What’s New in the 2017 DB‑Engines Rankings? Latest Database Releases and Features

RDBMS Family

Oracle Autonomous Database (18c) : Oracle released the 12.2.0.1 (12cR2) build, marketed as 18c, which introduces a fully self‑optimising, self‑managing, self‑repairing cloud database service.

MySQL 5.7.20 : Released 2017‑10‑16. Key changes include enhanced audit logging, deprecation of tx_isolation and tx_read_only in favour of transaction‑isolation and transaction‑read‑only, new NO_INIT=true data‑directory initialisation flag, Docker image loss bug fix, security hardening, X Plugin update, and replication improvements. Release notes: https://dev.mysql.com/doc/relnotes/mysql/5.7/en/news-5-7-20.html

MariaDB 10.2.9 : Adds the alpha‑stage MyRocks storage engine, window functions, recursive CTEs (available from 10.2.2), updates InnoDB to 5.7.18 and TokuDB to 5.6.37‑82.2. Release notes: https://mariadb.com/kb/en/mariadb/mariadb-1026-release-notes/

SQL Server 2017 : First simultaneous Windows and Linux release. Supports Docker, Kubernetes and OpenShift. Introduces Read‑scale Availability Groups, machine‑learning extensions for R and Python, and a native Linux Docker engine.

DB2 V11.1.1.2 : Improves crash recovery (faster pureScale node recovery, allows connections during recovery, faster transaction rollback) and adds richer JSON functions.

PostgreSQL 10 : Highlights include parallel query execution, logical replication/subscription, FDW push‑down, sharding enhancements, and SCRAM‑SHA‑256 authentication.

Greenplum 5.1 : GPORCA optimizer gains better short‑query performance, supports partitioned‑table index sub‑nodes, adds external data engine PXF, and introduces SHA‑256 data verification for gptransfer.

NoSQL Family

MongoDB 3.6 (December 2017 release) : Provides Change Streams (Oplog‑based pub/sub), retryable writes, JSON Schema validation, expressive array updates, and $lookup for richer joins. MongoDB also celebrated its Nasdaq listing.

Redis Cluster : Recent advances (summarised by the Redis China User Group) include four major integration improvements for clustering, high‑availability, and scaling.

ArangoDB 3.3 (milestone) : Introduces a beta replication engine with a redesigned protocol, automatic failover, and a global asynchronous replication API. Repository: https://github.com/arangodb/arangodb

NewSQL Family

TiDB 1.0 (GA) : GA release on 2017‑10‑16. Enhances MySQL compatibility, SQL optimizer, stability and performance. Integrated with Tencent Cloud and UCloud; source code at https://github.com/pingcap/tidb.

CockroachDB 1.1 : Adds fast CSV data import, cluster‑wide SHOW QUERIES and CANCEL QUERIES for operational management.

Time‑Series

InfluxDB 1.3.6 : Go‑based distributed time‑series database with horizontal scaling. Core features: time‑series functions (max, min, sum, etc.), real‑time metrics aggregation, and generic event handling. Download portal: https://portal.influxdata.com/downloads

Big Data Ecosystem

Hadoop 3.0 beta : First beta of the 3.0 line (released 2017‑10‑03). Fixed 576 bugs, performance improvements over 2.8.x. Not recommended for production until GA; see https://hadoop.apache.org/docs/r3.0.0-beta1/hadoop-project-dist/hadoop-common/release/3.0.0-beta1/CHANGES.3.0.0-beta1.html for full changelog.

Druid 0.10.1 : Over 100 enhancements including TopN performance boost, push‑down LIMIT for Group‑By, precise query timeout handling, Amazon S3A support, Protobuf 3.0 ingestion, HTTP firehose, improved disk‑space management for real‑time indexing, new Kafka metric emitter, column‑value filter, and extensive SQL improvements.

SnappyData 1.0.0 : 100 % compatible with Spark 2.1‑2.2, supports all Spark ML/MLlib algorithms, automatic Dataset conversion, and provides Stream Table structures for real‑time analytics.

Domestic Databases Overview

DM (DaMeng) V7.1.6.3 : Java‑VM‑based high‑performance RDBMS with improved Oracle compatibility, performance optimisations, and bug fixes.

OceanBase 1.4.51 : Deployed in major Chinese banks and Alipay. Adds three‑zone‑five‑center HA architecture, read‑only replicas, checksum protocol, synonym support, and recycle‑bin for accidental object removal.

SequoiaDB v2.8.3 Enterprise : Financial‑grade distributed DB supporting OLTP, NoSQL, and object storage. Integrated with Spark for analytics and BI pipelines; recognised as a Spark certified distributor.