Bug 1021822 - Improve performance of ENCRYPT protocol
Summary: Improve performance of ENCRYPT protocol
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: JBoss Data Grid 6
Classification: JBoss
Component: JGroups
Version: 6.2.0
Hardware: Unspecified
OS: Unspecified
unspecified
medium
Target Milestone: ---
: ---
Assignee: Tristan Tarrant
QA Contact: Martin Gencur
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2013-10-22 07:31 UTC by Martin Gencur
Modified: 2015-03-25 12:25 UTC (History)
4 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2015-03-25 12:25:45 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Issue Tracker JGRP-1722 0 Major Resolved Improve performance of ENCRYPT protocol 2014-07-11 11:17:50 UTC

Description Martin Gencur 2013-10-22 07:31:51 UTC
See description in the JIRA issue.

Comment 2 JBoss JIRA Server 2013-10-24 05:36:27 UTC
Bela Ban <bela> made a comment on jira JGRP-1722

See if ENCRYPT could be moved *down* the stack, e.g. just below the reliable retransmission protocols ({{NAKACK}, {{UNICAST}}). I seem to recall though that ENCRYPT has a dep on GMS.

Comment 3 JBoss JIRA Server 2013-10-24 05:36:52 UTC
Bela Ban <bela> made a comment on jira JGRP-1722

See if ENCRYPT could be moved *down* the stack, e.g. just below the reliable retransmission protocols (NAKACK, UNICAST). I seem to recall though that ENCRYPT has a dep on GMS.

Comment 4 JBoss JIRA Server 2013-10-24 06:21:22 UTC
Martin Gencur <mgencur> made a comment on jira JGRP-1722

The replication works also with ENCRYPT placed under NAKACK while AUTH is placed under GMS. My functional test for the ISPN server passes with this configuration. Data is successfully replicated.

Comment 5 JBoss JIRA Server 2013-10-24 08:25:54 UTC
Bela Ban <bela> made a comment on jira JGRP-1722

I think this is only true if (a) there is no communication between joiner and coordinator needed or (b) the transport is reliable.
* (a) If we have a shared keystore or username/password encryption, then this works
* (b) If we need to send a key request to the coord, and the coord sends a response, then either the transport needs to be reliable or UNICAST needs to be below ENCRYPT, or else these requests are sent *unreliably* and might get lost. A possible solution for (b) could be to assume that the message sending is unreliable and send the key request to the coordinator until a valid key is received. In this case, ENCRYPT could be moved under NAKACK/UNICAST.

[~mgencur] Either you used TCP as transport, or you were just lucky that the message wasn't dropped. Under real stress, there are no guarantees that a message isn't dropped.

Comment 6 JBoss JIRA Server 2014-03-07 16:45:59 UTC
Bela Ban <bela> updated the status of jira JGRP-1722 to Resolved

Comment 7 Martin Gencur 2014-12-03 14:19:45 UTC
This needs further analysis. Moving target to 6.5. This will be revisited after 6.4


Note You need to log in before you can comment on or make changes to this bug.