Got many web applications I wish to bind into a network for a better
user experience. The “User” model would then be shared so that any
registered user may enjoy all websites without setting the same profile
and password.
Here are the design options:
One rails app serving the user model to other websites using active
resource: I don’t like very much this option since I’m afraid it would
be too slow…
Running an LDAP server. Typical rails attributes (created_at,
updated_at) and typecasting may be an issue or am I wrong ?
OpenID: I like this technology and its principles but I’m afraid
migrating tens of thousands accounts from different websites and without
altering their passwords would be a pain in the… for the development
team as for the user who maybe won’t apreciate this new authentication
process with back and forth redirection between the openid server and
the web application…
One rails app serving the user model to other websites using active
resource: I don’t like very much this option since I’m afraid it would
be too slow…
What we’ve done is have one rails app serving the user models, however
the client rails apps then cache the data locally, so no speed concerns.
Updates are relatively infrequent so we can be very stupid about
invalidating the cache.
Can you provide more details on how you created a rails app that
serves the user model to other apps? I’m investigating this option
right now and I can’t find anything about the pros/cons or even how to
implement it. I know Active Resource is the key but how do you cache
all the session data in the local app? Any help or direction would be
appreciated. I don’t want to go too far down this path if it isn’t
going to work out.
Can you provide more details on how you created a rails app that
serves the user model to other apps? I’m investigating this option
right now and I can’t find anything about the pros/cons or even how to
implement it. I know Active Resource is the key but how do you cache
all the session data in the local app? Any help or direction would be
appreciated. I don’t want to go too far down this path if it isn’t
going to work out.
We’ve got a table that stores what we’ve got from the remote app.
Another row in another table states whether the cache is valid, when
someone edits information in the app that owns the user data it pings
all the other apps to tell them they should dump their cache (this
just flips a flag in the cache status table). Next time the app needs
to get some data it checks whether the cache is valid and if not
copies the data over again.
I suppose a key thing is the frequency of updates. It’s not that
frequent for us, so we can be very clumsy with our cache: changing one
user triggers a refresh of the info for all users. This may or may not
work for you.
We started out with activeresource (or rather a bastardisation of it
on the client side since we were running rails 1.2 at the time), but
it turned out that for the amount of data we were moving it was a bit
slow (i can’t quite remember the timings, perhaps 2-3 seconds) and the
slow bit was parsing the xml. We now just dump attribues with
Marshal.dump and send that down the wire.
Fred
This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.
