[ 
https://issues.apache.org/jira/browse/CASSANDRA-4482?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13427347#comment-13427347
 ] 

Jonathan Ellis edited comment on CASSANDRA-4482 at 8/2/12 2:34 PM:
-------------------------------------------------------------------

This is how the original Dynamo paper describes maintaining merkle trees.  The 
problem that Acunu doesn't mention is that this forces you to do read and 
rehash all the rows sharing the tree leaf with row X, whenever any row X is 
updated.  So you are trading sequential i/o for random i/o... not a good move, 
unless you assume SSD or a small dataset (and even then, you're rehashing many 
rows on each update, not just one, so it's far from clear that this is a good 
trade).

I'm a bigger fan of the continuous repair approach enabled by CASSANDRA-3912, 
and discussed in CASSANDRA-2699 (although i think 2699 overcomplicates things).
                
      was (Author: jbellis):
    This is how the original Dynamo paper describes maintaining merkle trees.  
The problem that Acunu doesn't mention is that this forces you to do read and 
rehash all the rows sharing the tree leaf with row X, whenever any row X is 
updated.  So you are trading sequential i/o for random i/o... not a good move, 
unless you assume SSD or a small dataset (and even then, you're rehashing many 
rows on each update, not just one).

I'm a bigger fan of the continuous repair approach enabled by CASSANDRA-3912, 
and discussed in CASSANDRA-2699 (although i think 2699 overcomplicates things).
                  
> In-memory merkle trees for repair
> ---------------------------------
>
>                 Key: CASSANDRA-4482
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-4482
>             Project: Cassandra
>          Issue Type: New Feature
>            Reporter: Marcus Eriksson
>
> this sounds cool, we should reimplement it in the open source cassandra;
> http://www.acunu.com/2/post/2012/07/incremental-repair.html

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to