Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
AES is the symmetric block cipher standardised by NIST in 2001 and used almost everywhere today, from TLS sessions and encrypted disks to Wi-Fi, backups, and database columns.

What Is AES Encryption?
AES, the Advanced Encryption Standard, is a symmetric block cipher that encrypts data in 128-bit blocks using a key of 128, 192, or 256 bits. The same key encrypts and decrypts, which makes it fast enough for bulk data and unsuitable for exchanging secrets with a stranger. It is published as FIPS 197, and it is almost certainly protecting your TLS sessions, your laptop’s disk, and your Wi-Fi traffic as you read this.
Its origin is unusual for a government standard. NIST ran an open international competition between 1997 and 2000; fifteen candidates were analysed in public, and the winner was Rijndael, designed by the Belgian cryptographers Joan Daemen and Vincent Rijmen. The standard was signed in November 2001 and the design has resisted serious cryptanalysis ever since.
AES transforms each block through a series of rounds: 10 for a 128-bit key, 12 for 192 bits, and 14 for 256 bits. Every round applies four operations to a 4×4 byte state:
Since Intel’s Westmere processors in 2010, most CPUs implement these steps as dedicated AES-NI instructions. Hardware acceleration is the reason AES-256 on a modern server costs a few cycles per byte, and the reason the performance argument for weaker ciphers disappeared.
The best published attack on full AES-128, a biclique technique from 2011, reduces the work factor to roughly 2^126 operations. That is a theoretical result with no practical meaning. AES-128 is not going to be brute forced, and AES-256 is not meaningfully safer against classical computers.
Where key size does matter is longevity and policy. Grover’s algorithm would, on a large quantum computer, roughly halve the effective strength of a symmetric key, which leaves AES-256 comfortable and AES-128 merely adequate. The NSA’s CNSA suite and several national schemes therefore mandate AES-256 for long-lived secrets, and that is the sensible default for anything you expect to still matter in 2040.
AES on its own encrypts exactly one block. A mode of operation turns that into something that handles real messages, and choosing badly undoes the cipher’s strength:
| Mode | Properties | Verdict |
|---|---|---|
| ECB | Each block encrypted independently | Never use; identical plaintext yields identical ciphertext |
| CBC | Chained, needs a random IV, no integrity | Legacy only, and vulnerable to padding oracles |
| CTR | Turns AES into a stream cipher, no integrity | Fine as a building block, never alone |
| GCM | CTR plus an authentication tag | Default choice for network traffic |
| XTS | Designed for fixed-size sectors | Default choice for disk encryption |
GCM carries one sharp edge: the 96-bit nonce must never repeat under the same key. Repeat it and an attacker can recover the authentication subkey and forge messages. Random nonces are safe for roughly 2^32 messages per key, which sounds generous until a container image ships with a fixed key and restarts a million times.
In fifteen years of assessments we have never broken AES. We have repeatedly broken the systems around it. The recurring findings are predictable: keys hard-coded in source control or an APK, a fixed IV reused for every record, CBC without a MAC so that error messages leak a padding oracle, encrypted database columns with the key stored in the same database, and backups encrypted with a passphrase printed in the runbook.
Side channels matter too. Table-driven software implementations leak key material through cache timing, as Daniel Bernstein demonstrated in 2005, which is why constant-time or hardware-backed implementations are the rule for anything multi-tenant. Where keys protect regulated data, keep them in a hardware security module or a managed key service rather than in application configuration. Our web application security services test exactly these integration points.
On hardware with AES-NI the difference is a few per cent of throughput, so cost is rarely the deciding factor. AES-256 gives a larger margin against future cryptanalysis and against Grover’s algorithm, and several compliance schemes require it for sensitive data. Choose AES-256 for anything stored long term, and do not lose sleep over AES-128 in a short-lived TLS session.
The algorithm itself has not. Published attacks on full AES shave only a few bits off the theoretical work factor and remain far beyond any achievable computation. Every real-world compromise involving AES has come from implementation flaws: reused nonces, predictable keys, missing integrity protection, or side-channel leakage from non-constant-time code.
CBC provides confidentiality but no integrity, so an attacker can alter ciphertext and see how the application reacts. Padding oracle attacks exploit that behaviour to decrypt data without the key, and they have affected TLS, ASP.NET, and many bespoke APIs. Authenticated modes such as AES-GCM detect tampering before decryption, which removes the whole class of problem.
No. Encryption at rest protects data on disk and encryption in transit protects it on the wire, but the application must decrypt data to work with it, and an attacker with valid credentials or code execution sees the plaintext. Encryption has to sit alongside access control, logging, and monitoring rather than replace them.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.