×
Community Blog Lindorm: Alibaba Cloud's Newest Cloud-Native Multi-Model Database

Lindorm: Alibaba Cloud's Newest Cloud-Native Multi-Model Database

On May 25th 2020, Alibaba Cloud announced the commercial launch of cloud-native multi-model database Lindorm.

Lindorm - A Cloud-Native Multi-Model Database

On May 25th 2020, Alibaba Cloud announced the commercial launch of cloud-native multi-model database Lindorm, a cloud-native multi-model database service applicable for any scale. Lindorm is compatible with multiple open-source interfaces such as HBase, Phoenix, OpenTSDB, and Solr, supporting time series model, wide-column table model and queue model. It provides capabilities such as SQL query, aggregation computing, retrieval analysis, and serverless elasticity to realize the storage and analysis for structured, semi-structured and unstructured data, as well as shared storage for multi-model data.

All these capabilities make it the preferred choice of database for commercial, internet, IoT, social network, and online gaming applications. In fact, Lindorm has taken place of HBase in Alibaba Group to support the core business with a scale of hundreds of PB.

Key Benefits

Superb Cost Efficiency

Lindorm supports high concurrency and multiple storage types with core capabilities as heterogeneous hybrid storage, automatic cold and hot separation, and high compression ratio.

Elastic Serverless

Lindorm saves all the effort to manage servers and capacity levels, provides the separation of storage and Computing, and can automatically and elastically expand to PB-levels storage and tens of millions TPS.

High Stability and Availability

As the core basic storage of Alibaba, Lindorm is designed based on a high-availability architecture to eliminate single point of failure (SPOF). It has successfully supported Double Eleven Shopping Festival for more than 10 years.

Open and Integrated

Compatible with a wide variety of open-source standard interfaces, Lindorm supports switching between open source systems and seamlessly connecting with multiple computing engines such as Spark, Flink, DLA, etc.

Multi-model Capability

Wide-column table model and time series model are supported, as well as the storage and analysis of structured, semi-structured and unstructured data. It has significant performance in IoT scenario especially for the multi-model data store, computing and analytics of device metadata, device operation data (time series data), device logs, etc.

Features

Multi-model Capability

Wide-column table model and time series model are supported, compatible with various standard interfaces such as HBase, OpenTSDB, SQL, Solr, etc. With multi-model data sharing storage, the development of applications is more flexible and convenient.

  • Wide-column Table Model: Lindorm supports the wide column data model, with flexible dynamic column capability (Schema-less), compatible with HBase protocol, seamless connection with HBase and big data related scenarios, easy to achieve massive big data storage.
  • Time Series Model: Support time series data model, compatible with OpenTSDB protocol, with high performance and high compression processing of time series data by optimizations for continuous writing of massive device data, suitable for IoT, operation and maintenance monitoring and other scenarios.
  • SQL Interface Compatible: Suitable for the shift from traditional relational databases with consistent user experience, which has lower cost and higher expansion capabilities.
  • Solr Interface Compatible: Search model supported, can realize real-time full-text search easily based on distributed full-text indexes.

Global Secondary Index

In addition to meeting the characteristics of global distribution, strong consistency, on-demand index, and redundancy, Lindorm is also compatible with the schemaless model to accelerate non-primary key queries.

Integrated Storage and Retrieval

The built-in multi-model engine and search engine are integrated for the integrated use of storage and retrieval, and satisfy the needs of data storage, multi-dimensional query, full-text index and etc., in massive data scenarios requiring unified mixed access.

Separation of Cold and Hot Data

Automatic hot and cold data recognition, flexible adjustment of the hot and cold separation criteria, completely transparent to applications, which results in a great reduction of the cost of cold data storage and the better performance of hot data access.

Strong Consistency and High Availability

Based on multi-copy technology with multiple availability zones, Lindorm supports the deployment of clusters across Availability Zones, performs automatic recovery of AZ-level failures, and ensures strong consistency of data; in addition, the ultimate consistent mode with advanced optimized performance/availability is available as an optional choice.

Seamless Connections Throughout the Ecosystem

Lindorm can be seamlessly connected with storage computing systems such as Spark, Flink, MySQL, DLA, MaxCompute, etc., and allow simple and easy-to-use product experiences.

  • Data Channel: Through link services like BDS/DTS, for online real-time synchronization and full historical data between Lindorm and common storage systems (HBase, MySQL, SLS, etc.) can be achieved.
  • Computational Analysis: Lindorm provides unified standard data interfaces and on-demand conversion between data formats, and supports open computing engines such as Spark, DLA, and Hive for real-time interactive analysis of data and complex batch analysis.
  • Data Visualization: Lindorm provides the interface with QuickBI and DataV to easily perform visualized data access and analysis.

Intelligent Diagnosis

With the help of its own LDInsight tool, users can automatically diagnose common problems, such as slow requests, hot spots, performance diagnosis, capacity analysis, and schema design, etc., and provide processing suggestions and automatic resolving capabilities.

Enterprise-level Capabilities

  • Security: support multiple network access controls such as the Internet, VPC, and security groups, and provides multiple security protection methods such as account authentication, authority, encryption, and audit.
  • Backup and Recovery: support on-demand and regular backup and recovery. Superscale archived data can be recovered fast and point-in-time without affecting production availability, which helps to meet corporate and government regulations.
  • Low Latency: The response time is as low as single-digit milliseconds under tens of millions of concurrencies.
0 0 0
Share on

ApsaraDB

63 posts | 5 followers

You may also like

Comments