Overview

ApsaraDB for HBase is a highly optimized NoSQL database (enterprise edition is available as Lindorm) that is compatible with the community edition of HBase. Thanks to this compatibility and integration with Spark, Phoenix, and Solr, ApsaraDB for HBase is easy to use and offers superior stability and cost performance. ApsaraDB for HBase can easily support high throughput and high concurrency scenarios, providing support for real-time volumetric data storage, full-text indexing, lightweight SQL queries, as well as time space and series queries. Compared with its open-source counterpart, ApsaraDB for HBase is better optimized, all the way down to the kernel, with superior read/write performance, disaster recovery capabilities, storage efficiency, and response latency. In terms of the numbers, read/write performance is marked by a 3 to 7 times improvement, RPO is reduced to less than 1 minute, 99% percentile latency has been reduced by 90%, MTTR is reduced by 90%, and the compression ratio has increased 13 fold. ApsaraDB for HBase guarantees a service uptime of 99.9%. ApsaraDB for HBase is suitable for high-demand industry applications like risk control, recommendation, advertising, IoT, VoT, feed streaming, and data visualization scenarios. Internally at Alibaba, ApsaraDB for HBase has already provided support for several of Alibaba Group’s core businesses, including Taobao, Alipay, and Cainiao.

Benefits

• High Storage Reliability Built with a distributed cluster architecture with six data backups and at least three replicas to guarantee a data reliability of 99.99999999%.
• High Availability Real-time availability monitoring and single-point failover are supported to guarantee the continuity of your workloads. Service process monitoring and automatic recovery can help you recover processes within a few seconds.
• High-Efficiency Operations and Maintenance ApsaraDB for HBase is equipped with a unified platform for visualized database management, monitoring, and alerting. Its kernel has been automatically upgraded for high-efficiency operations and maintenance, and the corresponding web console, API, and multi-language SDKs were designed for the easy cluster management.
• Significantly Improved Performance Developed as an improved version of the community edition of HBase. Alibaba Cloud has significantly optimized the kernel, allowing for read and write performance to be increased by 3 to 7 times, with 99% percentile latency reduced by 90%, and MTTR reduced by 90%.

Features

Optimized Kernel and Architecture

High availability architecture with unlimited cluster scaling and a deeply optimized kernel. High-Availability Architecture ApsaraDB for HBase adapts a high availability architecture where masters run as backups for each other. High reliability is guaranteed by real-time availability detection. Regions can be switched within a few seconds when a core node fails. Reduced Storage Costs High-compression ratios, cold storage, as well as hot and cold data separation are supported to reduce the storage costs by over 50%. Cluster Scalability Each core node can respond to up to 100 thousand queries per second and provide up to 8 TB of storage space. Disk and cluster sizes can be expanded as needed. A cluster can be scaled up to 1,000 nodes to handle 10 million queries per second and store petabytes of data. Low Read/Write Latency SSDs are used to support high-speed reads and writes. An individual data entry smaller than 0.2 KB can be read or written with a 99.9th percentile latency of 3 milliseconds and an average latency of 1 millisecond. Data Backup Cluster data backup and restore, and real-time incremental data backup are supported. The Recovery Time Objective (RTO) is reduced to less than 1 minute. Dual-cluster Disaster Recovery Primary and secondary clusters are used for automatic failover. Data between two clusters is synchronized in real time.

Support for Various Scenarios

SQL, time series, time space, and data retrieval are supported. Support for SQL The phoenix SQL component supports secondary indexes and standard SQL syntax. Support for Solr The built-in Solr component supports full-text indexes for complex searches. This component is provided for synchronizing data from ApsaraDB for HBase to Solr. Native Secondary Indexes The native secondary indexes can be read and written six times faster than phoenix indexes. No external component needs to be installed. Support for Time Series The OpenTSDB component supports time series data. Support for Time Space The GeoMesa component supports time space data. Real-Time Medium-Sized Object (MOB) Storage The real-time storage and access of objects smaller than 10 MB is supported.

Fully-Managed HBase Analytics Engine

The Analytics Engine is designed to meet user needs in data streaming and data analytics. Support for Data Streaming Support for ingesting data from Kafka, Log Service, and Message Queue allows for alerting and Extract, Transform, Load (ETL) processing. Support for Various Data Sources Support for various data sources, including HBase, Object Storage Service (OSS), RDS, and MongoDB allows for complex data analytics. High Performance Support for operator pushdown and column pruning.

Efficient Operations and Maintenance

A visualized and easy-to-use O&M platform is provided that automatically upgrades the system to the latest version. Cluster Monitoring on Cloud Cluster information is monitored in real time, so that you can obtain up-to-date cluster information. Monitored metrics include CPU utilization, IOPS, connections, and disk space. Alerts are sent when anomalies are detected. Visualized Management Platform A visualized management platform is provided so that you can easily scale out clusters, modify configurations, and restart clusters. Database Kernel Version Management Automatic upgrades are supported to fix vulnerabilities at the earliest time, eliminating the need to manage kernel versions manually. HBase settings are optimized to maximize the utilization of system resources.