Checkpoints

Not Supported

nessDB does not support recovery and checkpoint.

Compression

Naïve (Page-Level)

nessDB uses QuickLZ, a compression library written in C, to compress the key-value pair when storing the pair into tree nodes.

Concurrency Control

Not Supported

There are two mutexes for the meta-data and SST (Small-Splittable Tree), and no tuple level latch. A thread must get the latch before merging the SST nodes.

Data Model

Key/Value

nessDB is a key-value DBMS, which supports Redis-like protocol, such as SET, GET, DEL, EXISTS. To be specific, both key and value are general raw data and are represented by a pointer with length.

Foreign Keys

Not Supported

nessDB is a key-value DBMS so it does not support foreign key.

Indexes

Log-Structured Merge Tree

nessDB used LSM-tree (Log-Structured Merge Tree) in v1.8, Small-Splittable Tree in v2.0. Those data structures are suitable for high-performance insertion compared to B-tree.

Isolation Levels

Not Supported

nessDB uses a single mutex for the whole SST structure instead of tuple-level latches.

Joins

Not Supported

nessDB does not support join. It is a key-value DBMS with only interfaces such as SET, GET, DEL, EXISTS.

Query Execution

Tuple-at-a-Time Model

nessDB is a key-value DBMS and each query only involves a single key-value pair.

Query Interface

Custom API

nessDB is an embedded DBMS with custom C API, as well as wrappers for Java, Python and Lua.

Storage Architecture

Disk-oriented

nessDB is a disk-oriented DBMS. All data are stored in non-volatile storage after shutdown.

Storage Model

N-ary Storage Model (Row/Record)

nessDB uses a Small-Splittable Tree (SST), a B-Epsilon-tree with small height (level), for storage engine. SST has multiple levels with exponential expanding size. For level 0, the data is un-sorted and the other levels are sorted. This data structure is designed for random write workloads.

Storage Organization

Heaps

For a tree node file, there is a header which stored count of the key-value pair in each level. The header is followed by several var-length key-value pairs with data offset. The following diagram shows the structure of a tree node file, and the first line represents the header part.

| L0 count | L1 count | L2 count | L3 count |
| Key 0 length | key 0 data | data offset | operation flag|
| Key 1 length | key 1 data | data offset | operation flag|
...
| Key N length | key N data | data offset | operation flag|

Stored Procedures

Not Supported

System Architecture

Embedded

Views

Not Supported