Trustworthy Data Analytics in the Cloud Using SGX

Explore a 22-minute IEEE conference talk on VC3, a groundbreaking system enabling secure MapReduce computations in the cloud using Intel SGX. Delve into the innovative approach that maintains code and data confidentiality while ensuring result correctness and completeness. Learn how VC3 operates on unmodified Hadoop, keeping it and other large components out of the Trusted Computing Base. Discover the system's use of SGX processors for memory isolation and new protocols securing distributed computations. Examine the optional enforcement of region self-integrity invariants to prevent unsafe memory access attacks. Analyze experimental results comparing VC3's performance to unprotected Hadoop, revealing negligible average runtime overhead for base security guarantees and minimal increases with added integrity measures.

Introduction
Outline
MapReduce
Intel SGX
Challenges
Partitioning
Security Invariants
Evaluation
Conclusion
Questions