User manual VIA KT133A BROCHURE

[. . . ] VIA Technologies and Centaur Technology make no representations or warranties with respect to the accuracy or completeness of the contents of this publication or the information contained herein, and reserves the right to make changes at any time, without notice. VIA Technologies and Centaur Technology disclaim responsibility for any consequences resulting from the use of the information included herein. VIA and VIA C3 are trademarks of VIA Technologies, Inc. CentaurHauls is a trademark of Centaur Technology, Inc. . [. . . ] They are predictable in the sense described above (i. e. , forward prediction if the initial conditions are known). Like all software algorithms, they are reproducible: they produce the same results starting with the same input Their security relies on an assumption that adversaries do not have enough computational power to perform certain mathematical operations. As a result, their irreversibility cannot be proved, although, in practice, this is not a major problem. There is a way to improve the situation here. Rather than relying solely on a software generator to produce all of the randomness, a software generator can be used in conjunction with a source of true randomness (entropy). Instead of using the previous result as the seed for the next number to be calculated, output from the independent source of entropy can be used as the seed. Even if the entropy source is imperfect (provides less than one bit of entropy per output bit), the results in this case can be more robust than what could be generated by either the software or the entropy source alone. An example is using a hardware random number generator (see the next section) to produce the seed for a software algorithm. Although a variety of designs are used, the software algorithms are often based on cryptographic hash functions. A hash algorithm mixes up its input bits in a fashion such that its output bits should be unpredictable unless the entire input is known. Thus, a good hash algorithm effectively destroys any statistical correlations among the input bits and makes it computationally infeasible to recover the input from the output. (In the Microsoft CryptoAPI and Linux examples, a SHA-1 algorithm is applied to the entropy seeds). Other common problems with entropy generators on computers are that they require hooks in the operating system, they are difficult to test, they often require some user involvement, and they are slow (since they are based on macro physical events). An unusual example of a physically-based random source is available through www. random. org, which continually provides generated random numbers using atmospheric noise (sampled as radio frequency noise). The rate varies, but this generator typically produces about 40, 000 bits per second. True Hardware Generators By now you should be convinced that good security requires good random numbers, and while it is possible to generate pretty good statistically random numbers using software, it is difficult, slow, and subject to numerous security pitfalls. The only truly random generator is some mechanism that detects quantum behavior at the sub-atomic level. This is because randomness is inherent in the behavior of sub-atomic particles (remember quantum mechanics, Heisenberg's Uncertainty Theorem, etc. ). A quantum based hardware generator is actually practical. Examples that have been used are: VIA KT133A Security Application Note - 15 The interval between the emission of particles during radioactive decays. This technique is used at a web 18 site to generate random bits that are available on-line from the decay of Krypton-85. This source generates only 30 bytes per second and requires a cumbersome (and dangerous?) pile of hardware. [. . . ] We have tested a number of KT133A chips in a variety of multitasking environments, with as many as 10 tasks concurrently running version of our ACE test package. No ACE failures have been observed in production steppings of the VIA KT133A. Cryptography Research, Inc. , has recently completed a (brief) security evaluation of the Advanced Cryptography Engine. A copy of the report is available at: www. somesite. org From their summary: "Marketing quote from CRI" Customers interested in testing the ACE for themselves should obtain a copy of the VIA KT133A Advanced Cryptography Engine Programmers Guide for programming details. [. . . ]