Jaguar -- Fastest NoSQL Database for Enterprises

Jaguar is a scalable, fast NoSQL database for big data. It uses an 
innovative array-indexing technology in data access path. Compared to traditional 
B+Tree in relational database software, Jaguar employs plain arrays in data indexing.
Arrays have excellent data locality, read-ahead, and caching properties. The core
storage and indexing engine is 10X - 50X faster than B+Tree, therefore table join 
and range search are extremely fast.

Jaguar uses the simplest data model: array. All data items are placed in a huge
flat sparse array sorted by its primary keys. Also secondary-indexes are maintained as data 
is updated in the data array. The indexes are also stored in arrays.

Even in a complex distributed enviroment, data is still logically presented as an array
to the user. From the perspective of users or programmers, a table is just an array of
sorted data items, and so is an index. Physically the table is striped over a number of
hosts in the cluster. No other data platforms can allow programmers to directly access 
the underlying data structure.