Bug #46546 | Large performance hit when using YaSSL over OpenSSL | ||
---|---|---|---|
Submitted: | 4 Aug 2009 16:02 | Modified: | 25 Jul 2019 11:20 |
Reporter: | Andrew Hutchings | Email Updates: | |
Status: | Won't fix | Impact on me: | |
Category: | MySQL Server: General | Severity: | S3 (Non-critical) |
Version: | 5.0.83 | OS: | Any |
Assigned to: | CPU Architecture: | Any |
[4 Aug 2009 16:02]
Andrew Hutchings
[4 Aug 2009 16:57]
Andrew Hutchings
Callgrind for YaSSL client and server
Attachment: callgrind.out.14693 (application/octet-stream, text), 389.11 KiB.
[4 Aug 2009 16:57]
Andrew Hutchings
Callgrind for OpenSSL client when using YaSSL server
Attachment: callgrind.out.14786 (application/octet-stream, text), 295.74 KiB.
[4 Aug 2009 17:02]
Andrew Hutchings
Running the last 2 attachments through kcachegrind shows that most of the time is taken with TaoCrypt's RecursiveMultiply
[7 Aug 2009 8:33]
Sergei Golubchik
Andrew, could you submit a bug report to yassl ? I suppose the only thing *we* can do is to build with openssl when possible.
[9 Aug 2009 7:09]
Andrew Hutchings
My initial benchmarks show that CyaSSL is many times faster (RSA decryption is nearly 10x faster than YaSSL). If licensing supports this I suggest looking into moving to this. Especially since this would help with the C++/C issues with APIs.
[16 Mar 2012 9:27]
MySQL Verification Team
http://www.yassl.com/yaSSL/Blog/Entries/2011/4/28_MySQL_SSL_Performance_Benchmarking.html
[25 Jul 2019 11:20]
Erlend Dahl
We will remove the support for YaSSL in an upcoming release.