The Fact and Fiction of Homomorphic Encryption
The approach's promise continues to entice cryptographers and academics. But don't expect it to help in the real world anytime soon.
The history of homomorphic encryption stretches back to the late 1970s. Just a year after the RSA public-key scheme was developed, Ron Rivest, Len Adleman, and Michael Dertouzos published a report called "On Data Banks and Privacy Homomorphisms." The paper detailed how a loan company, for example, could use a cloud provider (then known as a commercial time-sharing service) to store and compute encrypted data. This influential paper led to the term "homomorphic encryption."
What Is Homomorphic Encryption?
Homomorphic encryption describes any encryption scheme in which a particular operation on two ciphertext values gives an encrypted result, which, when decrypted, maps to the result the operation would have been in plaintext. Because the keys are needed only during the initial encryption and final decryption, complete privacy of the inputs and outputs is maintained during the computation process.
The purpose of homomorphic encryption is to allow computation on encrypted data. The process describes the conversion of data into ciphertext that can be analyzed and worked with as if it were still in its original form.
Why It Works (At Least in Theory)
Homomorphic encryption can be a significant asset to your business compliance and data privacy efforts. We all know the regulatory landscape changed dramatically in 2018. The EU's General Data Protection Regulation (GDPR) came into force last May, the California Consumer Privacy Act (CCPA) is scheduled to be implemented on January 1, 2020, and as many as 40 other states are considering data privacy laws. GDPR was notable because of its penalty of 4% of global revenue for serious infractions, such as not having sufficient customer consent to process data. The CCPA is a "mini-GDPR" with unprecedented power for consumers to control the collection, use, and transfer of their own data.
These new regulations came about as the result of significant breaches and abuses of data over the past few years. There is an undeniable data breach fatigue these days. But not every breach comes about as the result of outside attackers. There are other threat models to consider. Insider threats and privileged access account for about 35% to 60% of breaches, according to industry reports. Anthem and MyFitnessPal fell victim to this type of attack. Even using traditional encryption (encryption at rest and transparent data encryption, for example), database administrators have access to all of your data in the clear. They have access to the crown jewels.
Homomorphic encryption can also be a business enablement tool. It can allow cloud workload protection ("lift and shift" to cloud), cloud/aggregate analytics (or privacy preserving encryption), information supply chain consolidation (containing your data to mitigate breach risk), and automation and orchestration (operating and triggering off of encrypted data for machine-to-machine communication).
Where Homomorphic Encryption Falls Short
Homomorphic encryption originally slowed mathematical computations down to a crawl. The initial performance impact was 100 trillion times slower (not a typo). There has been significant performance improvement since then, but the latest figure is that about 50,000 end-to-end transactions can be performed in a certain range of time. That amount is still too small in today's fast-moving world.
Furthermore, homomorphic encryption requires application modifications. You need prior knowledge of what type of computation was being performed (additive, multiplication, etc.) Businesses with less predictable or more free-form operations will have to rewrite or modify applications to make homomorphic encryption viable. Again, that's not feasible for businesses at scale.
Finally, there are still questions about the overall encryption strength (encryption entropy). Homomorphic encryption exposes valuable properties in achieving this mechanism without decrypting this data, and without access to keys. There are still some open questions about the encryption strength using a scheme like this.
Putting It All Together
Homomorphic encryption is a long way off from real-world enterprise implementation, but there has been substantial progress in the areas of differential privacy and privacy preservation techniques. There are tools that can deliver homomorphic-like encryption without the inherent drawbacks that homomorphic encryption brings, so that businesses can mandate a higher security standard without actually breaking processes or application functionality.
Businesses should not have to sacrifice speed for security — it's not a zero-sum dynamic. When done properly, security can actually accelerate your business. Eliminating that friction between security and business leadership builds trust between departments and creates better security outcomes.
The promise of practical homomorphic encryption continues to entice cryptographers and academics. Although it is a rapidly developing area to watch, in practice, its poor performance to date makes homomorphic encryption impractical to implement in enterprise environments in the near future.
Related Content:
About the Author
You May Also Like
DevSecOps/AWS
Oct 17, 2024Social Engineering: New Tricks, New Threats, New Defenses
Oct 23, 202410 Emerging Vulnerabilities Every Enterprise Should Know
Oct 30, 2024Simplify Data Security with Automation
Oct 31, 2024