CS579/ECE599 Assignment Homework 2
Python算法代写 Please complete this assignment (200 pts total) and submit your report/program code on Canvas (all files compressed in one .zip without
Please complete this assignment (200 pts total) and submit your report/program code on Canvas (all files compressed in one .zip without the .bin files)
For side-channel analysis, it is often necessary to deal with substantial amounts of data that typically exceed the memory available in a computer. Consequently, it is not possible to load all data into memory and compute the result directly. Instead, the result must be computed in incremental steps. In addition, working with large data sets can introduce numerical instabilities. Of course, this can lead to significant problems and make a successful analysis impossible. Therefore, you are asked to investigate the numerical stability of your preferred programming environment to avoid follow-up problems in subsequent homework assignments.
1. Python算法代写
The file measurement_data_uint8.bin contains 1 billion samples of type uint8. This is representative of actual measurement data, as many oscilloscopes have 8 bit of resolution due to the Analog-to-Digital Converter (ADC) being used. Your task is to compute the mean and variance of this data by applying different methods. Note: this is about incrementallyprocessing the data and updating intermediate variables, to then compute the final result.
Do not use any readily available functions for mean/variance (90 pts)
a) Naïve approach ∗ (5 pts)
b) Welford’s algorithm (cf. Wikipedia or other online reference) (10 pts)
c) One-pass arbitrary order method † (35 pts)
d) Histogram method ‡ (35 pts)
e) Compare the runtime of the different methods (5 pt)
Note: you can confirm your result by comparing the mean with the standard deviation. Aside from the shift of the decimal point, the initial four digits are the same and will be easily recognized as part of the hacker culture (“1337”).
For the next two tasks, please consider the following files and the notation in the appendix:
- traces_10000x50_int8.bin, containing = I0|| . . . ||IN, whereas = 10000 and = 50
- plaintext_10000x16_uint8.bin, containing the corresponding plaintexts = P0|| . . . ||PN
∗https://en.wikipedia.org/wiki/Algorithms_for_calculating_variance
† https://eprint.iacr.org/2015/207.pdf, only Section 4 and Appendix A
‡ https://eprint.iacr.org/2017/624.pdf, only Section 3 and 4
2. Python算法代写
In the lecture, we learned that not only key extraction may be a viable goal of a side-channel analysis but also leakage detection. In particular, leakage detection may help to choose points for more computationally demanding attacks or as part of a security certification process to confirm the lack of points of interest. The most basic leakage detection is the SNR (cf. Appendix). To develop a natural understanding of what is happening, first compute the numerator and denominator separately. (20 pts)
a) Compute the numerator (‘signal’) of the traces and plot the result. (5 pts)
b) Compute the denominator (‘noise’) of the traces and plot the result. (5 pts)
c) Compute the whole SNR and plot the result. (10 pts)
Caveat: many software libraries treat 0 differently when computing, e.g., variance. In numpy, you should use, e.g., where=means != 0 to exclude zeros in your computation.
3. Python算法代写
Perform a Correlation Power Analysis (CPA) on the traces to extract they key. Do not use any external libraries that provide readily available functions for side-channelanalysis.
Please report the execution time of your attack. (90 pts)
Please write all your programs in one of the following languages/environments: Python/Jupyter (strongly recommended), C/C++, Rust, Java, Matlab/Octave. Thoroughly document your code using comments. Your .zip file should contain your code, instructions how to make it run (if needed), the figures, etc.
更多代写：cs代写加拿大 美国gre代考 英国monash.代考 英语论文开头 apa格式标题 simpletense代写
