Release of grapho-oriented DBMS Nebula Graph 3.2

published Open DBMS Nebula Graph 3.2 , designed to effectively store large sets of interconnected data forming a graph that can include billions of nodes and trillion ties. The project is written in C ++ and spreads under the Apache 2.0 license. Client libraries for contacting the DBMS are prepared for the GO, Python and Java.

DAM is used Distributed architecture without separation of resources (shared-nothing), which implies the launch of independent and self-sufficient processing processes of Graphd requests and storaged storage processes. The meta-service is engaged in the orchestration of data movement and the provision of meta-information about the column. To ensure the consistency of data, a protocol is used on the basis of the algorithm raft .

.

The main features of Nebula Graph:

• Ensuring security through providing access to only authenticated users whose powers are set Through the system control system based on roles (RBAC).
• possibility connecting different types of storage motors. Supporting the expansion of the language formation language with new algorithms.
• Ensuring the minimum delays in reading or recording data and maintaining high throughput. At testing in a cluster of one Graphd node and three storaged units A 632 GB database, including a graph of 1.2 billion peaks and 8.4 billion rib delay were at a several milliseconds level, and the throughput amounted to 140 thousand requests per second.
• linear scalability.
• SQL-like query language, sufficiently powerful and easy to perceive. Operations such as GO (bidirectional bypass of the top of the count), Group by, Order by, Limit, Union, Union Distinct, Intersect, Minus, Pipe (using the result from the previous request) are supported. Indices and variables determined by the user are supported.
• Ensuring high availability and disruption resistance.
• Support for creating snapshots with a cut of the state of the database to simplify the creation of backup copies.
• readiness for industrial use (is already used in the infrastructure of JD, Meituan and Xiaohongshu).
• The ability to change the storage and update scheme without stopping or influence on the operations performed.
• TTL support to limit the time of life.
• commands to control settings and storage hosts.
• tools for managing work and planning work (work is still supported from work).
• Operations for searching for the full path and the shortest path between the given peaks.
• OLAP interface for integration with third -party analytics platforms.
• Data imports from CSV or Spark.