Policy-Compliant Cloud Storage Systems
This paper presents GDPRuler, a trusted middleware system that enables verifiable GDPR compliance for unmodified key-value stores on untrusted clouds by deploying a policy-enforcing monitor within a Confidential Virtual Machine, achieving strong compliance guarantees with low performance overhead and efficient metadata management.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine you have a very fast, super-efficient warehouse (a Key-Value Store) where you keep millions of boxes. Each box has a simple label (a key) and contains your personal items (the data). This warehouse is great for speed, but it's located in a building owned by a stranger (the Cloud). You are worried that the building owner might peek inside your boxes, change the labels, or refuse to let you take your things back when you ask.
Furthermore, there are strict new laws (GDPR) that say:
- You can only look at your own boxes.
- You can only use your boxes for specific reasons you agreed to (like "shopping," not "marketing").
- If you say "delete this," it must be gone forever, and there must be a record that it happened.
- You must be able to prove to a police officer (the Regulator) that you followed these rules.
The problem is that this fast warehouse wasn't built with these rules in mind. It's too simple to understand complex legal instructions, and the building owner (the cloud provider) is technically powerful enough to ignore your rules if they wanted to.
The Solution: GDPRuler
The paper introduces GDPRuler, which acts like a super-secure, invisible security guard standing right between you and the warehouse.
Here is how it works, using simple analogies:
1. The "Glass Box" (Confidential Virtual Machine)
Usually, if you hire a security guard, the building owner could still peek over the guard's shoulder or even bribe them. GDPRuler puts the guard inside a Confidential Virtual Machine (CVM).
- The Analogy: Imagine the guard is working inside a transparent, unbreakable glass box that is locked from the outside. Even the building owner cannot see inside the box, touch the guard, or change what the guard is doing. The guard can verify their own identity to you using a special digital ID card (Remote Attestation) to prove they are real and haven't been tampered with.
2. The "Smart Translator" (Policy Language)
The warehouse only speaks "Key" and "Value." It doesn't speak "Law."
- The Analogy: GDPRuler has a translator that turns your legal wishes (like "Only use this data for shopping") into simple instructions the warehouse understands. When you ask the warehouse to do something, the translator checks your ID and your rules first. If you try to use the data for "marketing" when your rule says "shopping only," the translator stops the request before it even reaches the warehouse.
3. The "Tiny Sticker" (Compact Metadata)
To enforce rules, the system needs to attach a list of rules to every single box. If this list were a giant book, it would slow the warehouse down to a crawl.
- The Analogy: GDPRuler writes the rules on a tiny, invisible sticker that fits perfectly on the box without adding bulk. It uses a special code (bitmaps) so the sticker is super small. This means the warehouse stays fast, even though every box now has a rule attached to it.
4. The "Special Index" (Metadata Indexing)
If you want to find all boxes belonging to "Alice" or all boxes used for "Marketing," a normal warehouse would have to open every single box to check the label. This is slow.
- The Analogy: GDPRuler builds a special library card catalog just for these rules. Instead of opening every box, the guard just looks up "Alice" in the catalog, gets a list of the specific boxes, and goes straight to them. This makes finding your data or deleting it 13 to 182 times faster than without the catalog.
5. The "Tamper-Proof Diary" (Audit Logs)
The law says you need a record of every time someone touches your data. If the building owner could erase this diary, the rules wouldn't matter.
- The Analogy: GDPRuler keeps a diary inside the glass box. Every time a box is opened, moved, or deleted, a note is written in the diary. This diary is sealed with a special wax stamp (cryptographic signature) that breaks if anyone tries to change a word. If a regulator asks to see the diary, the guard shows them the sealed pages, proving that no one has faked the history.
Does it slow things down?
The authors tested this system with real-world traffic.
- Speed: The system is about 61% as fast as the warehouse running alone. Most of this slowdown (about 30%) is just because the "glass box" (the secure environment) takes a little extra time to work. The actual "security guard" work (checking rules) only adds about 8% more delay.
- Space: The tiny stickers (metadata) take up less than 20% extra space on the shelves.
- Logging: Writing the diary entries is so efficient that it barely slows down the warehouse at all (less than 2% slowdown).
The Bottom Line
GDPRuler proves you can have a fast, modern cloud warehouse and still strictly follow privacy laws, without needing to rebuild the warehouse or change your software. It does this by placing a trusted, unbreakable security guard in a glass box who translates your legal rights into fast, enforceable actions, all while keeping a tamper-proof record of everything that happens.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.