Home / Knowledge hub / Homomorphic encryption
Knowledge hub
Homomorphic encryption
Encryption that lets you compute on data without ever decrypting it.
What it means
Homomorphic encryption is a form of encryption that allows calculations on data while it stays encrypted. The result is also encrypted, and only the holder of the key can read it. The party doing the work never sees the plain values.
Some schemes support only one kind of operation, such as adding numbers. Fully homomorphic schemes support arbitrary calculations, which was shown to be possible in 2009, but they cost far more computing power than working on plain data.
Why it matters
It offers a way to use outside computing power on sensitive data, for example analysing records held by a cloud provider without giving the provider the readable data. That is attractive in health, finance and any setting where sharing raw data is restricted.
It is also easy to over-promise. A product page that says data is processed encrypted may only mean that data is encrypted at rest and in transit. The difference matters when you answer a security questionnaire or a customer's legal team.
How it shows up
In real applications you meet it rarely, mostly in research, specialist privacy products and a few narrow uses such as private lookups or summed totals. Most web apps rely on encryption at rest and in transit and decrypt data to work on it.
The mistakes are about claims and misuse. Teams assume it hides how much data is processed or who asked. Keys are mishandled, which defeats the scheme. A home-made version is built from unpadded RSA. Ask what is actually encrypted during processing, and who holds the key.
Questions people ask
Homomorphic encryption, answered
Is homomorphic encryption the same as end-to-end encryption?
No. End-to-end encryption keeps data unreadable in transit between two ends, and those ends decrypt it to use it. Homomorphic encryption lets a third party compute on the data without decrypting it.
They solve different problems and can be combined.
Is fully homomorphic encryption practical?
For some narrow tasks, yes. For general use it is still much slower and heavier than normal processing, so it is chosen only where the privacy benefit justifies the cost.
Expect specialist libraries and careful design rather than a switch you turn on.
Does it protect against a compromised application?
Only in part. If an attacker controls the client that holds the key, or the app that asks for results, the secrecy is gone. The scheme protects the server doing the computing, not the whole system.
Ordinary access control still applies around it.
Is it part of post-quantum cryptography?
Not exactly. Many modern schemes are built on lattice problems that are also used in quantum-safe cryptography, but being homomorphic and being quantum-resistant are separate properties.
Check the specific scheme's documentation.
Should a small software company use it?
Usually not yet. It is worth a look if you must process regulated data on infrastructure you do not trust and cannot avoid it. Otherwise standard encryption, strict access rules and data minimisation do more for less effort.
Start with the simpler controls.
Related
Terms that sit next to this one
Symmetric encryption
Encryption where the same secret key locks and unlocks the data.
Public key cryptography
Encryption with a pair of keys: one shared openly, one kept private.
Encryption at rest
Stored data scrambled so a stolen disk is useless without the key.
Quantum-safe cryptography
Encryption designed to stay secure against future quantum computers.
Data minimisation
Collecting and keeping only the personal data you actually need.
What is actually encrypted while it runs?
Whatever the cryptography promises, the app around it still has to enforce who can ask for what. We find vulnerabilities in web apps and APIs and prove them. Ask us to run it on one of your applications.
Or start with a $199 pilot on one application: thirty days, success criteria agreed before day one, credited against the annual if you convert.