Help Center> CloudTable Service> Service Overview> What Is CloudTable
View PDF

What Is CloudTable

CloudTable Service (CloudTable for short) is a fully managed NoSQL data storage service based on Apache HBase. It offers strong consistency and single-digit millisecond latency, making it a great fit for storing massive amounts of structured, semi-structured, time series, and spatiotemporal data. It can be widely used in Internet of Things (IoT), Internet of Vehicles (IoV), finance, smart city, meteorology, and other industries.

The CloudTable cluster mode provides users with dedicated clusters, which are available upon subscription. This mode is applicable to users with high service throughput and low delay requirements.

Architecture

The following figure shows the architecture of CloudTable.

Figure 1 Architecture
  • Open ecosystem: CloudTable is compatible with the native APIs of HBase, OpenTSDB, and GeoMesa.
  • HBase: Key-value data models and high availability (HA) of the architecture are supported. Two HMaster nodes in active/standby mode realize real-time HA detection. Failover between regions in seconds can address unexpected computing unit failures and ensure service HA. Separation of storage and computing ensures high data reliability. A multi-backup storage mechanism is used, ensuring no less than 99.999999% storage reliability.
  • Time series database: OpenTSDB is integrated to provide efficient read, write, query, and computing of time series data. Read/write performance is improved by 30% to 60%. Interpolation, downsampling, and aggregation analysis capabilities are supported. High compression ratio (10:1) lowers costs.
  • Spatiotemporal big data: GeoMesa (an open source suite of tools that enables large-scale geospatial querying and analytics) is integrated to store and analyze spatiotemporal data and provide functions such as tracking query, region distribution statistics, region query, density analysis, aggregation, and origin-destination (OD) analysis, surpassing traditional spatial database functions.
  • Storage layer: HA is supported. Faulty disks do not affect data read/write and reliability. The computing and storage are separated. Storage is charged by usage. This costs less.
  • Cluster management: A visualized management platform provides functions such as cluster management, configuration parameter modification, and online computing unit scale-out. In addition, the platform supports visualized monitoring and O&M and provides common monitoring information (such as CPU usage, IOPS, and incoming and outgoing traffic) and custom alarm rules, simplifying system O&M.
  • Security: CloudTable uses Identity and Access Management (IAM) to implement identity authentication Underlying computing units of CloudTable run on a Virtual Private Cloud (VPC) network. VPC isolates networks logically and provides a secure and isolated network environment. Security groups are used to implement access control in and between security groups. You can customize security group rules to control access to CloudTable.