisation's network needs, and justify each against a specific attack.

A guest phone that could see the billing laptops

During a routine check at the Nashik Road store, Neha Kulkarni joins the guest Wi-Fi on her phone, the network customers use while they wait. To her alarm, her phone can see every billing laptop in the store and even the shared drive at head office. Tulsi Health has what engineers call a flat network: every device can reach every other device, so a stranger's phone sits one step away from MedLedger and the supplier contracts. Fixing this takes network security mechanisms, the tools that decide which traffic may go where, watch for trouble, and protect data that has to cross the public internet. Firewalls, intrusion detection and prevention, VPNs and segmentation each do one of those jobs.

One term first. A device on a network offers different services through numbered doors called ports. A website answers on port 443 for HTTPS, a shared drive on port 445, a router's settings page on another. Traffic is described by where it comes from, where it is going and which port it knocks on, and every mechanism below makes decisions from those details.

Firewalls: the guard at the society gate

A firewall is a device or program that checks each connection against a list of rules and either allows it or blocks it. A rule might say "billing laptops may reach Rathi Softech's update server on port 443" or "nothing from the guest Wi-Fi may reach the billing laptops". The firewall reads the rules from the top and applies the first one that matches. A careful design ends the list with default deny: anything not explicitly allowed is blocked, so a forgotten case fails safely instead of slipping through.

Think of the guard at a housing society's gate. Residents walk in, a delivery for a named flat is let through, and a stranger who says he is "here to check the water tank" with no resident expecting him is turned away. The guard does not need to know every possible troublemaker, only the short list of who is allowed. Firewalls come in a few forms: a network firewall at the edge where the office meets the internet, a host firewall running on each laptop, and a web application firewall that sits in front of a website and examines the web requests themselves.

Delivery for B-4 ✓ rule allows it "Here to check the water tank" ✕ no rule allows it Gate guard Gate rules ✓ residents ✓ deliveries to a named flat ✓ guests a resident expects ✕ everyone else Visitor register 10:02 rider B-4 10:05 rider B-4 10:07 rider B-4 ! 9th visit this hour A wing B wing Rules at the gate = firewall Register that spots odd patterns = IDS

The gate rules decide who gets in, like a firewall. The register notices that something allowed is happening strangely often, like an IDS.

A firewall has limits too. It judges traffic by its addresses and ports, so a phishing email arriving through the allowed mail service, or an employee misusing access they legitimately have, passes straight through. That is why firewalls are one layer among several rather than the whole answer.

Sort the connections below for the Nashik Road store's firewall. Ask of each one whether a real job needs it; if none does, the default answer is deny.

IDS and IPS: the register and the guard who acts

An intrusion detection system (IDS) watches network traffic, or activity on computers, and raises an alert when something looks like an attack. It works in two ways. Signature-based detection compares traffic with patterns of known attacks, like a guard checking faces against photos of known troublemakers. Anomaly-based detection learns what normal looks like and flags anything unusual, like the register showing the same delivery rider visiting flat B-4 nine times in an hour. Signatures catch known attacks reliably but miss new ones; anomaly detection can spot the new, but also flags harmless surprises.

An intrusion prevention system (IPS) does the same watching but sits directly in the path of the traffic, so it can block what it detects immediately instead of only reporting it. That power cuts both ways. An IPS that wrongly decides a customer's order looks like an attack, a false positive, blocks a real customer. An IDS that sends hundreds of alerts a day that nobody reads protects no one.

IDS: watches and alertsGets a copy of the trafficRaises an alert for a personCannot slow or break the appUseless if nobody reads alertsIPS: watches and blocksSits in the traffic's pathDrops the attack immediatelyA false alarm blocks real usersNeeds careful tuning

Many organisations run detection first, learn what normal looks like, and switch rules to blocking once they trust them.

VPNs: a sealed armoured van on a public road

When Neha works from home, her laptop has to reach the shared drive at head office across the public internet, through her home Wi-Fi and her internet provider. A virtual private network (VPN) creates an encrypted tunnel across that public path, between her laptop and a VPN gateway at head office, which checks who she is before letting the tunnel in. It is like a sealed armoured van on a public road: the road belongs to everyone, but nobody on it can see or change what the van carries. The same idea links each store to head office, a site-to-site VPN, so store traffic never crosses the internet unprotected.

A VPN protects the journey, not the endpoints. It does nothing about malware already on the laptop or a phishing link clicked inside the tunnel. Because a VPN is a door into the internal network, its login needs two-step verification; a stolen password alone should never open it. Consumer VPN apps sold for personal browsing simply move your trust from the café's network to the VPN company.

Segmentation: separate floors with separate keys

Network segmentation divides one network into separate zones and puts firewall rules between them, so traffic can cross from one zone to another only where a rule allows it. Think of an office building where each floor needs its own key: a visitor on the ground floor can use the reception and the washroom, but the lift will not take them to the floor with the server room. If a laptop in one zone is infected, the damage stays in that zone instead of spreading everywhere.

Tulsi Health networkfirewall between everyzoneno way inGuest Wi-Fiinternet onlyupdates, appStore countersbilling laptopsmail, driveOfficestaff laptopsnamed traffic onlyServers + backupshared drive, backups

The guest phone that once saw everything now reaches only the internet, and the backup zone accepts only the traffic it needs.

Principles that hold the design together

A secure network architecture is the overall plan that combines these mechanisms, and a few principles guide it. Defence in depth means several layers, so one failure is not fatal. Least privilege means each zone, device and person can reach only what their job needs. Default deny means unknown traffic is refused. Inside does not mean trusted: a device on the office network still has to prove who it is, because attackers often start from one infected laptop already inside. And everything is logged and watched, because a defence that nobody checks slowly decays.

Put those principles to work below. Choose controls for Tulsi Health within the budget, run the five attacks, and see which control acted on each one; notice which attacks only a preventive control can stop.

Follow your curiosity.

↑ ↓ to explore · Enter to open · Escape to close