db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "David W. Van Couvering" <David.Vancouver...@Sun.COM>
Subject Re: in-memory
Date Tue, 11 Oct 2005 22:47:39 GMT
Sorry, what I should have said was that with the relaxed durability 
approach, we really need a mode where if there is a crash you can reboot 
the db in a consistent state, which as I understand it doesn't exist 
today.  If you are a pure in-memory solution, you have automatic 
"consistency" because your data disappears if the VM goes down. :)

David

Daniel John Debrunner wrote:
> David W. Van Couvering wrote:
> 
> 
>>I have to agree, we need a more useful "memory-based" solution that
>>guarantees consistency if not durability.
> 
> 
> So just to repeat myself, the current relaxed durability matches a
> memory based solution in terms of consistency, ie. no problems.
> 
> Durability is similar for both though with the current disk-based scheme
> and relaxed durability there are cases where the database will exist
> after a re-boot, where with the in-memory solution the data will be gone.
> 
> Not saying an in-memory solution might not be useful, but the relaxed
> durability will give similar performance benefits since the data will be
> cached in memory, either within Derby or within the OS filesystem.
> 
> Dan.
> 
> 

Mime
View raw message