--
You received this message because you are subscribed to the Google Groups "Redis DB" group.
To post to this group, send email to redi...@googlegroups.com.
To unsubscribe from this group, send email to redis-db+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/redis-db?hl=en.
> Any other comments? redis is not ideal for the scenerio I outlined? maybe
> someone can tell me why so I can understand? :)
For sure Redis *can* be used to implement something like Disqus, but
can does not mean must have.
Depends on many parameters. Is a comment system DB-bound? Or given
that you can cache the whole stuff until there is no a new comment
added this is a non issue so it's better to use another on disk store
that is perfect for many reads and seldon writes?
What is the amount of comments?
And so forth. But for sure you can use Redis efficiently for such a task.
I could do this: store all the comments for a given post as a single
hash if the total messages are <= a given threshold (the majority).
This is very space efficient in Redis.
Switch to a different graph-alike representation just for uncommonly
long comments.
Five minutes idea. People that design this systems for work know better :)
Cheers,
Salvatore
--
Salvatore 'antirez' Sanfilippo
http://invece.org
"We are what we repeatedly do. Excellence, therefore, is not an act,
but a habit." -- Aristotele
> Switch to a different graph-alike representation just for uncommonly
> long comments.
I mean uncommonly long threads, not comments. Sorry for the mistake.
Cheers,
Salvatore
do you mind if I add Disqus as "testimonial" in the new Redis site ;)
Linking the Disqus logo to the analytics page in this page:
http://redis.io/topics/whos-using-redis
Cheers,
Salvatore
--
Cool thank you :)
Salvatore
I completely agree for being conservative and don't do switches if
it's not worth it. Also memcached does very well what it advertises
doing.
But I think there is a reason for a big percentage of memcached users
to consider switching to Redis, that is, you no longer cache and
invalidate, but cache and *modify* your old value as the DB side
changes, without the need of recreating it from scratch. So for
caching HTML fragments this is not an issue, but for caching
everything more complex, like objects (use Hashes), timelines (use
Lists), and so forth, Redis can completely shift the paradigm, and
remove a lot of load from the DB. Persistence is another interesting
feature for many users using Redis just like a cache. On the other
side memcached benefits from being more specialized, for instance it
uses less data for caching very small strings, from what I saw in some
test I did time ago.
Cheers,
--
You received this message because you are subscribed to the Google Groups "Redis DB" group.
To post to this group, send email to redi...@googlegroups.com.
To unsubscribe from this group, send email to redis-db+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/redis-db?hl=en.