Key points are not available for this paper at this time.
The explosive data growth in high-performance computing (HPC) puts pressure on systems to process huge amounts of data and poses threats to the security of important data during transmission. Traditional data protection methods such as encryption inevitably attracts intermediate intercepting entities' attention. As a means of hiding information within other irrelevant data (carrier), steganography is used to transmit critical data without arousing the attention of regulators. Error-bounded lossy compression is a data reduction technique that effectively alleviates system pressures due to large data volumes. In this paper, we propose a steganography scheme based on the lossy compressor SZ named StegaZ. StegaZ performs steganography while compressing data by selecting random bits for insertion based on the password entered by the user. StegaZ does not affect unaware users' normal usage of the decompressed dataset. The experimental results show StegaZ preserves more than 99.6% of the original compression ratio and achieves a PSNR of 100% when selecting an appropriate dataset. Additionally, it imposes minimal compression bandwidth overhead, sometimes even able to obtain a higher compression bandwidth than the original.
Shan et al. (Tue,) studied this question.