couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Antony Blakey <antony.bla...@gmail.com>
Subject Re: Replication and forms of weak consistency
Date Mon, 16 Feb 2009 04:48:52 GMT

On 16/02/2009, at 2:55 PM, Chris Anderson wrote:

> Maybe someone could tell the story of
> what happens if you just use CouchDB, with it's weak consistency

'weak consistency' is a class of guarantees, not a guarantee per se,  
so CouchDB doesn't 'have' weak consistency. In a replicated (as  
opposed to local) context, I think it provides virtually no  
consistency guarantees.

> , and
> don't bother even thinking about these issues. Is the naive approach
> doomed? Does it suffice to tell the developers who have more relaxed
> demands, "keep your documents independent - eg don't use multi-doc
> transactions" or can consistency issues bite back in the single-doc
> case as well?

It would suffice to say "don't presume that just because you update A  
before you update B your code will see A if it can see B". This is the  
essence of not having Monotonic Writes. It's not about transactions at  
all. Isolated write groups are something that would a) optimise  
replication; and b) be useful to application developers (especially  
with a fail-on-commit option), but they aren't transactions in the  
traditional sense.

The current replication system is fragile for single-doc cases,  
because it depends on replication succeeding i.e. AFAICT there's no  
guarantee that a given document will ever replicate, unless  
replication completes. With Monotonic Writes, the anti-entropy nature  
of replication will guarantees that document updates will be seen as  
replication makes incremental progress. But I've still not analysed  
the replication code, so I don't know whether document replication is  
FIFO or not.

I think a Monotonic Writes guarantee makes applications easier to  
write, and given that it can be provided, why not do it? By adopting a  
formal model, it means we can draw on existing research, and by  
adopting a formal conceptual framework, we can mnore fruitfully  
discuss these issues.

Antony Blakey
-------------
CTO, Linkuistics Pty Ltd
Ph: 0438 840 787

One should respect public opinion insofar as is necessary to avoid  
starvation and keep out of prison, but anything that goes beyond this  
is voluntary submission to an unnecessary tyranny.
   -- Bertrand Russell



Mime
View raw message