Dependable Industrial Crypto Computing

This paper presents a work on how to assure the dependability of a crypto system built across on and off the blockchain by using the proposed adaptive checkpoint and rollback algorithm, and a prototype is developed for demonstration purpose. The theoretical background of the proposed checkpoint and rollback algorithm is studied to characterize the variables affecting the dependability such as security, authenticity and reliability with respect to the rates of hit by any events of those issues, the rates to detect and diagnose, and then the rate to vote for a consensus whether to trigger a rollback or not. Based on the variables characterization in a stochastic manner, then steady state probabilities and state transition probabilities are derived in order to assure the ultimate effective dependability of each individual dependability variable (i.e., security, authenticity and reliability), then finally to assure the dependability in a compound manner with each variable assigned a weight depending on the nature of the systems specifications. Based on the theoretical study, a protype of a crypto system is built to demonstrated the underlying architecture and operations and to justify the need for such system to take synergistic advantages from both on- and off-chain blockchains, with an experimental result of a benefit in gas fee which is the most exigently addressed issue today in blockchain systems especially in Ethereum network of blockchains. An astonishing gas fee saving results are demonstrated. It is observed that the crypto system benefits more if more computationally intensive transactions are executed off-chain while vice versa.