Optimizing Failure Rates in Trustee-Based Recovery Systems

Table of Links Abstract and 1. Introduction Abstract and 1. Introduction Related Work 2.1 The Alternative-Authenticator Approach 2.2 The Original-Authenticator Approach The Proposed Secret Backup Approaches 3.1 Notations 3.2 Assumptions 3.3 The Direct-Escrow Method 3.4 Our Proposed Algorithms Security and Reliability Analysis 4.1 Security Analysis 4.2 Reliability Analysis 4.3 Recovery Failure Rate Analysis 4.4 Real World Parameters 4.5 Failure Rate Optimization of (k,n) Comparison Conclusion, Acknowledgment, and References Related Work 2.1 The Alternative-Authenticator Approach 2.2 The Original-Authenticator Approach Related Work 2.1 The Alternative-Authenticator Approach 2.1 The Alternative-Authenticator Approach 2.2 The Original-Authenticator Approach 2.2 The Original-Authenticator Approach The Proposed Secret Backup Approaches 3.1 Notations 3.2 Assumptions 3.3 The Direct-Escrow Method 3.4 Our Proposed Algorithms The Proposed Secret Backup Approaches 3.1 Notations 3.1 Notations 3.2 Assumptions 3.2 Assumptions 3.3 The Direct-Escrow Method 3.3 The Direct-Escrow Method 3.4 Our Proposed Algorithms 3.4 Our Proposed Algorithms Security and Reliability Analysis 4.1 Security Analysis 4.2 Reliability Analysis 4.3 Recovery Failure Rate Analysis 4.4 Real World Parameters 4.5 Failure Rate Optimization of (k,n) Security and Reliability Analysis Security and Reliability Analysis 4.1 Security Analysis 4.1 Security Analysis 4.2 Reliability Analysis 4.2 Reliability Analysis 4.3 Recovery Failure Rate Analysis 4.3 Recovery Failure Rate Analysis 4.4 Real World Parameters 4.4 Real World Parameters 4.5 Failure Rate Optimization of (k,n) 4.5 Failure Rate Optimization of (k,n) Comparison Comparison Comparison Conclusion, Acknowledgment, and References Conclusion, Acknowledgment, and References Conclusion, Acknowledgment, and References Appendix Appendix 4.4 Real World Parameters We adopt appropriate parameters to estimate the failure rate F by referring to available statistical or published research data. We assume a reasonable range of values for analysis for those with no referenced data. The real-world data references are more realistic, although sometimes they differ from theoretical numbers. For example, a long enough password is theoretically hard to guess, but in reality, people tend to use some common words for convenience. Therefore, a test on real-world password sets is closer to the realistic security level of the password approach [42]. Then in equation 2, only the trustee unavailable rate U is to be identified. Although the actual U value is known only until our method's real deployment, we assume 0.1% failed trustees for later calculations. 4.5 Failure Rate Optimization of (k,n) We then compute the probabilities P (insecurity rate), Q (unreliability rate), and F (failure rate) using the values quoted in the previous subsection. For example, we show in Fig. 3 the failure rates of different recovery thresholds k's for the choice of 6 trustees. Therefore, for each choice of n, one may calculate the optimum k. Hence, we show in Fig. 4 the minimum failure rate and the corresponding k value for each option of the number of trustees n. Practically, one may determine the target allowable failure rate and use this table to choose the optimal parameters of (k, n). Authors: (1) Wei-Hsin Chang, Deepmentor Inc. (mailggwhc@gmail.com); (2) Ren-Song Tsay, Computer Science Department, National TsingHua University, Hsinchu, Taiwan (rstsay@cs.nthu.edu.tw). Authors: Authors: (1) Wei-Hsin Chang, Deepmentor Inc. (mailggwhc@gmail.com); (2) Ren-Song Tsay, Computer Science Department, National TsingHua University, Hsinchu, Taiwan (rstsay@cs.nthu.edu.tw). This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv available on arxiv