sebgaj

Hyperscalers like AWS, Azure or Google transition their datacenters to a confidential computing stack to protect their cloud against the growing invasion of hacks (eg. implemented by Russian, Chinese, or North Korean special hacking teams).

Do you think CC is the right mitigation to all the cloud threats?

This project enables you to engage in secure and cost-effective conversations while safeguarding sensitive data and files:

- prompt ChatGPT with extensive documents or a large number of documents, overcoming the limitations of GPT-3's context capacity. For instance, it enables tokenization of materials such as books, electronic patient records, or contracts.

- opt in or out of pseudonymizing prompts, ensuring privacy by sanitizing personal information such as names, email addresses, and street names. If uncomfortable with potentially sensitive content within prompts, automated suggestions can be manually post-edited for added security and peace of mind.

Confidential Cloud computing is the answer and hyperscalers work on that. You literally put your data in a black box.

I would take any database (eg Mongo, Maria, …) patch out the „delete“ queries, and run the DB in a confidential cloud like enclaive.cloud. Leveraging attestation and confidential compute, one can add a proof that the DB does not allow for erasure. Running the DB in a (kubernetes) cluster on multiple clouds would ensure the fault tolerance known from blockchain ledgers.

Fully agree. The architecure design is weak, given the fact that TEE-based approaches allow the hardening of process isolation.

What bothers me is not that the hack is done in 3min, but the capability to open the front trunk and door while driving

HashiCorp vault is the de-facto key management solution to rotate keys in docker, docker swarm or k8s environments. So far, HashiCorp Vault allowed for data-at-rest encryption of keys and secrets.

The open sourced project "always encrypted Hashicorp vault" by team enclaive makes sure keys and secrets are encrypted at runtime (aka data-in-use encryption) without changing or limiting the functionality and performance of the vault.

Always encrypted Hashicorp vault is an attractive alternative to HSMs, allowing businesses to drastically reduces security expenses.

Hyperscalers are silently working on technologies to make cloud the safest places for workloads. Among the technologies is confidential cloud computing, isolating for the very first time applications from the infrastracture.

If folks are interested in running applications in Intel SGX enclaves in the cloud, giving your workload extra security and privacy power, follow this tutorial.

My argument is based on the increased entropy of cryptographic keys replacing user-chosen passwords. So brute-forcing and social engineering becomes harder.

Clearly more secure than password based authentication. I hope it will finally kill phishing attacks.

One of the top tier open source conferences in Europe holds a track on confidential computing. Videos available after the conference.

Researchers found security pitfalls in IBM’s cloud infrastructure

Video explains the principles of confidential cloud computing. A new technology hyperscalers are slowly adding to their offerings.

Nice job COSIC folks!