>From my point of view the biggest difference between the two systems is the
>following: Cassandra can always accept writes. It uses its own local storage
>(no dependence on HDFS or anything like that) and P2P data replication. In
>contrast, HBase depends on HDFS so is unavailable if the filesystem has a
>problem, and furthermore if regions are in transition due to node failure or
>load balancing then writes to the affected region are temporarily blocked
>until it is available again. In exchange for always accepting writes,
>depending on what fails where and when, reads in Cassandra may only retrieve
>stale data, or the reads of some clients can be inconsistent with reads of
>others. Eventually all reads will return the last written value. In contrast,
>HBase and Hypertable and Google's implementation are strongly consistent
>systems. A read always returns the last written value.
In the proposed HBase replication system (HBASE-1295), each peer cluster is
strongly consistent but edits will be asynchronously propagated between them --
we call this regional consistency.
- Andy
On Tue, Sep 1, 2009 at 11:45 AM, charles du <[email protected]> wrote:
> Hi:
>
> Does anyone have experience with both Cassandra and HBase? To me, they
> target at a similar problem. I am wondering what are main differences
> between these two, like reliablity/performance/features?
>
> Thanks.
>
> --
> tp
>
