Canonical's decision to move away from CouchDB

[Riyad Kalla <rk...@gmail.com>]

REF:
http://www.h-online.com/open/news/item/Canonical-dropping-CouchDB-from-Ubuntu-One-1382809.html

> Lenton says Canonical has worked with "the company behind CouchDB" to
make the database scale
> in a particular way. "Our situation is rather unique and we were unable
to resolve some of the issues
> we came across," said Lenton, who pointed out they needed to scale to
millions of users and scale
> down to a "reasonable load on small client machines".

Does anyone happen to know exactly what was the blocker(s) for Canonical?
e.g. client load time for small mobile devices?

Re: Canonical's decision to move away from CouchDB

[Randall Leeds <ra...@gmail.com>]

On Tue, Nov 22, 2011 at 18:31, Alex Besogonov <al...@gmail.com> wrote:
> I had the problems with Erlang runtime on mobile devices. Specifically, on iOs
> devices. So I started my SofaDB project to replicate the CouchDB replication
> protocol in a simple C++ library.
>
> Consequently, I found that the protocol for conflict resolution is not
> documented anywhere, including the source code :)

I'm most of the way through a JavaScript implementation for PouchDB
(https://github.com/mikea/pouchdb). I'll make sure to comment in
thoroughly. Also, I'll make sure to put some of my attention to
getting that in a wiki whenever we decide to hold our first wiki
sprint (which I'll organize shortly).

>
> On Tue, Nov 22, 2011 at 11:46 AM, Riyad Kalla <rk...@gmail.com> wrote:
>> REF:
>> http://www.h-online.com/open/news/item/Canonical-dropping-CouchDB-from-Ubuntu-One-1382809.html
>>
>>> Lenton says Canonical has worked with "the company behind CouchDB" to
>> make the database scale
>>> in a particular way. "Our situation is rather unique and we were unable
>> to resolve some of the issues
>>> we came across," said Lenton, who pointed out they needed to scale to
>> millions of users and scale
>>> down to a "reasonable load on small client machines".
>>
>> Does anyone happen to know exactly what was the blocker(s) for Canonical?
>> e.g. client load time for small mobile devices?
>>
>

I'd be really interested to hear more from Canonical. I'm disappointed
if they dropped it already. It's only getting much better and they
already put significant work into tying it into their desktop
platform. Finally, I'm not sure what would be a good replacement for
their use case.

Re: Canonical's decision to move away from CouchDB

[Alex Besogonov <al...@gmail.com>]

I had the problems with Erlang runtime on mobile devices. Specifically, on iOs
devices. So I started my SofaDB project to replicate the CouchDB replication
protocol in a simple C++ library.

Consequently, I found that the protocol for conflict resolution is not
documented anywhere, including the source code :)

On Tue, Nov 22, 2011 at 11:46 AM, Riyad Kalla <rk...@gmail.com> wrote:
> REF:
> http://www.h-online.com/open/news/item/Canonical-dropping-CouchDB-from-Ubuntu-One-1382809.html
>
>> Lenton says Canonical has worked with "the company behind CouchDB" to
> make the database scale
>> in a particular way. "Our situation is rather unique and we were unable
> to resolve some of the issues
>> we came across," said Lenton, who pointed out they needed to scale to
> millions of users and scale
>> down to a "reasonable load on small client machines".
>
> Does anyone happen to know exactly what was the blocker(s) for Canonical?
> e.g. client load time for small mobile devices?
>