Lindorm:Features

Efficient reading and writing for time series data

LindormTSDB provides efficient and concurrent reading and writing capabilities. You can read millions of data points per second and write tens of millions of data points per second.

• Data writing

  You can use one of the following methods to write data to LindormTSDB:

  • Use the INSERT statement of SQL. We recommend that you use Java Database Connectivity (JDBC) to connect to LindormTSDB if you choose this method.

  • Use the InfluxDB line protocol to write data.

  • Use the OpenTSDB API to write data.

  • Use Prometheus to write monitoring data to the remote storage.

• Data query

  You can execute SQL statements to query data or use the OpenTSDB API to query data. You can also use the data query feature in the Lindorm console to visualize the queried data based on data groups, downsampling, and spatial aggregation.

Data management

• Database management

  Time series data stored in different databases is isolated. You can execute SQL statements to specify the interval between time-based partitions, cold/hot data boundary, time-to-live (TTL) values of data for these databases. For more information, see Database management.

• User and permission management

  You can verify the identities of the users that access your time series data and check the permissions that the users have on the storage that stores the time series data. You can execute SQL statements or use the Lindorm console to manage the users in LindormTSDB and the permissions of these users. For more information, see Manage users and permissions.

Efficient time series data storage technologies

• Time series data compression

  LindormTSDB uses an efficient compression technology to reduce the average storage space occupied by each data point to 1 to 2 bytes. This reduces the overall storage usage by 90% and increases the speed of data writing.

• Cold data archiving

  LindormTSDB allows you to archive cold data for long-term business and store the archives in cost-effective storage. This can reduce the storage costs by more than 70%.

Computing capabilities for time series data

LindormTSDB provides various computing functions to allow you to process time series data. LindormTSDB supports multiple data processing operations, such as downsampling, data interpolation, and spatial aggregation, which allow you to perform complex queries.

Monitoring and O&M

LindormTSDB provides an O&M system that you can use to maintain your Lindorm instances. You can view the status, performance metrics, and storage usage of your instances in real time. You can also configure alert rules and alert notification methods. This way, you can identify resource bottlenecks at the earliest opportunity.

Data security and cluster security

LindormTSDB provides the following solutions to ensure the security of your data and Lindorm instances:

• LindormTSDB allows you to access your instances over virtual private clouds (VPCs) to ensure the security of the instances.

• LindormTSDB allows you to configure whitelists. You can configure a whitelist for an instance and add the IP addresses of the clients that you want to allow to access the instance to the whitelist. This further secures your instances and data. If a client and a Lindorm instance are deployed in the same VPC, but the IP address of the client is not included in the whitelist of the instance, the client cannot access the instance.

• LindormTSDB allows you to manage user permissions at the database level. You can use whitelists together with permission control to manage data access in a more fine-grained manner.

• By default, LindormTSDB creates multiple replicas of data to ensure data availability.

In-database machine learning

LindormTSDB provides an out-of-the-box in-database machine learning service (Lindorm ML). You can execute SQL statements to perform the full process of machine learning to explore more data value.

The Lindorm ML service has the following benefits:

• Easy-to-use: You can execute standard SQL statements to perform the full process of machine learning even if you do not have professional knowledge for machine learning.

• Data retention: You do not need to export data to external platforms for machine learning. This improves the efficiency of machine learning and helps your data meet regulation requirements.

• Enterprise features: In Lindorm ML, models and data are both stored in databases. Therefore, you can use the enterprise features provided for databases in machine learning, such as permission management, data audit, and data encryption.

Lindorm ML supports common time series forecasting and anomaly detection algorithms.