Malware Detection and Protection
Explain how signature-based and behaviour-based detection catch malware, and decide when real-time protection and quarantine should be trusted or escalated to IT.
A pop-up that saved Sameer's afternoon
Sameer Joshi in accounts downloads what looks like a supplier's March invoice. Before he can open it, a box appears in the corner of his screen: "Threat blocked. Moved to quarantine." In those two seconds, a program called anti-malware checked the new file, recognised it as malicious software and locked it away. Anti-malware is software that finds, blocks and removes malware, meaning any program written to harm a device or its owner. How it decides that a file is dangerous comes down to two very different methods, and knowing both explains why this protection works so often and why it still sometimes misses.
Three ideas in one pop-up: the file was checked the moment it arrived, it matched a known fingerprint, and it was locked away rather than deleted.
Antivirus and anti-malware
The older name for this software is antivirus, from the days when the main danger was the virus, a program that copies itself into other files. Today's threats include Trojans (harmful programs disguised as useful ones, like Sameer's fake invoice), spyware that secretly records what you type, and ransomware that locks files for a ransom. Modern products catch all of these, so the broader name anti-malware tool fits better, though most people still say antivirus and mean the same thing.
Windows, macOS and Android all ship with a built-in anti-malware tool, and companies often add a paid product that IT can manage from one screen. When the tool finds something, it usually puts the file in quarantine: a locked folder where the file cannot run but is not yet deleted. That way, if the tool made a mistake about a genuine file, IT can restore it. Quarantine is the hospital isolation ward of the computer: separate the patient first, then decide.
Signature-based detection: the "wanted" poster
The first method works like a police station's wall of "wanted" posters. Every time security researchers find a new piece of malware, they record a signature for it: a pattern that identifies that exact program. Very often the signature is a hash, a short fingerprint calculated from every byte of a file. The same file always gives the same fingerprint, and changing even one character gives a completely different one. Signature-based detection calculates the fingerprint of each new file and compares it against a list of millions of known bad fingerprints, which the tool downloads fresh several times a day.
The check is fast and very accurate for malware already on the list. A file that is not on the list simply passes.
This is fast and rarely wrong. If the fingerprint matches, the file really is that known malware, so genuine files are almost never blamed. The weakness is the flip side: it only recognises what is already on the poster wall. Brand-new malware has no poster yet, and criminals routinely make small changes to old malware so that its fingerprint changes. A thief who shaves off his moustache walks straight past a poster drawn with one.
See this for yourself below. Treat each box as the contents of a file. The two start out almost identical, differing in only one character, yet their fingerprints share almost nothing, which is exactly why a tiny change slips past a signature list.
Behaviour-based detection: the alert constable
The second method ignores what a program looks like and watches what it does. An alert constable on night patrol does not need a poster to be suspicious of someone trying every shutter handle on a street. Behaviour-based detection watches running programs for actions that malware typically performs and genuine programs rarely do: renaming hundreds of files a minute, deleting backup copies, switching off the anti-malware itself, copying itself into the list of programs that start automatically, or quietly sending data to an unknown server.
Its great strength is that it can catch malware nobody has seen before, because a brand-new ransomware still has to encrypt files to do its job. Its weakness is false alarms, or false positives: genuine programs sometimes behave strangely too. A backup tool also reads every file, and a MedLedger update also changes program files. Behaviour-based tools therefore weigh several signs together, and they need the program to start doing something before they can judge it, so a little harm may already be done.
Each method covers the other's blind spot, which is why modern anti-malware tools use both.
Sort the detections below into the method that made them. Ask yourself one question for each card: did the tool recognise the file itself, or did it notice something the program was doing?
Real-time protection: always on duty
Checking files only once a week would leave six days for malware to work. Real-time protection, also called on-access scanning, checks files at the moment something touches them: when a file is downloaded, saved, opened, run or copied from a pen drive. That is how Sameer's fake invoice was caught before he could double-click it. A scheduled scan, which looks through the whole disk at a quiet time such as 1 am, is the backup check that finds anything that arrived while the signature list was older.
Real-time protection runs all day; signature lists refresh several times; a full scan sweeps the disk once at night.
Malware knows how much real-time protection matters, so one of the first things it often tries is to switch it off. That is another reason everyday staff use standard accounts, which cannot turn protection off. Staff sometimes turn it off themselves to make a slow laptop faster or to install something that keeps getting blocked. If a genuine file is blocked, the right move is to tell IT, who can check it and restore it from quarantine, not to disable the guard for everyone.
What anti-malware cannot do
Anti-malware is one strong layer, not a complete defence. It cannot stop Farida from typing her password into a convincing fake login page, because no malicious file is involved. It is only as current as its last update, and a laptop that has been offline for a month carries a month-old poster wall. It cannot protect a device on which it has been switched off. At Tulsi Health it works alongside updates, standard accounts, careful staff and offline backups, so that whatever slips past the poster or the patrol still meets the next layer.