Download presentation
Presentation is loading. Please wait.
Published byAusten Eaton Modified over 8 years ago
1
Scalable Verifiable Encrypted Search Encrypted Search with Third Party Support and Protection From Dishonest Data Stores
2
What is Encrypted Search? Data is “big” nowadays and storage is often outsourced Clients must be able to store and access their data Inevitable leaks plague this concept and make outsourcing impossible for sensitive data “Encrypted search” is a family of protocols where: data is encrypted on the client side and stored elsewhere data is searchable and accessible by the client using encrypted search keys the server does not know the contents of the data or the keys
3
+Scalable? In the past, encrypted search has been implemented but did not perform well. Solutions with O(>n) do not scale to large datasets. Recently (in this decade) papers have been published outlining a sub-linear protocol. Another nuance is that clients must be able to allow third parties to search their data. This is a challenging cryptography probably that has also been recently solved. Experimental implementations for proposed solutions exist.
4
+Verifiable? Current solutions assume an honest but curious server Inferences can be made from leaked information, such as number of documents that match the query and relative size of documents So, even if the server responds accurately, it may be maliciously harvesting information about clients’ data It is not unreasonable to assume that there will be dishonest servers as well, e.g. those that return less documents than actually matched No implementation exists for verifiable encrypted search, but a paper outlining a solution has been published by Professor Rosario Gennaro.
5
Our Goal Add “verifiability” to the OSPIR-OXT protocol developed in a joint project by IBM, MIT, and UC Urvine Make Scalable Verifiable Encrypted Search easy to use Current implementation is a mess of scripts written in several languages Requires a specific system for compilations and run time Very difficult to create new schemas and generate data We want this to be available to everyone
6
Current stage in the process Got IBM Espada project that implements the OSPIR-OXT protocol to compile and run. Now figuring out how to generalize data generation to any schema.
7
Thanks! Johannes Christ and Daniel Intskirveli
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.