I may be wrong but I didn't think CouchDB was all that hard to install and maintain. Do you have some specific concerns about deployment and maintenance?
I definitely feel you about the self joins. If I had had something like CouchDB availableto me about 2 years ago a number of different projects would have been much much simpler. On Wed, Sep 10, 2008 at 9:39 PM, Brad King <[EMAIL PROTECTED]> wrote: > This is more of a business related post than technical, so bear with > me. I love how couchdb works: its elegant, it works how I think, it > can accept any reasonably structured data. My decision to make is > whether or not to take our company onto this new ground for a mission > critical system: customer product data. Technically we call it > "inventory" but I plan to keep all of the quantity control in > Postgres, so its really just product data. We have some really big > customers, with hundreds of thousands of products. It has to work, it > can't fail. This data feeds everything else in our system. > > I have a vision for this great marriage of Postgres for inventory > quantity managment and CouchDB for customer specific product data > catalogs, but I'm struggling with what I think our Operations team can > realistically manage for keeping these CouchDB systems up, running , > backed up, and easily deploying new servers for new customers. > > Is anyone else in the SaaS business taking the plunge here with > CouchDB. Am I insane for even considering it? I'm just tired of > writing self joins against unindexed fields in sql to handle > variations in product schema per customer, it just doesn't make sense. >
