Scott Fluhrer edited untitled.tex  almost 9 years ago

Commit id: 3f89108a19464d27205642bceab88398f5d0444c

deletions | additions      

       

However, there are ways to avoid this issue; we will review two different protocols, and give a possibility of how this can be addressed in both of them.  One note: the above point multiplication analysis assumed a balanced base-48 notation, while the below will assume a standard base-48 notation. This is because standard base-48 notation is easier to do arithmetic in, while it is not difficult to convert to a balanced notation, if that would be helpful to the point multiplication logic.  \subsection{The case of ECDH/ECIES}  The easiest case to handle is the case of ECDH/ECIES. In these cases, the integer that we multiply by is just a random number that we pick, and has no correlation with any other value (with the exception that we multiply two different points by the same integer).  In this case, we can avoid the initial problem (how do we convert the binary integer into base-48 without giving a side channel attack) simply by selecting the initial random number in base-48. That is, we never explicitly express the multiplier in binary; instead, we pick a series of random base 48 digits.  \begin{itemize}  \item ECDSA  \item Performance estimates