This topic describes the latest and previous specifications of Time Series Database (TSDB) for InfluxDB® instances and the prices for different specifications.
Price for storage
TSDB for InfluxDB® supports the following three types of disks: SSDs, ultra disks, and PL1 Enhanced SSDs (ESSDs). The following table describes the three types of disks.
|SSDs||SSDs are block-level Elastic Block Storage (EBS) devices that use a distributed storage architecture to ensure low latency, high performance, high durability, and high reliability. You can create, release, and resize SSDs based on your requirements at any time. We recommend that you use SSDs as the storage of TSDB for InfluxDB® instances.|
|Ultra disks||Ultra disks are general block-level Elastic Block Storage (EBS) devices that use a distributed storage architecture to provide performance and throughput that are slightly lower than those of SSDs while ensuring high durability and high reliability. You can create, release, and resize ultra disks based on your requirements at any time.|
|PL1 ESSDs||ESSDs are developed by Alibaba Cloud based on the next-generation distributed block storage architecture to deliver ultra-high storage performance. ESSDs utilize 25 Gigabit Ethernet and remote direct memory access (RDMA) to ensure low one-way latency and high random read/write capabilities.|
For more information about the performance levels of disks, see EBS performance. The throughput of a TSDB for InfluxDB® instance is also related to the number of write requests that are sent to the instance at the same time.
The following table describes the unit prices for disks in different regions.
|Region||Unit price for ultra disks||Unit price for SSDs|
|Malaysia (Kuala Lumpur)||0.34||1.21|
|Chinese mainland regions||0.35||1|
|China (Hong Kong)||0.35||1.27|
|US (Silicon Valley)||0.35||0.35|
Prices for different specifications
- The performance metrics listed in the following table are estimated based on the empirical data of specific test cases. The performance metrics are for your reference only when you purchase the services. The actual system performance varies based on scenarios, system loads, data structures, and query complexity.
- Benchmark reference data for write performance testing: A single write request contains 500 points, and all data is written to only one database. A total of 50,000 series are written. Each series includes five tag key-value pairs and five field key-value pairs.
- Benchmark reference data for read performance testing: Each query scans 10 series and 10,000 raw points.
Performance metrics for reference
- In the following table, the Number of write requests per second row lists the number of write requests per second for each specification.
- In the following table, the Number of written points per second row lists the number of points that are written per second.
- The query performance is measured by the number of queries per second, which is listed in the Number of query requests per second row. Each query scans 10 series and 10,000 raw points in the tests.
|Edition||Standard Edition||High-availability Edition|
|Specification||4 cores, 16 GB||8 cores, 32 GB||16 cores, 64 GB||32 cores, 128 GB||64 cores, 256 GB||4 cores, 16 GB||8 cores, 32 GB||16 cores, 64 GB||32 cores, 128 GB||64 cores, 256 GB|
|Number of write requests per second||~100||~160||~300||~500||~800||~160||~240||~400||~600||~1000|
|Number of written points per second||~50000||~80000||~150000||~250000||~400000||~80000||~120000||~200000||~300000||~500000|
|Number of query requests per second||~50||~80||~150||~250||~400||~170||~300||~480||~780||~1250|
|China||Chinese mainland regions||541||1074||2140||4271||8535||1503||2986||5951||11881||23742|
|China (Hong Kong)||802||1604||3208||6415||12830||2230||4462||8922||17844||35689|
|Asia Pacific||Japan (Tokyo)||888||1776||3550||7100||14200||2470||4939||9875||19750||39502|
|Malaysia (Kuala Lumpur)||762||1524||3047||6094||12187||2120||4239||8474||16951||33902|
|US (Silicon Valley)||824||1647||3294||6588||13176||2291||4582||9163||18326||36652|
- TSDB for InfluxDB® High-availability Edition instances store three replicas of data across three nodes. However, the storage capacity that you select on the purchase page is the capacity of the SSDs of one node. The actual storage capacity that is allocated to each instance of High-availability Edition is three times the standard SSD capacity that you specify on the purchase page for a single node. Therefore, the prices for High-availability Edition in the preceding table are calculated based on a unit price that is three times the unit price for SSDs that you specify on the purchase page.
- To provide better services, TSDB for InfluxDB® instances with 2 CPU cores and 8 GB of memory are no longer available. If you have already purchased TSDB for InfluxDB® instances with this specification, you can continue using these instances.