Encryption¶

Transform a plaintext under a key into a ciphertext from which the message cannot be feasibly recovered without the decryption key — a computational asymmetry gated by a secret, so confidentiality depends only on who holds the key rather than on the channel.

Version: v2 · 2026-09-19 · History
Domain-specific #: 330
Domain group: Applied Sciences & Engineering
Origin domain: Computer Science & Software Engineering
Subdomain: Cryptography → Computer Science & Software Engineering

Core Idea¶

Encryption transforms a plaintext under a key into a ciphertext from which the original cannot be feasibly recovered by any party lacking the decryption key. The defining commitment is computational asymmetry gated by a secret: encrypting and decrypting with the key are efficient, while inverting the ciphertext without it requires a problem believed infeasible (discrete log, factorization). Two key arrangements exist — symmetric (one shared key) and asymmetric (a published-encrypt/private-decrypt pair) — often combined in hybrid schemes like TLS.

How would you explain it like I'm…

Secret Scramble Lock

Encryption is like scrambling a message into secret gibberish that only someone with the right key can unscramble. With the key, it's quick to lock and unlock. Without the key, figuring it out would take way, way too long to be worth trying.

Scrambling With a Key

Encryption turns a readable message into scrambled text using a secret called a key. With the right key, unscrambling is fast, but without it, figuring out the message is so hard that no one can practically do it. Sometimes both people share the same key. Other times there are two linked keys: a public one anyone can use to lock messages to you, and a private one only you have to unlock them. Good encryption also notices if someone tampered with the scrambled message.

Key-Locked Message Scrambling

Encryption transforms a readable message (plaintext) into scrambled ciphertext using a key. It relies on an asymmetry: with the key, encrypting and decrypting are fast, but without it, recovering the message means solving a math problem believed to be far too hard to do in practice. In symmetric encryption, like AES, one shared key both locks and unlocks, so everyone who has it must be trusted. In public-key encryption, like RSA, a public key encrypts and a private key decrypts, so you can publish one without exposing the other. Real systems such as TLS often combine both, and modern schemes also check that the ciphertext wasn't tampered with. Often the weakest point is not the math but how keys are created, shared and stored.

Encryption maps plaintext under a key to ciphertext, built on a computational asymmetry. Encryption and keyed decryption are efficient, while inversion without the key reduces to a problem assumed hard at the key sizes in use, such as factoring or discrete logarithms for classical public-key schemes. Symmetric schemes like AES-GCM and ChaCha20-Poly1305 use one shared key, so security depends on distributing it safely and on every holder staying uncompromised. Asymmetric schemes like RSA-OAEP and ECIES use a linked key pair: the public key encrypts and can be published, while the private key decrypts and stays with its owner. This solves key distribution at the cost of speed. Hybrid designs, as in TLS and Signal's Double Ratchet, use public-key methods to establish an ephemeral symmetric session key that encrypts the bulk data. Modern schemes are AEAD (authenticated encryption with associated data), providing both confidentiality and integrity. Without authentication, chosen-ciphertext attacks against decryption oracles can leak plaintext. Large quantum computers running Shor's algorithm would break RSA and elliptic-curve schemes, which motivated NIST's post-quantum standards such as CRYSTALS-Kyber. In practice, key management is the dominant attack surface.

Scope of Application¶

Encryption lives across the cryptographic-security cluster — the disciplines sharing both the keyed-transformation mathematics and the adversary-with-bounded-compute threat model.

Computer security — at-rest, in-transit (TLS), and end-to-end (Signal) confidentiality plus hybrid schemes.
Signals intelligence and diplomacy — substitution ciphers, one-time pads, Enigma: the same math by hand.
DRM and content protection — media encrypted under license keys.
Authentication-adjacent constructions — digital signatures, MACs, secure-boot chains anchored in a root key.
Post-quantum migration — the same reasoning re-sized against a quantum adversary (Kyber, Dilithium).

Clarity¶

Naming encryption separates it from encoding (reversible, no secret, conceals nothing) and hashing (irreversible, no recovery). It forces the right question — is there a secret, and does the operation reverse? — whose answer fixes the intervention family. It also makes confidentiality a local property: instead of asking who can read a message across every node and cable, the engineer asks the single question "who holds the key?", relocating the audit from the channel to key custody.

Manages Complexity¶

The raw confidentiality problem is unbounded — every node, cable, disk, and operator the bytes touch. Encryption collapses that surface to one quantity: who holds the key. The algorithm drops out (published standards assumed sound), leaving the key-management surface and threat model. The key arrangement reduces to a trust-context branch, the hardness guarantee to one comparison (assumed-hard problem versus adversary), and confidentiality-versus-integrity to whether AEAD is bundled in.

Abstract Reasoning¶

Encryption licenses boundary-drawing (classify the transformation and trust context before choosing a defense); a diagnostic move localizing confidentiality to key custody, not topology; threat-model reasoning that sizes every guarantee against a specified adversary and watches the integrity gap (chosen-ciphertext attacks, AEAD); and forward-secrecy reasoning about whether past traffic survives a key compromise.

Knowledge Transfer¶

Within the cryptographic-security cluster encryption transfers as mechanism — the transformation classification, the symmetric/asymmetric/hybrid branch, key-custody localization, threat-model sizing, and forward secrecy carry intact across computer security, classical cryptanalysis, DRM, and signatures, because these are one substrate. Beyond it, "genetic encoding" or "the diploma encrypts your transcript" borrows only concealment while dropping the secret-gated transformation. The parents that travel are Information Asymmetry, Access Control, and Authentication; the asymmetric-inverse-difficulty lesson travels as trapdoor-function reasoning, a math/CS structure encryption instantiates.

Relationships to Other Abstractions¶

Current abstraction Encryption Domain-specific

Parents (2) — outgoing typed relations

Encryption is a kind of Encoding And Decoding Prime

Encryption is Encoding and Decoding specialized to a secret-gated scheme whose code conceals content from parties lacking the decryption key.
Encryption is a decomposition of Asymmetry Prime

Removing cipher vocabulary leaves a directed, non-interchangeable cost relation between a key-holder and a party lacking the key.

Children (1) — abstractions linked here by typed relations

Public-Key Cryptography Domain-specific includes, conditional Encryption

Public-Key Cryptography contains Encryption in its confidential-transmission branch, where anyone encodes to a public key and only its private mate decodes.

Hierarchy paths (2) — routes to 2 parentless roots

Encryption → is a kind of Encoding And Decoding → includes Transformation → is a kind of, typical Function (Mapping)

Show alternative path (1)

Neighborhood in Abstraction Space¶

Encryption sits in a sparse region of the domain-specific corpus (66^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (2769 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25