Hypertable was designed to solve the horizontal scalability problem. If you have more data than can fit onto a single machine, or you need resources from more than one machine to efficiently serve your data, then Hypertable is one possible solution. Hypertable differs from Cassandra in that Hypertable is based on the
Bigtable design, whereas Cassandra is based on the
Dynamo design. The primary difference between these two designs is that Bigtable keeps data physically ordered by a primary key chosen by the application and is consistent, whereas Dynamo is a distributed hash table (DHT) an employs eventual consistency. Because Bigtable keeps data physically sorted by a meaningful key, it is better suited to applications that require efficient range scans (analytics, etc.). If all you need is efficient single-key lookups and you're not too concerned about consistency, then Cassandra may be a good choice.
HBase and Hypertable can be used to solve the same problems.