Re: Announcing Newt DB, the amphibious database
Jim Fulton <[email protected]>
| Newsgroups | gmane.comp.web.zope.zodb |
|---|---|
| Message-ID | <CAPDm-FgJjJDecPJKcsLaX8zDouUk-gvLvAoRNQLAquRvHm=Dng@mail.gmail.com> |
On Wed, Feb 22, 2017 at 3:20 PM, Mike Metcalfe <[email protected]> wrote: > > > On Tuesday, 24 January 2017 20:35:04 UTC+2, Jim Fulton wrote: >> >> Newt addresses some of ZODB's biggest issues: >> >> - Data are visible and queryable in Postgres and usable by non-Python >> applications. >> >> >> - Postgres provides powerful search and indexing capabilities. >> >> >> - Because indexing is done by Postgres: >> - You don't need to have catalogs and their index structures in >> your application. For some applications, this can reduce the number of >> objects by an order of magnitude. >> - Conflict errors become much less frequent, because the most >> common source of conflicts, catalog indexes and supporting int-id managers, >> are no more. >> >> The only time I've had problems with ZODB is in a high-write environment. > Can newt help? > Maybe. If you were using catalogs and used Postgres indexes instead, you could end up writing a lot less, and have far fewer conflicts, which would help. Have you tried RelStorage with Postgres? I'm guessing that there are opportunities to scale Postgres a lot farther than ZEO. Newt can slow writes down, if you have a lot of indexes. :) Although if you're replacing catalogs with Postgres indexes it's probably a big net win. You can also mitigate index burden on writes by using the asynchronous JSON converter: http://www.newtdb.org/en/latest/topics/updater.html Jim -- Jim Fulton http://jimfulton.info -- You received this message because you are subscribed to the Google Groups "zodb" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.