[
https://issues.apache.org/jira/browse/HDFS-6606?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14115547#comment-14115547
]
Alejandro Abdelnur commented on HDFS-6606:
------------------------------------------
Yi, the approach LGTM. I'll ping [~atm] in case I'm missing something.
My feedback on the patch:
*CryptoInputStream.java*:
* read(BB), the {{if (n >- 0) THEN ELSE}} section at the end, why do we need
this now and not before? or it is a bug in CIS?
*DataTransferSaslUtils.java*:
* requestedQopContainsPrivacy(), should’t we trim() the values in the set?
just in case somebody added a whitespace, or this is not possible?
* sendSaslMessageAnNegotiationCipherOptions(), can payload be NULL? or that is
an error if we get here? if the later we should throw an exception.
* createStreamPair(), we are using the same key and IV for both streams. we
should either use 2 different KEYs or 2 different IVs, not to repeat the use of
a KEY/IV pair. or simply exchanging 2 cipheroptions, one for IN and one for OUT.
* the use wrapping/unwrapping seems too asymmetrical how is implemented, is it
possible to have 2 symmetric methods used at the same level on both sides?
*hdfs.proto*:
* ChiperOption, I don’t think the fields should be optional, they are all
required, no?
*SaslDataTransferServer.java*:
* why do we need 2 confs?, what was wrong with the original one?
> Optimize HDFS Encrypted Transport performance
> ---------------------------------------------
>
> Key: HDFS-6606
> URL: https://issues.apache.org/jira/browse/HDFS-6606
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: datanode, hdfs-client, security
> Reporter: Yi Liu
> Assignee: Yi Liu
> Attachments: HDFS-6606.001.patch, HDFS-6606.002.patch,
> OptimizeHdfsEncryptedTransportperformance.pdf
>
>
> In HDFS-3637, [~atm] added support for encrypting the DataTransferProtocol,
> it was a great work.
> It utilizes SASL {{Digest-MD5}} mechanism (use Qop: auth-conf), it supports
> three security strength:
> * high 3des or rc4 (128bits)
> * medium des or rc4(56bits)
> * low rc4(40bits)
> 3des and rc4 are slow, only *tens of MB/s*,
> http://www.javamex.com/tutorials/cryptography/ciphers.shtml
> http://www.cs.wustl.edu/~jain/cse567-06/ftp/encryption_perf/
> I will give more detailed performance data in future. Absolutely it’s
> bottleneck and will vastly affect the end to end performance.
> AES(Advanced Encryption Standard) is recommended as a replacement of DES,
> it’s more secure; with AES-NI support, the throughput can reach nearly
> *2GB/s*, it won’t be the bottleneck any more, AES and CryptoCodec work is
> supported in HADOOP-10150, HADOOP-10603 and HADOOP-10693 (We may need to add
> a new mode support for AES).
> This JIRA will use AES with AES-NI support as encryption algorithm for
> DataTransferProtocol.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
