This paper describes some software optimizations for AES-GCM over short messages, applicable for modern processors that have dedicated instructions. By processing two (short) messages in parallel, we achieve better performance than by processing twice, back-to-back, a single (short) message. Additional performance is gained if the using application collects several messages, sorts them by order of length, and the feeds them (in pairs) to the two-message AES-GCM function. For example, our experiments carried out on the latest Intel processor (micro architecture codename Skylake), over a realistic distribution of message lengths, our optimization achieves up to 1.95x speedup, compared to OpenSSL.
|Title of host publication||Information Technology – New Generations - 14th International Conference on Information Technology|
|Number of pages||5|
|State||Published - 2018|
|Event||14th International Conference on Information Technology - New Generations, ITNG 2017 - Las Vegas, United States|
Duration: 10 Apr 2017 → 12 Apr 2017
|Name||Advances in Intelligent Systems and Computing|
|Conference||14th International Conference on Information Technology - New Generations, ITNG 2017|
|Period||10/04/17 → 12/04/17|
Bibliographical noteFunding Information:
Acknowledgements This research was supported by the ISRAEL SCIENCE FOUNDATION (grant No. 1018/16).
© Springer International Publishing AG 2018.
- Software optimizations
ASJC Scopus subject areas
- Control and Systems Engineering
- Computer Science (all)