|
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
|
Re: [PhillyOnRails] Amazon's SimpleDB, let's talk about it
|
:) It'll definitely be a boon for pre-funded startups who have those
sorts of data needs and don't have the cash to build out their own
hadoop cluster. And I'm sure the more frugal developers will figure out
some way to swap out data into s3 for archiving and longer term
storage. But yeah, 10G per domain and no way to map/reduce over the
data is pretty weak.
Toby DiPasquale wrote:
On Fri, Dec 14, 2007 at 02:41:07PM -0800, Cliff Moon wrote:
Lots of similarities to bigtable/hbase. It certainly won't replace an
RDBMS or even couch in your architecture. Particularly because of the
eventual consistency limitations. If your application called for
semi-structured data, frequent reads, and only batch updates then it'd
come in real handy.
Boy, that sounds just like this application I keep hearing about... what
is it called? Ah, that was the name of it: Powerset. I guess that's why
Jim and boys in Powerlabs are busy building HBase for Yahoo! and the rest
of the Hadoop userbase.
You're right, of course, Cliff, but SimpleDB is the weakest of all three
in that area given its size limitations and expense at this time. However,
those are things Amazon can handle easily with time so I'm hopeful it can
become a real contender for CouchDB. It will never be a threat to HBase
unless they allow arbitrary queries via a framework like map/reduce.
_______________________________________________
To unsubscribe or change your settings, visit:
http://lists.phillyonrails.org/mailman/listinfo/talk
|
|