Trusted Computing

Trusted Computing is a technology system that aims to improve the security and trustworthiness of computer systems, ensuring that the entire computing environment and process from hardware to software levels are trustworthy.

The main goal of trusted computing is to create a computing environment that can self-verify its own integrity and trust status, thereby enhancing the security and reliability of communication and business processes in the network environment.

We use the IMA files on the machines to perform remote computation and verification continuously. First, the current node (worker role) acts as the verifier and obtains the latest verification data from the blockchain. It finds the top N nodes with high trust scores as Light Verifiers and sends verification requests.

When the light verifier verifies the current node successfully, it will add points to the current node. When the verification fails, the light verifier will deduct points from the worker node and impose a verification frequency penalty, making it unable to be verified again in a short time.

When the light verifier completes the verification of the worker, it will obtain the top N nodes with high scores from the blockchain again and broadcast the results. When these N nodes receive the broadcast, they will merge the received information with the local information and judge if they are the node with the highest score (then they are the full verifier), then they will put the results on the chain.

Above we described the verification based on IMA.

In the practice process, we also considered the whitelist of internal commands of the machine. When there is an executable file that is not in the whitelist in the IMA file, we think that this node is at great risk.

For trusted verification, we still think it is a verification failure. In addition to punishment, we will update the whitelist.

We need to emphasize that the worker, light verifier, and full verifier roles are not fixed in the network. As the trust score changes, the role of the node will change. But what can be determined is that every node in the network contains the worker role. Only in this way can all the nodes in the network run within the trusted theory.