Free
Message: I am just trying to understand

There are others as well that measure ......

From US8572440 B1(signet)

"Some techniques offer data access security through the use of data related to physical characteristics of the storage media. These systems rely on precise measurement of solid state media process characteristics. Examples of these include variations in the remnant charge of EEPROM's, or variations in row/column addressing circuits. These and other characteristics have been used for authentication when the memory is presented, and the authenticating characteristics are measured. Such techniques are described in detail in Fernandez (U.S. Pat. No. 5,644,636).

The critical issue which cannot be addressed by the prior art are applications where data security is not simply shifted to attacks on the reader mechanisms and where data must not be duplicated or separated even by trusted and/or authorized users."

Measuring different levels of a ceratin chemical composition of chips would, IMO, fall into the above catogory.

US8572440 B1 is about natural or induced defects, and it's not just about semiconductor defects, ....but material composition defects as well.

"Also, while not required in all aspects, the manufacturer of the memory array could introduce or allow for higher error rates on the memory array, thereby improving the number of defective locations and the uniqueness of the reference map. As such, while the memory array would have a reduced storage, the memory array would also be more secure and identifiable according to aspects of the invention. A further advantage is that the manufacturer costs could be reduced since, with a higher error rate being allowed, fewer memory arrays would be deemed unacceptable and the manufacturing tolerances could be loosened to allow more of the defective locations to be introduced into the resulting memory array. By way of example, the lessening of manufacturing tolerances to purposefully introduce errors randomly in semiconductor, magnetic or optical would provide a greater abundance of intermittent error (i.e., while one can over voltage a memory location to force it to ground or power, but it cannot be made intermittent where it loses it's data over a couple of clock cycles). For low cost magnetic equivalents, such introduction would equate to inadequate distribution of the ferric material in a magnetic strip and disk drives would be more like semiconductors (i.e. process contamination, etc.). Optical memory could allow bits being inscribed onto the aluminum in a loose fashion such that a segment of information (likely in a known place) will be read differently at times as the tolerance of the reader has been disregarded to create interspersed errors. However, other methods can be used to introduce or allow such errors at the manufacturing stage in other aspects of the invention."

doni


Share
New Message
Please login to post a reply