Thanks Bourg. >> Even if Chimera doesn't implement a JCE provider due to its impractical >> nature, do you think it would be possible to reuse the standard Java types >> (Cipher, SecureRandom) instead or redefining them? The implementation could >> be instantiated directly (Cipher cipher = new OpensslCipher()) instead of >> calling Cipher.getInstance(). JCE architecture is quite heavy weight for such abstractions. In JCE design rational, the JCE Cipher class is not designed to be inherited or implemented by others. It's implementation class for final use and wraps to the under layer SPI mechanisms.
>> Another question, will there still be an incentive to use Chimera once Java >> 9 with the AES improvements is widely used? Should we consider Chimera as a >> temporary solution to the current Java limitations? There are some still some improvements will not happen in JDK 9 release. For example, Openssl GCM will still 4x - 6x faster. Of source optimization is one aspect, the stream or io classes build on it are also useful. Also consider that the release JDK9 will not available in some time 2017. The wider adoption of JDK9 will be much later than that time I guess. Thanks, Haifeng -----Original Message----- From: Emmanuel Bourg [mailto:ebo...@apache.org] Sent: Friday, February 26, 2016 11:03 PM To: Commons Developers List <dev@commons.apache.org> Cc: common-...@hadoop.apache.org Subject: Re: [crypto][chimera] Next steps Le 24/02/2016 03:41, Chen, Haifeng a écrit : > Hope this information helps. Thanks Bourg for these insightful questions. It does, thank you for the insights. Even if Chimera doesn't implement a JCE provider due to its impractical nature, do you think it would be possible to reuse the standard Java types (Cipher, SecureRandom) instead or redefining them? The implementation could be instantiated directly (Cipher cipher = new OpensslCipher()) instead of calling Cipher.getInstance(). Another question, will there still be an incentive to use Chimera once Java 9 with the AES improvements is widely used? Should we consider Chimera as a temporary solution to the current Java limitations? Emmanuel Bourg --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
