Inside TDengine

This chapter briefly describes some of the internal designs of TDengine.

📄️ Architecture

Cluster and Basic Logical Units

📄️ Storage Engine

The core competitiveness of TDengine lies in its outstanding write and query performance. Compared to traditional general-purpose databases, TDengine was focused from the beginning on deeply exploring the unique aspects of time-series data scenarios. It fully utilizes the time-ordered, continuous, and highly concurrent characteristics of time-series data, and has independently developed a set of writing and storage algorithms specifically tailored for time-series data.

📄️ Query Engine

TDengine, as a high-performance time-series big data platform, has its querying and computing capabilities as one of the core components. The platform offers a rich set of query processing features, including not only regular aggregation queries but also advanced functions such as time-series data window queries and statistical aggregations. These query and computation tasks require close cooperation between taosc, vnode, qnode, and mnode. In a complex supertable aggregation query scenario, multiple vnodes and qnodes may need to share the responsibilities of querying and computing. For definitions and introductions of vnode, qnode, mnode, please refer to System Architecture

📄️ Data Subscription Engine

Data subscription, as a core feature of TDengine, provides users with the ability to flexibly obtain the data they need. By deeply understanding its internal principles, users can more effectively utilize this feature to meet various real-time data processing and monitoring needs.

📄️ Stream Processing Engine

Stream Computing Architecture

📄️ Data Caching

In modern Internet of Things (IoT) and Industrial Internet of Things (IIoT) applications, efficient data management is crucial for system performance and user experience. To address the real-time read and write demands in high concurrency environments, TDengine has designed a complete caching mechanism, including write cache, read cache, metadata cache, and file system cache. These caching mechanisms are closely integrated to optimize data query response speed and improve data writing efficiency, while ensuring data reliability and high system availability. By flexibly configuring cache parameters, TDengine offers users the best balance between performance and cost.

📄️ Data Compression

Data compression is a technology that reorganizes and processes data using specific algorithms without losing effective information, aiming to reduce the storage space occupied by data and improve data transmission efficiency. TDengine employs this technology in both the storage and transmission processes to optimize the use of storage resources and accelerate data exchange.