When creating a Tencent Cloud CVM, the instance type specified by the user determines the hardware configuration of the instance. Each instance type offers varying compute, memory, and storage capabilities. Users can choose an appropriate instance type based on the scale of the application they need to deploy and run. These instance families consist of different combinations of CPU, memory, storage, heterogeneous hardware, and network bandwidth, allowing you to flexibly select the right resources for your application.
For stable businesses, it is recommended to choose the monthly subscription billing mode, with longer purchase durations being more cost-effective. For sudden business peaks, you can opt for the pay-as-you-go billing mode, allowing you to create or terminate instances at any time and pay based on the actual usage of the instances. Billing is accurate to the second, minimizing costs to the greatest extent.
Instance Type
Tencent Cloud instance families are categorized as follows:
Local Disk Types | Subtype | Description |
Standard instances provide a balance of compute, memory, and network resources to accommodate most applications. | ||
Memory optimized instances feature large memory and are suitable for applications that require extensive memory operations, searches, and computations, such as high-performance databases and distributed memory caching. | ||
This instance family features high random IOPS, high throughput and low latency, making it suitable for I/O-intensive applications that require high disk read/write performance and low latency, such as high-performance databases. | ||
This instance family is equipped with massive storage resources and features high throughput, making it suitable for throughput-intensive applications such as Hadoop distributed computing, massive log processing, distributed file systems, and large data warehouses. | ||
Featuring up to 3.8GHz turbo frequency, these instances provide the highest single-core computing performance. They are suitable for compute-intensive applications such as batch processing, high-performance computing, and large-scale game servers. | ||
- | This family is equipped with heterogeneous hardware such as GPU and FPGA to deliver real-time, fast parallel computing and floating-point computing capabilities. It is suitable for high-performance applications such as deep learning, scientific computing, video encoding/decoding, and graphics workstations. | |
With the lowest per core-hour cost, this family is suitable for compute-intensive applications that frequently deploy ultra-large amounts of computing nodes for short periods of time. Examples include rendering, genetic analysis, and solid-state pharmaceutical research. | ||
- | Bare Metal Cloud Servers are high-performance bare metal cloud servers developed based on Tencent Cloud's latest virtualization technology. They combine the flexibility and elasticity of cloud servers with the high stability and robust computing power of physical machines. These servers seamlessly integrate with the full range of Tencent Cloud products. | |
- | High-performance computing clusters use bare metal cloud servers as nodes, interconnected via RDMA, providing high-bandwidth and ultra-low latency network services. These clusters can meet the parallel computing requirements of large-scale high-performance computing, artificial intelligence, and big data recommendation applications. |
Note:
Some instances have been discontinued. For more information, see Instance Specifications (Discontinued). Glossary
vCPU: Each vCPU corresponds to a hyperthread of an Intel Xeon processor core or a hyperthread of an AMD EPYC processor core.
Send/Receive Packets: the maximum number of data packets that the instance can process per second for both sending and receiving, without differentiating private or public network traffic.
Private Network Bandwidth Capacity: The maximum amount of data (bit) that can be transmitted per second within the instance's private network.
Multi-queue Technology: multiple send/receive packet queues are supported on a virtual ENI, with the sending and receiving of the data packets of each queue capable of being simultaneously processed by different CPUs
Note:
If you purchase a larger instance specification, it is recommended to use multi-queue technology to achieve higher performance. If you use a single queue, some network metrics may not reach their maximum values.
To adjust the ENI queues, please refer to the Steps in the ENI Multi-Queue Configuration Error documentation.
Number of Queues: the quantity of send/receive packet queues supported by each virtual ENI (N number of queues indicates that N number of receiving queues and N number of sending queues can be simultaneously supported)
Instance Limits
There is a limit to the total number of instances that can be launched within a region. For more information on these limitations, please refer to CVM Instance Purchase Limits.
Instance system disk and data disk mounting limitations: To ensure optimal disk IO performance, Tencent Cloud imposes restrictions on the size and type of data disks purchased with instances. Please refer to the disk configuration module for the corresponding instance family. If necessary, you can bypass the instance disk limitations by purchasing separate cloud disks.
Note that the private network bandwidth capacity of an instance specification is the maximum private network bandwidth limit of the corresponding instance. If the CVM private network traffic exceeds this limit, random packet loss may happen within the private network for your instances.
The availability of instance specifications may vary from region to region. Some configurations may be sold out. Please see the purchase page for the latest information.
The network packet sending and receiving data in this document are based on pure network forwarding test scenarios. For testing methods, please refer to Network Performance Testing. For performance data in actual business scenarios, it is essential to conduct tests before estimating capacity.
Based on your business scenarios, the instance families can be divided into the following categories:
Standard Instance Family
Standard instances provide a balance of compute, memory, and network resources to accommodate most applications.
Standard S6
As the latest generation of standard instances, S6 instances are powered by an all-new optimized virtualization platform, and provide balanced and stable compute, memory, and network resources. These instances are ideal for many applications.
S6 instances come with the new Intel® Xeon® processors with the latest DDR4 memory. They are network optimized by default, offering a private network throughput of up to 19 million pps and a private network bandwidth of up to 100 Gbps.
Features
A new-generation Star Lake two-way server developed by Tencent Cloud, with the 3rd generation Intel® Xeon® scalable processor
Intel® Xeon® Ice Lake processor, with a CPU clock rate of 2.7 GHz and a turbo frequency of 3.3 GHz
Support the processor-to-memory ratios of 1:2 and 1:4
Up to 100 Gbps private network bandwidth for ultra high packet throughput to meet extremely high private network transfer requirements
The network performance of an instance corresponds to its specifications; the higher the specifications, the stronger the network forwarding performance, and the higher the internal network bandwidth cap.
Supports hyper-threading
Use Cases
Standard S6 instances are applicable to the following use cases:
Enterprise applications of different types and sizes
Small and medium-sized database systems, caches, and search clusters
Computing clusters, memory-intensive data processing
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Requirements
S6 instances are pay-as-you-go.
Supports only VPC.
We recommend you use the instance together with the TencentOS Server operating system to enjoy the best application performance.
S6 instances are available for purchase with configurations as shown in the table below. Ensure that the selected S6 instance size meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Instances can achieve up to 100 Gbps network bandwidth, depending on the support of the instance's operating system kernel version and runtime environment. When the PPS exceeds 10 million and the bandwidth is greater than 50 Gbps, the kernel protocol stack may cause significant network performance degradation. In this case, the netperf test bandwidth value may not meet expectations. You can use the DPDK method to mask the differences in the cloud server kernel protocol stack and obtain the true network performance of the instance. For testing methods, please refer to High-Throughput Network Performance Testing.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
S6.MEDIUM2 | 2 | 2 | 300,000 | 250,000 | 2 | 2 | 2.7GHz | - |
S6.MEDIUM4 | 2 | 4 | 300,000 | 250,000 | 2 | 2 | 2.7GHz | - |
S6.MEDIUM8 | 2 | 8 | 300,000 | 250,000 | 2 | 2 | 2.7GHz | - |
S6.LARGE8 | 4 | 8 | 0.6 million | 0.5 million | 4 | 4 | 2.7GHz | - |
S6.LARGE16 | 4 | 16 | 0.6 million | 0.5 million | 4 | 4 | 2.7GHz | - |
S6.2XLARGE16 | 8 | 16 | 1.2 million | 1 million | 8 | 7 | 2.7GHz | - |
S6.2XLARGE32 | 8 | 32 | 1.2 million | 1 million | 8 | 7 | 2.7GHz | - |
S6.4XLARGE32 | 16 | 32 | 2.5 million | 2,000,000 | 16 | 13 | 2.7GHz | - |
S6.4XLARGE64 | 16 | 64 | 2.5 million | 2,000,000 | 16 | 13 | 2.7GHz | - |
S6.8XLARGE64 | 32 | 64 | 5,000,000 | 4,000,000 | 32 | 26 | 2.7GHz | - |
S6.8XLARGE128 | 32 | 128 | 5,000,000 | 4,000,000 | 32 | 26 | 2.7GHz | - |
S6.12XLARGE96 | 48 | 96 | 7,000,000 | 6 million | 48 | 39 | 2.7GHz | - |
S6.12XLARGE192 | 48 | 192 | 7,000,000 | 6 million | 48 | 39 | 2.7GHz | - |
S6.16XLARGE216 | 64 | 216 | 10 million | 8 million | 48 | 50 | 2.7GHz | - |
S6.32XLARGE432 | 128 | 432 | 15 million | 16,000,000 | 48 | 100 | 2.7GHz | - |
Standard SA3
Standard SA3 instances utilize the latest generation of AMD EPYCTM processors, offering a wide range of specifications and exceptional cost-performance ratio. These instances provide a balanced combination of compute, storage, and network resources, making them the optimal choice for numerous applications.
Standard SA3 instances utilize the latest AMD EPYCTM Milan processors, feature the latest DDR4 memory, and come with default network optimization. They offer a maximum intranet transmission capacity of 19 million pps and support up to 100 Gbps intranet bandwidth.
Features
Provides high performance with high reliability, security, and stability based on the Star Lake servers developed by Tencent Cloud
2.55GHz AMD EPYC™ Milan processor, with a turbo frequency of 3.5GHz, equipped with the latest generation of eight-channel DDR4, ensuring stable memory computing performance.
Higher specification, SA3.58XLARGE940, offering 232 vCPUs and 940 GB memory
Support the processor-to-memory ratios of 1:2 and 1:4
Supports up to 100 Gbps of private network bandwidth for ultra-high packet throughput to meet extremely high private network transmission requirements
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Standard SA3 instances are applicable to the following use cases:
Enterprise applications of different types and sizes
Search and other computing clusters
Video encoding/decoding, video rendering, and other applications sensitive to single-core performance
Requirements
SA3 instances are pay-as-you-go.
Supports only VPC.
We recommend you use the instance together with the TencentOS Server operating system to enjoy the best application performance.
SA3 instances are available for purchase with various configurations. Please refer to the instance specifications below. Ensure that the SA3 instance size you choose meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Instances can achieve up to 100 Gbps network bandwidth, depending on the support of the instance's operating system kernel version and runtime environment. When the PPS exceeds 10 million and the bandwidth is greater than 50 Gbps, the kernel protocol stack may cause significant network performance degradation. In this case, the netperf test bandwidth value may not meet expectations. You can use the DPDK method to mask the differences in the cloud server kernel protocol stack and obtain the true network performance of the instance. For testing methods, please refer to High-Throughput Network Performance Testing.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
SA3.MEDIUM2 | 2 | 2 | 300,000 | 250,000 | 2 | 1.5 | 2.55GHz | - |
SA3.MEDIUM4 | 2 | 4 | 300,000 | 250,000 | 2 | 1.5 | 2.55GHz | - |
SA3.MEDIUM8 | 2 | 8 | 300,000 | 250,000 | 2 | 1.5 | 2.55GHz | - |
SA3.LARGE8 | 4 | 8 | 0.5 million | 250,000 | 4 | 2 | 2.55GHz | - |
SA3.LARGE16 | 4 | 16 | 0.5 million | 250,000 | 4 | 2 | 2.55GHz | - |
SA3.2XLARGE16 | 8 | 16 | 8 million | 0.5 million | 8 | 4 | 2.55GHz | - |
SA3.2XLARGE32 | 8 | 32 | 8 million | 0.5 million | 8 | 4 | 2.55GHz | - |
SA3.4XLARGE32 | 16 | 32 | 1.5 million | 1.1 million | 16 | 7 | 2.55GHz | - |
SA3.4XLARGE64 | 16 | 64 | 1.5 million | 1.1 million | 16 | 7 | 2.55GHz | - |
SA3.8XLARGE64 | 32 | 64 | 2.5 million | 2.2 million | 32 | 14 | 2.55GHz | - |
SA3.8XLARGE128 | 32 | 128 | 2.5 million | 2.2 million | 32 | 14 | 2.55GHz | - |
SA3.12XLARGE96 | 48 | 96 | 4,000,000 | 3,300,000 | 48 | 21 | 2.55GHz | - |
SA3.12XLARGE192 | 48 | 192 | 4,000,000 | 3,300,000 | 48 | 21 | 2.55GHz | - |
SA3.16XLARGE128 | 64 | 128 | 5.2 million | 4.4 million | 48 | 28 | 2.55GHz | - |
SA3.16XLARGE256 | 64 | 256 | 5.2 million | 4.4 million | 48 | 28 | 2.55GHz | - |
SA3.20XLARGE160 | 80 | 160 | 6.5 million | 5,500,000 | 48 | 35 | 2.55GHz | - |
SA3.20XLARGE320 | 80 | 320 | 6.5 million | 5,500,000 | 48 | 35 | 2.55GHz | - |
SA3.24XLARGE192 | 96 | 192 | 7.8 million | 6,600,000 | 48 | 42 | 2.55GHz | - |
SA3.24XLARGE384 | 96 | 384 | 7.8 million | 6,600,000 | 48 | 42 | 2.55GHz | - |
SA3.29XLARGE216 | 116 | 216 | 9.5 million | 8 million | 48 | 50 | 2.55GHz | - |
SA3.29XLARGE470 | 116 | 470 | 9.5 million | 8 million | 48 | 50 | 2.55GHz | - |
SA3.40XLARGE320 | 160 | 320 | 13,000,000 | 11,000,000 | 48 | 69 | 2.55GHz | - |
SA3.40XLARGE640 | 160 | 640 | 13,000,000 | 11,000,000 | 48 | 69 | 2.55GHz | - |
SA3.58XLARGE432 | 232 | 432 | 19 million | 16,000,000 | 48 | 100 | 2.55GHz | - |
SA3.58XLARGE940 | 232 | 940 | 19 million | 16,000,000 | 48 | 100 | 2.55GHz | - |
Standard SR1
Standard SR1 instances are the latest generation of ARM-based standard instances, built on a newly optimized virtualization platform. They provide balanced, stable compute, memory, and network resources, making them the optimal choice for a wide range of applications.
Note:
Currently, this model is available for sale in Guangzhou Zone 6. You can visit the purchase page to buy it. Other regions will be gradually opened, so please stay tuned for product updates.
Features
Featuring Ampere® Altra® processors, with a base frequency of 2.8GHz and a consistent maximum frequency of 3.0GHz across all cores.
Support the processor-to-memory ratios of 1:2 and 1:4
Data disks can support Premium Cloud Disk, SSD Cloud Disk, and Enhanced SSD Cloud Disk.
Support up to 25 Gbps of private network bandwidth
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Standard SR1 instances are applicable to the following use cases:
Enterprise applications of different types and sizes
ARM Simulation Test
Embedded development and testing
CPU-based Machine Learning Inference
Requirements
SR1 instances are pay-as-you-go.
The network of SR1 instances must be based on a VPC.
SR1 instances are available for purchase with configurations as detailed in the table below. Ensure that the SR1 instance size you choose meets the minimum CPU and memory requirements for your operating system and applications. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or opt for other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
SR1.SMALL1 | 1 | 1 | 250,000 | 1 | 1.5 | 2.8GHz | - |
SR1.SMALL2 | 1 | 2 | 250,000 | 1 | 1.5 | 2.8GHz | - |
SR1.MEDIUM4 | 2 | 4 | 250,000 | 2 | 1.5 | 2.8GHz | - |
SR1.MEDIUM8 | 2 | 8 | 250,000 | 2 | 1.5 | 2.8GHz | - |
SR1.LARGE8 | 4 | 8 | 250,000 | 2 | 1.5 | 2.8GHz | - |
SR1.LARGE16 | 4 | 16 | 250,000 | 2 | 1.5 | 2.8GHz | - |
SR1.2XLARGE16 | 8 | 16 | 0.5 million | 2 | 3 | 2.8GHz | - |
SR1.2XLARGE32 | 8 | 32 | 0.5 million | 2 | 3 | 2.8GHz | - |
SR1.4XLARGE32 | 16 | 32 | 1.1 million | 4 | 6 | 2.8GHz | - |
SR1.4XLARGE64 | 16 | 64 | 1.1 million | 4 | 6 | 2.8GHz | - |
SR1.8XLARGE64 | 32 | 64 | 2.2 million | 8 | 12 | 2.8GHz | - |
SR1.8XLARGE128 | 32 | 128 | 2.2 million | 8 | 12 | 2.8GHz | - |
SR1.16XLARGE128 | 64 | 128 | 4.5 million | 16 | 16 | 2.8GHz | - |
Standard S5
Standard S5 instances are the next-generation standard instances, based on a newly optimized virtualization platform. They provide balanced, stable compute, memory, and network resources, making them the optimal choice for many applications.
Features
Intel® Xeon® Cascade Lake or Intel® Xeon® Cooper Lake processor, with a CPU clock rate of 2.5 GHz and turbo frequency of 3.1 GHz
New Intel Advanced Vector Extension (AVX-512) instruction set
1:2 or 1:4 processor to memory ratio
It can support up to 29 Gbps intranet bandwidth and ultra-high network packet transmission and reception capabilities, meeting the extremely high intranet transmission demands.
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Standard S5 instances are applicable to the following scenarios:
Enterprise applications of different types and sizes
Small and medium-sized database systems, caches, and search clusters
Computing clusters, memory-intensive data processing
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Requirements
S5 instances are pay-as-you-go.
The network of S5 instances must be based on a VPC.
S5 instances support purchase configurations; please refer to the instance specifications below. Ensure that the selected S5 instance size meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
S5.SMALL1 | 1 | 1 | 250,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
S5.SMALL2 | 1 | 2 | 250,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
S5.SMALL4 | 1 | 4 | 250,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
S5.MEDIUM2 | 2 | 2 | 250,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
S5.MEDIUM4 | 2 | 4 | 300,000 | 250,000 | 2 | 1.5 | 2.5GHz | - |
S5.MEDIUM8 | 2 | 8 | 300,000 | 250,000 | 2 | 1.5 | 2.5GHz | - |
S5.LARGE4 | 4 | 4 | 0.5 million | 250,000 | 2 | 1.5 | 2.5GHz | - |
S5.LARGE8 | 4 | 8 | 0.5 million | 250,000 | 2 | 1.5 | 2.5GHz | - |
S5.LARGE16 | 4 | 16 | 0.5 million | 250,000 | 2 | 1.5 | 2.5GHz | - |
S5.2XLARGE16 | 8 | 16 | 8 million | 250,000 | 2 | 3.0 | 2.5GHz | - |
S5.2XLARGE32 | 8 | 32 | 8 million | 250,000 | 2 | 3.0 | 2.5GHz | - |
S5.4XLARGE32 | 16 | 32 | 1.5 million | 300,000 | 4 | 6.0 | 2.5GHz | - |
S5.4XLARGE64 | 16 | 64 | 1.5 million | 300,000 | 4 | 6.0 | 2.5GHz | - |
S5.6XLARGE48 | 24 | 48 | 2,000,000 | 400,000 | 6 | 9.0 | 2.5GHz | - |
S5.6XLARGE96 | 24 | 96 | 2,000,000 | 400,000 | 6 | 9.0 | 2.5GHz | - |
S5.8XLARGE64 | 32 | 64 | 2.5 million | 0.6 million | 8 | 12 | 2.5GHz | - |
S5.8XLARGE128 | 32 | 128 | 2.5 million | 0.6 million | 8 | 12 | 2.5GHz | - |
S5.12XLARGE96 | 48 | 96 | 4,000,000 | 900,000 | 12 | 17.0 | 2.5GHz | - |
S5.12XLARGE192 | 48 | 192 | 4,000,000 | 900,000 | 12 | 17.0 | 2.5GHz | - |
S5.16XLARGE256 | 64 | 256 | 5,000,000 | 1.2 million | 16 | 23.0 | 2.5GHz | - |
S5.21XLARGE320 | 84 | 320 | 6 million | 1,600,000 | 16 | 29.0 | 2.5GHz | - |
Standard Storage Optimized S5se
Standard Storage Optimized S5se instances are based on a newly optimized virtualization platform, comprehensively enhancing cloud disk performance. They utilize the latest Intel® Xeon® Cascade Lake processors and DDR4 memory, providing exceptional cloud disk performance.
Features
Supports Premium Cloud Disk and SSD as the system disk
Supports ulTra SSD as the data disk
Instance cloud disk storage performance reaches up to 1.1 million IOPS and 4 GB/s read/write bandwidth.
2.5 GHz Intel® Xeon® Cascade Lake processor, with a turbo frequency 3.1 GHz and the latest generation 6-channel DDR4 for stable memory computing performance
New Intel Advanced Vector Extension (AVX-512) instruction set
1:2 or 1:4 processor to memory ratio
Support up to 25 Gbps of private network bandwidth
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Standard Storage Optimized S5se instances are applicable to the following scenarios:
Enterprise applications of different types and sizes
IO-intensive applications including large databases, NoSQL databases, audio/video processing services, and Elasticsearch cluster
Requirements
S5se instances are pay-as-you-go.
Supports only VPC.
See below for available S5se specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
S5se.LARGE16 | 4 | 16 | 0.5 million | 250,000 | 2 | 1.5 | 2.5GHz | - |
S5se.2XLARGE16 | 8 | 16 | 8 million | 250,000 | 2 | 3.0 | 2.5GHz | - |
S5se.2XLARGE32 | 8 | 32 | 8 million | 250,000 | 2 | 3.0 | 2.5GHz | - |
S5se.4XLARGE32 | 16 | 32 | 1.5 million | 300,000 | 4 | 6.0 | 2.5GHz | - |
S5se.4XLARGE64 | 16 | 64 | 1.5 million | 300,000 | 4 | 6.0 | 2.5GHz | - |
S5se.8XLARGE64 | 32 | 64 | 2.5 million | 0.6 million | 8 | 12.0 | 2.5GHz | - |
S5se.8XLARGE128 | 32 | 128 | 2.5 million | 0.6 million | 8 | 12.0 | 2.5GHz | - |
S5se.17XLARGE316 | 68 | 316 | 5,000,000 | 1.2 million | 16 | 25.0 | 2.5GHz | - |
Instance Types
Standard SA2 instances are the latest generation of instances featuring AMD EPYC™ processors, offering a wide range of specifications with exceptional cost-performance ratios, ensuring your workloads receive industry-leading value. These instances provide a balanced mix of compute, memory, and network resources, making them the optimal choice for many applications.
SA2 instances come with the new AMD EPYC™ ROME processors with the latest DDR4 memory. They are network optimized by default, offering a private network throughput up to 7,500,000 pps and support up to 25 Gbps of private network bandwidth.
Features
Provides high performance with high reliability, security, and stability based on the Star Lake servers developed by Tencent Cloud
2.6 GHz AMD EPYC™ ROME processor, with a turbo frequency 3.3 GHz and the latest generation 8-channel DDR4 for stable memory computing performance
Higher specification, SA2.45XLARGE464, offering 180 vCPUs and 464 GB memory
Supported CPU/RAM ratios: 1:2 and 1:4
Support up to 25 Gbps of private network bandwidth
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Standard SA2 instances are applicable to the following use cases:
Enterprise applications of different types and sizes
Search and other computing clusters
Video encoding/decoding, video rendering, and other applications sensitive to single-core performance
Requirements
SA2 instances are pay-as-you-go.
Supports only VPC.
SA2 instances support purchasing configurations; please refer to the instance specifications below. Ensure that the SA2 instance size you choose meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
SA2.SMALL1 | 1 | 1 | 250,000 | 1 | 1.5 | 2.6GHz | - |
SA2.SMALL2 | 1 | 2 | 250,000 | 1 | 1.5 | 2.6GHz | - |
SA2.SMALL4 | 1 | 4 | 250,000 | 1 | 1.5 | 2.6GHz | - |
SA2.MEDIUM2 | 2 | 2 | 250,000 | 1 | 1.5 | 2.6GHz | - |
SA2.MEDIUM4 | 2 | 4 | 300,000 | 2 | 1.5 | 2.6GHz | - |
SA2.MEDIUM8 | 2 | 8 | 300,000 | 2 | 1.5 | 2.6GHz | - |
SA2.LARGE4 | 4 | 4 | 0.5 million | 2 | 1.5 | 2.6GHz | - |
SA2.LARGE8 | 4 | 8 | 0.5 million | 2 | 1.5 | 2.6GHz | - |
SA2.LARGE16 | 4 | 16 | 0.5 million | 2 | 1.5 | 2.6GHz | - |
SA2.2XLARGE16 | 8 | 16 | 0.7 million | 2 | 1.5 | 2.6GHz | - |
SA2.2XLARGE32 | 8 | 32 | 0.7 million | 2 | 1.5 | 2.6GHz | - |
SA2.4XLARGE32 | 16 | 32 | 1 million | 4 | 3.0 | 2.6GHz | - |
SA2.4XLARGE64 | 16 | 64 | 1 million | 4 | 3.0 | 2.6GHz | - |
SA2.8XLARGE64 | 32 | 64 | 1.4 million | 8 | 5.0 | 2.6GHz | - |
SA2.12XLARGE96 | 48 | 96 | 2,100,000 | 12 | 7.0 | 2.6GHz | - |
SA2.16XLARGE128 | 64 | 128 | 2.8 million | 16 | 9.0 | 2.6GHz | - |
SA2.20XLARGE160 | 80 | 160 | 3.5 million | 16 | 12.0 | 2.6GHz | - |
SA2.22XLARGE224 | 90 | 224 | 3.75 million | 16 | 13.0 | 2.6GHz | - |
SA2.24XLARGE192 | 96 | 192 | 4.2 million | 16 | 14.0 | 2.6GHz | - |
SA2.32XLARGE256 | 128 | 256 | 5.6 million | 32 | 18.0 | 2.6GHz | - |
SA2.40XLARGE320 | 160 | 320 | 7.1 million | 32 | 23.0 | 2.6GHz | - |
SA2.45XLARGE464 | 180 | 464 | 7.5 million | 32 | 25.0 | 2.6GHz | - |
Standard S4
S4 instances provide a balance of compute, memory, and network resources, making them ideal for many applications.
S4 instances come with new Xeon® Skylake processors with the latest DDR4 memory. They are network optimized by default, offering a private network throughput up to 6,000,000 pps, and support up to 25 Gbps of private network bandwidth.
Features
2.4 GHz Intel® Xeon® Skylake 6148 processor, offering stable computing performance
New Intel Advanced Vector Extension (AVX-512) instruction set
The latest generation of 6-channel DDR4 memory with a bandwidth of 2,666 MT/s
Higher specification, S4.18XLARGE228, offering 72 vCPUs and 228 GB memory
1:2 or 1:4 processor to memory ratio
Support up to 25 Gbps of private network bandwidth
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Standard S4 instances are applicable to the following use cases:
Enterprise applications of different types and sizes
Small and medium-sized database systems, caches, and search clusters
Computing clusters, memory-intensive data processing
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Requirements
S4 instances are pay-as-you-go.
Supports only VPC.
S4 instances are available for purchase and configuration; please refer to the instance specifications below. Ensure that the selected S4 instance size meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
S4.SMALL1 | 1 | 1 | 250,000 | 1 | 1.5 | 2.4GHz | - |
S4.SMALL2 | 1 | 2 | 250,000 | 1 | 1.5 | 2.4GHz | - |
S4.SMALL4 | 1 | 4 | 250,000 | 1 | 1.5 | 2.4GHz | - |
S4.MEDIUM2 | 2 | 2 | 250,000 | 1 | 1.5 | 2.4GHz | - |
S4.MEDIUM4 | 2 | 4 | 300,000 | 2 | 1.5 | 2.4GHz | - |
S4.MEDIUM8 | 2 | 8 | 300,000 | 2 | 1.5 | 2.4GHz | - |
S4.LARGE8 | 4 | 8 | 0.5 million | 2 | 1.5 | 2.4GHz | - |
S4.LARGE16 | 4 | 16 | 0.5 million | 2 | 1.5 | 2.4GHz | - |
S4.2XLARGE16 | 8 | 16 | 8 million | 2 | 3.0 | 2.4GHz | - |
S4.2XLARGE32 | 8 | 32 | 8 million | 2 | 3.0 | 2.4GHz | - |
S4.4XLARGE32 | 16 | 32 | 1.5 million | 4 | 6.0 | 2.4GHz | - |
S4.4XLARGE64 | 16 | 64 | 1.5 million | 4 | 6.0 | 2.4GHz | - |
S4.6XLARGE48 | 24 | 48 | 2,000,000 | 6 | 8.0 | 2.4GHz | - |
S4.6XLARGE96 | 24 | 96 | 2,000,000 | 6 | 8.0 | 2.4GHz | - |
S4.8XLARGE64 | 32 | 64 | 2.5 million | 8 | 11.0 | 2.4GHz | - |
S4.8XLARGE128 | 32 | 128 | 2.5 million | 8 | 11.0 | 2.4GHz | - |
S4.12XLARGE96 | 48 | 96 | 4,000,000 | 12 | 16.0 | 2.4GHz | - |
S4.12XLARGE192 | 48 | 192 | 4,000,000 | 12 | 16.0 | 2.4GHz | - |
S4.16XLARGE256 | 64 | 256 | 5,000,000 | 16 | 22.0 | 2.4GHz | - |
S4.18XLARGE288 | 72 | 288 | 6 million | 16 | 24.0 | 2.4GHz | - |
Standard Network-optimized SN3ne
Standard SN3ne instances are a relatively new generation of network optimized instances. This family provides a balance of computing, memory, and network resources, with outstanding network throughput, and it is a good choice for many applications.
Standard Network Optimized SN3ne instances are equipped with the latest Xeon® Skylake processors with the latest DDR4 memory, default network optimization, and private network throughput up to 6,000,000 pps, with performance is nearly 8 times greater than the Standard S3 family. It can support up to 25 Gbps of private network bandwidth, with performance 2.5 times greater than the Standard S3 family.
Features
2.5 GHz Intel Xeon® Skylake 6133 processors with stable computing performance
The latest generation of 6-channel DDR4 memory with a bandwidth of 2,666 MT/s
Higher specification, SN3ne.18XLARGE228, offering 72 vCPUs and 228 GB memory
1:2 or 1:4 processor to memory ratio
Support up to 25 Gbps of private network bandwidth
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Standard Network Optimized SN3ne instances are applicable to the following scenarios:
Enterprise applications of different types and sizes
Small and medium-sized database systems, caches, and search clusters
Computing clusters, memory-intensive data processing
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Requirements
SN3ne instances are pay-as-you-go.
The network of SN3ne instances must be based on a VPC.
SN3ne instances are available for purchase with various configurations. Please refer to the instance specifications below. Ensure that the selected SN3ne instance size meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
SN3ne.SMALL2 | 1 | 2 | 250,000 | 1 | 1.5 | 2.5GHz | - |
SN3ne.MEDIUM2 | 2 | 2 | 250,000 | 1 | 1.5 | 2.5GHz | - |
SN3ne.MEDIUM4 | 2 | 4 | 300,000 | 2 | 1.5 | 2.5GHz | - |
SN3ne.LARGE4 | 4 | 4 | 0.5 million | 2 | 1.5 | 2.5GHz | - |
SN3ne.LARGE8 | 4 | 8 | 0.5 million | 2 | 1.5 | 2.5GHz | - |
SN3ne.LARGE16 | 4 | 16 | 0.5 million | 2 | 1.5 | 2.5GHz | - |
SN3ne.2XLARGE16 | 8 | 16 | 8 million | 2 | 3.0 | 2.5GHz | - |
SN3ne.2XLARGE32 | 8 | 32 | 8 million | 2 | 3.0 | 2.5GHz | - |
SN3ne.3XLARGE24 | 12 | 24 | 1 million | 3 | 4.0 | 2.5GHz | - |
SN3ne.4XLARGE32 | 16 | 32 | 1.5 million | 4 | 6.0 | 2.5GHz | - |
SN3ne.4XLARGE64 | 16 | 64 | 1.5 million | 4 | 6.0 | 2.5GHz | - |
SN3ne.6XLARGE48 | 24 | 48 | 2,000,000 | 6 | 8.0 | 2.5GHz | - |
SN3ne.6XLARGE96 | 24 | 96 | 2,000,000 | 6 | 8.0 | 2.5GHz | - |
SN3ne.8XLARGE64 | 32 | 64 | 2.5 million | 8 | 11.0 | 2.5GHz | - |
SN3ne.8XLARGE128 | 32 | 128 | 2.5 million | 8 | 11.0 | 2.5GHz | - |
SN3ne.12XLARGE96 | 48 | 96 | 4,000,000 | 12 | 16.0 | 2.5GHz | - |
SN3ne.12XLARGE192 | 48 | 192 | 4,000,000 | 12 | 16.0 | 2.5GHz | - |
SN3ne.16XLARGE128 | 64 | 128 | 5,000,000 | 16 | 22.0 | 2.5GHz | - |
SN3ne.16XLARGE256 | 64 | 256 | 5,000,000 | 16 | 22.0 | 2.5GHz | - |
SN3ne.18XLARGE288 | 72 | 288 | 6 million | 16 | 24.0 | 2.5GHz | - |
Standard S3
Standard S3 instances are a relatively new generation of standard instances. This family provides a balance of computing, memory, and network resources, and it is a good choice for many applications.
Standard S3 instances are equipped with the latest Xeon® Skylake processors with the latest DDR4 memory and support for up to 10 Gbps of private network bandwidth.
Features
2.5 GHz Intel Xeon® Skylake 6133 processors with stable computing performance
The latest generation of 6-channel DDR4 memory with a bandwidth of 2,666 MT/s
Larger instance size, S3.20XLARGE320, offering 80 vCPUs and 320 GB of memory
1:2 or 1:4 processor to memory ratio
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Standard S3 instances are applicable to the following use cases:
Enterprise applications of different types and sizes
Small and medium-sized database systems, caches, and search clusters
Computing clusters, memory-intensive data processing
Requirements
S3 instances are pay-as-you-go.
S3 instances can be launched in classic networks and VPCs.
S3 instances are available for purchase and configuration. Please refer to the instance specifications below. Ensure that the selected S3 instance size meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
S3.SMALL1 | 1 | 1 | 200,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
S3.SMALL2 | 1 | 2 | 200,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
S3.SMALL4 | 1 | 4 | 200,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
S3.MEDIUM2 | 2 | 2 | 200,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
S3.MEDIUM8 | 2 | 8 | 250,000 | 250,000 | 2 | 1.5 | 2.5GHz | - |
S3.LARGE4 | 4 | 4 | 0.45 million | 350,000 | 4 | 1.5 | 2.5GHz | - |
S3.LARGE8 | 4 | 8 | 0.45 million | 350,000 | 4 | 1.5 | 2.5GHz | - |
S3.LARGE16 | 4 | 16 | 0.45 million | 350,000 | 4 | 1.5 | 2.5GHz | - |
S3.2XLARGE16 | 8 | 16 | 850,000 | 0.7 million | 8 | 1.5 | 2.5GHz | - |
S3.2XLARGE32 | 8 | 32 | 850,000 | 0.7 million | 8 | 1.5 | 2.5GHz | - |
S3.3XLARGE24 | 12 | 24 | 850,000 | 1 million | 12 | 1.5 | 2.5GHz | - |
S3.3XLARGE48 | 12 | 48 | 850,000 | 1 million | 12 | 1.5 | 2.5GHz | - |
S3.4XLARGE32 | 16 | 32 | 850,000 | 1,300,000 | 16 | 2.0 | 2.5GHz | - |
S3.4XLARGE64 | 16 | 64 | 850,000 | 1,300,000 | 16 | 2.0 | 2.5GHz | - |
S3.6XLARGE48 | 24 | 48 | 850,000 | 2,000,000 | 16 | 3.0 | 2.5GHz | - |
S3.6XLARGE96 | 24 | 96 | 850,000 | 2,000,000 | 16 | 3.0 | 2.5GHz | - |
S3.8XLARGE64 | 32 | 64 | 850,000 | 2,600,000 | 16 | 4.0 | 2.5GHz | - |
S3.8XLARGE128 | 32 | 128 | 850,000 | 2,600,000 | 16 | 4.0 | 2.5GHz | - |
S3.12XLARGE96 | 48 | 96 | 850,000 | 3.9 million | 16 | 6.0 | 2.5GHz | - |
S3.12XLARGE192 | 48 | 192 | 850,000 | 3.9 million | 16 | 6.0 | 2.5GHz | - |
S3.16XLARGE256 | 64 | 256 | 850,000 | 5.2 million | 16 | 8.0 | 2.5GHz | - |
S3.20XLARGE320 | 80 | 320 | 850,000 | 6.5 million | 16 | 10.0 | 2.5GHz | - |
Standard SA1
Standard SA1 instances are powered by AMD EPYC™ processors and offer a variety of specifications with exceptional cost-effectiveness, ensuring your workloads receive industry-leading price-performance. These instances provide a balanced mix of compute, memory, and network resources, making them the optimal choice for many applications.
Features
2.0 GHz AMD EPYC™ 7551 processors with a Turbo Boost clock of 2.55 GHz, offering stable computing performance
The latest generation of 8-channel DDR4 memory with a memory bandwidth of 2,666 MT/s
1:2 or 1:4 processor to memory ratio
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Standard SA1 instances are applicable to the following use cases:
Enterprise applications of different types and sizes
Small and medium-sized database systems, caches, and search clusters
Computing clusters, memory-intensive data processing
Requirements
SA1 instances are pay-as-you-go.
SA1 instances can be launched in classic networks and VPCs.
SA1 instances are available for purchase with various configurations. Please refer to the instance specifications below. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
SA1.SMALL1 | 1 | 1 | - | 250,000 | 1 | 1.5 | 2.0GHz | - |
SA1.SMALL2 | 1 | 2 | - | 250,000 | 1 | 1.5 | 2.0GHz | - |
SA1.SMALL4 | 1 | 4 | - | 250,000 | 1 | 1.5 | 2.0GHz | - |
SA1.MEDIUM2 | 2 | 2 | - | 250,000 | 1 | 1.5 | 2.0GHz | - |
SA1.MEDIUM4 | 2 | 4 | - | 250,000 | 2 | 1.5 | 2.0GHz | - |
SA1.MEDIUM8 | 2 | 8 | - | 250,000 | 2 | 1.5 | 2.0GHz | - |
SA1.LARGE4 | 4 | 4 | - | 350,000 | 4 | 1.5 | 2.0GHz | - |
SA1.LARGE8 | 4 | 8 | - | 350,000 | 4 | 1.5 | 2.0GHz | - |
SA1.LARGE16 | 4 | 16 | - | 350,000 | 4 | 1.5 | 2.0GHz | - |
SA1.2XLARGE16 | 8 | 16 | - | 0.7 million | 8 | 1.5 | 2.0GHz | - |
SA1.2XLARGE32 | 8 | 32 | - | 0.7 million | 8 | 1.5 | 2.0GHz | - |
SA1.4XLARGE32 | 16 | 32 | - | 1,300,000 | 16 | 1.5 | 2.0GHz | - |
SA1.4XLARGE64 | 16 | 64 | - | 1,300,000 | 16 | 1.5 | 2.0GHz | - |
Note:
In the SA1 instance specification list, the "-" in the network packet transmission and reception column indicates that there is no fixed commitment to network packet performance and processor frequency for that specification. If you have a strong demand for consistent performance, it is recommended to choose a model with a performance commitment.
Standard Network-optimized S2ne
The Standard Network Optimized S2ne instances are the optimal choice for applications with high network packet transmission capabilities, offering up to millions of network transmissions per second. They are recommended for large-scale gaming servers, video streaming, live broadcasting, and other scenarios with high network PPS requirements.
Note:
This instance type is currently only available to whitelisted users. Contact your account manager if you wish to purchase these instance type.
Features
Intel Xeon E5-2680 Broadwell (v4) processors with base frequency of 2.4 GHz and DDR4 memory, offering stable computing performance
Up to 48 cores and 192GB are available for sale.
1:2 or 1:4 processor to memory ratio
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
Scenarios that require sending and receiving massive network packets, such as game services, video services, and financial analysis
Enterprise applications of different types and sizes
Requirements
S2ne instances are pay-as-you-go.
S2ne instances can only be launched in a VPC.
See below for available S2ne specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
S2ne.SMALL2 | 1 | 2 | 120,000 | 1 | 1.5 | 2.4GHz | - |
S2ne.MEDIUM2 | 2 | 2 | 120,000 | 1 | 1.5 | 2.4GHz | - |
S2ne.MEDIUM4 | 2 | 4 | 150,000 | 2 | 1.5 | 2.4GHz | - |
S2ne.LARGE4 | 4 | 4 | 300,000 | 2 | 1.5 | 2.4GHz | - |
S2ne.LARGE8 | 4 | 8 | 300,000 | 2 | 1.5 | 2.4GHz | - |
S2ne.LARGE16 | 4 | 16 | 300,000 | 2 | 1.5 | 2.4GHz | - |
S2ne.2XLARGE16 | 8 | 16 | 0.6 million | 2 | 2.0 | 2.4GHz | - |
S2ne.2XLARGE32 | 8 | 32 | 0.6 million | 2 | 2.0 | 2.4GHz | - |
S2ne.3XLARGE24 | 12 | 24 | 900,000 | 3 | 2.5 | 2.4GHz | - |
S2ne.3XLARGE48 | 12 | 48 | 900,000 | 3 | 2.5 | 2.4GHz | - |
S2ne.4XLARGE32 | 16 | 32 | 1.2 million | 4 | 3.5 | 2.4GHz | - |
S2ne.4XLARGE64 | 16 | 64 | 1.2 million | 4 | 3.5 | 2.4GHz | - |
S2ne.6XLARGE48 | 24 | 48 | 1,800,000 | 6 | 5.0 | 2.4GHz | - |
S2ne.6XLARGE96 | 24 | 96 | 1,800,000 | 6 | 5.0 | 2.4GHz | - |
S2ne.8XLARGE64 | 32 | 64 | 2.4 million | 8 | 6.5 | 2.4GHz | - |
S2ne.8XLARGE128 | 32 | 128 | 2.4 million | 8 | 6.5 | 2.4GHz | - |
S2ne.12XLARGE192 | 48 | 192 | 3,600,000 | 12 | 9.5 | 2.4GHz | - |
Standard S2
Standard S2 instances are a relatively new generation of instances. This family provides a balance of computing, memory, and network resources, and it is a good choice for many applications.
Standard S2 instances are equipped with Intel® Xeon® Broadwell processors with DDR4 memory.
Features
Intel Xeon E5-2680 Broadwell (v4) processors with base frequency of 2.4 GHz and DDR4 memory
Up to 56 cores and 224 GB are available for sale.
1:2 or 1:4 processor to memory ratio
Balance of computing, memory, and network resources
Use Cases
This family is used for small and mid-size databases, data processing tasks that require additional memory and cache fleets, and for running backend servers for SAP, Microsoft SharePoint, cluster computing and other enterprise applications.
Requirements
S2 instances are pay-as-you-go, and can also be used as production instances of standard host HS20 in CDHs.
S2 instances can be launched in classic networks and VPCs.
S2 instances are available for purchase with configurations detailed below. Ensure that the selected S2 instance size meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
S2.SMALL1 | 1 | 1 | 200,000 | 250,000 | 1 | 1.5 | 2.4GHz | - |
S2.SMALL2 | 1 | 2 | 200,000 | 250,000 | 1 | 1.5 | 2.4GHz | - |
S2.SMALL4 | 1 | 4 | 200,000 | 250,000 | 1 | 1.5 | 2.4GHz | - |
S2.MEDIUM2 | 2 | 2 | 200,000 | 250,000 | 2 | 1.5 | 2.4GHz | - |
S2.MEDIUM4 | 2 | 4 | 250,000 | 250,000 | 2 | 1.5 | 2.4GHz | - |
S2.MEDIUM8 | 2 | 8 | 250,000 | 250,000 | 2 | 1.5 | 2.4GHz | - |
S2.LARGE4 | 4 | 4 | 0.45 million | 350,000 | 4 | 1.5 | 2.4GHz | - |
S2.LARGE8 | 4 | 8 | 0.45 million | 350,000 | 4 | 1.5 | 2.4GHz | - |
S2.LARGE16 | 4 | 16 | 0.45 million | 350,000 | 4 | 1.5 | 2.4GHz | - |
S2.2XLARGE16 | 8 | 16 | 0.5 million | 0.7 million | 8 | 1.5 | 2.4GHz | - |
S2.2XLARGE32 | 8 | 32 | 0.5 million | 0.7 million | 8 | 1.5 | 2.4GHz | - |
S2.3XLARGE24 | 12 | 24 | 0.5 million | 1 million | 8 | 2.5 | 2.4GHz | - |
S2.3XLARGE48 | 12 | 48 | 0.5 million | 1 million | 8 | 2.5 | 2.4GHz | - |
S2.4XLARGE32 | 16 | 32 | 0.5 million | 1,300,000 | 8 | 3.0 | 2.4GHz | - |
S2.4XLARGE64 | 16 | 64 | 0.5 million | 1,300,000 | 8 | 3.0 | 2.4GHz | - |
S2.6XLARGE48 | 24 | 48 | 0.7 million | 2,000,000 | 8 | 4.5 | 2.4GHz | - |
S2.6XLARGE96 | 24 | 96 | 0.7 million | 2,000,000 | 8 | 4.5 | 2.4GHz | - |
S2.8XLARGE64 | 32 | 64 | 0.7 million | 2,600,000 | 8 | 6.0 | 2.4GHz | - |
S2.8XLARGE128 | 32 | 128 | 0.7 million | 2,600,000 | 8 | 6.0 | 2.4GHz | - |
S2.14XLARGE224 | 56 | 224 | 0.7 million | 4.5 million | 8 | 10.0 | 2.4GHz | - |
Standard S1
Series 1 Standard S1 instances are cloud server types that cover a range of low to high core configurations, meeting users' flexible configuration needs and featuring moderate pricing and customizable options. Data disks can be paired with different types, such as local disks and SSD cloud disks, depending on the hardware specifications.
Features
The Standard S1 instance type has the following features:
Offer a range of CPUs from low to high core count for flexible configuration options
Intel Xeon CPUs and DDR3 memories
Storage can be configured with local disks and SSD cloud disks.
Balance of computing, memory, and network resources
Use Cases
Standard S1 instances are suitable for large, medium, and small-sized applications and databases.
Requirements
S1 instances are pay-as-you-go, and can also be used as production instances of standard host in CDHs.
S1 instances can be launched in basic networks and VPCs.
In the instance specification list, a "-" indicates that there is no fixed commitment for network packet transmission and processor frequency performance for that specification. If you have a strong demand for consistent performance, it is recommended to choose models with performance commitments.
S1 instances are available for purchase with configurations detailed below. Ensure that the selected S1 instance size meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
S1.SMALL1 | 1 | 1 | - | 250,000 | 1 | 1.5 | - | - |
S1.SMALL2 | 1 | 2 | - | 250,000 | 1 | 1.5 | - | - |
S1.SMALL4 | 1 | 4 | - | 250,000 | 1 | 1.5 | - | - |
S1.MEDIUM2 | 2 | 2 | - | 250,000 | 1 | 1.5 | - | - |
S1.MEDIUM4 | 2 | 4 | - | 250,000 | 2 | 1.5 | - | - |
S1.MEDIUM8 | 2 | 8 | - | 250,000 | 2 | 1.5 | - | - |
S1.MEDIUM12 | 2 | 12 | - | 250,000 | 2 | 1.5 | - | - |
S1.LARGE4 | 4 | 4 | - | 350,000 | 4 | 1.5 | - | - |
S1.LARGE8 | 4 | 8 | - | 350,000 | 4 | 1.5 | - | - |
S1.LARGE16 | 4 | 16 | - | 350,000 | 4 | 1.5 | - | - |
S1.2XLARGE8 | 8 | 8 | - | 0.7 million | 8 | 2.0 | - | - |
S1.2XLARGE16 | 8 | 16 | - | 0.7 million | 8 | 2.0 | - | - |
S1.2XLARGE32 | 8 | 32 | - | 0.7 million | 8 | 2.0 | - | - |
S1.3XLARGE24 | 12 | 24 | - | 1 million | 8 | 2.5 | - | - |
S1.3XLARGE48 | 12 | 48 | - | 1 million | 8 | 2.5 | - | - |
S1.4XLARGE16 | 16 | 16 | - | 1.4 million | 8 | 3.5 | - | - |
S1.4XLARGE32 | 16 | 32 | - | 1.4 million | 8 | 3.5 | - | - |
S1.4XLARGE48 | 16 | 48 | - | 1.4 million | 8 | 3.5 | - | - |
S1.4XLARGE64 | 16 | 64 | - | 1.4 million | 8 | 3.5 | - | - |
S1.6XLARGE48 | 24 | 48 | - | 2,000,000 | 8 | 5.0 | - | - |
S1.8XLARGE64 | 32 | 64 | - | 2,700,000 | 8 | 7.0 | - | - |
S1.12XLARGE96 | 48 | 96 | - | 4,000,000 | 8 | 10.0 | - | - |
Note:
In the S1 Instance Specification list, a "-" in the Network Packets and Base Frequency columns indicates that there is no fixed commitment for network packet transmission and processor base frequency performance. If you have a strong demand for consistent performance, it is recommended to choose a model with performance commitments.
Memory Optimized Instance Family
Memory Optimized instances feature large memory and are suitable for applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching.
MEM optimized MA3
As the latest generation of memory optimized instances, MA3 instances are powered by an all-new optimized virtualization platform and designed to deliver fast and stable performance for workloads that process large data sets in memory. They are ideal for applications that require high in-memory computing performance.
Features
Provides high performance with high reliability, security, and stability based on the Star Lake servers developed by Tencent Cloud
Equipped with AMD EPYC™ Milan processor, with a base frequency of 2.55GHz and a turbo frequency of 3.5GHz.
The processor-to-memory ratio is 1:8.
Purchase the lowest price for the same memory size
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
These instances are ideal for:
Applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching
External Hadoop clusters or Redis for fields such as computational genomics
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Requirements
MA3 instances are pay-as-you-go.
The network of MA3 instances must be based on a VPC.
We recommend you use the instance together with the TencentOS Server operating system to enjoy the best application performance.
MA3 instances are available for purchase with various configurations. Please refer to the instance specifications below. Ensure that the selected MA3 instance size meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
MA3.SMALL8 | 1 | 8 | 250,000 | 1 | 1.5 | 2.55GHz | - |
MA3.MEDIUM16 | 2 | 16 | 300,000 | 2 | 1.5 | 2.55GHz | - |
MA3.LARGE32 | 4 | 32 | 0.5 million | 4 | 2 | 2.55GHz | - |
MA3.2XLARGE64 | 8 | 64 | 8 million | 8 | 4 | 2.55GHz | - |
MA3.4XLARGE128 | 16 | 128 | 1.5 million | 16 | 7 | 2.55GHz | - |
MA3.8XLARGE256 | 32 | 256 | 2.5 million | 32 | 14 | 2.55GHz | - |
MEM optimized M6
As the latest generation of memory optimized instances, M6 instances are powered by an all-new optimized virtualization platform and designed to deliver fast and stable performance for workloads that process large data sets in memory. They are ideal for applications that require high in-memory computing performance.
Features
A new-generation Star Lake two-way server developed by Tencent Cloud, with the 3rd generation Intel® Xeon® scalable processor
Intel® Xeon® Ice Lake processor, with a CPU clock rate of 2.7 GHz and a turbo frequency of 3.3 GHz
The processor-to-memory ratio is 1:8.
Lowest price per GB of memory among instance types
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
These instances are ideal for:
Applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching
External Hadoop clusters or Redis for fields such as computational genomics
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Requirements
M6 instances are pay-as-you-go.
Supports only VPC.
We recommend you use the instance together with the TencentOS Server operating system to enjoy the best application performance.
M6 instances are available for purchase and configuration. Please refer to the instance specifications below. Ensure that the M6 instance size you choose meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Instances can achieve up to 100 Gbps network bandwidth, depending on the support of the instance's operating system kernel version and runtime environment. When the PPS exceeds 10 million and the bandwidth is greater than 50 Gbps, the kernel protocol stack may cause significant network performance degradation. In this case, the netperf test bandwidth value may not meet expectations. You can use the DPDK method to mask the differences in the cloud server kernel protocol stack and obtain the true network performance of the instance. For testing methods, please refer to High-Throughput Network Performance Testing.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
M6.SMALL8 | 1 | 8 | 250,000 | 250,000 | 1 | 1.5 | 2.7GHz | - |
M6.MEDIUM16 | 2 | 16 | 300,000 | 250,000 | 2 | 2 | 2.7GHz | - |
M6.LARGE32 | 4 | 32 | 0.6 million | 0.5 million | 4 | 4 | 2.7GHz | - |
M6.2XLARGE64 | 8 | 64 | 1.2 million | 1 million | 8 | 7 | 2.7GHz | - |
M6.4XLARGE128 | 16 | 128 | 2.5 million | 2,000,000 | 16 | 13 | 2.7GHz | - |
M6.8XLARGE256 | 32 | 256 | 5,000,000 | 4,000,000 | 32 | 26 | 2.7GHz | - |
M6.31MEDIUM470 | 62 | 470 | 10 million | 8 million | 48 | 50 | 2.7GHz | - |
M6.31XLARGE940 | 124 | 940 | 19 million | 16,000,000 | 48 | 100 | 2.7GHz | - |
Security-enhanced Memory Optimized M6ce
As the latest generation of memory optimized instances, M6ce instances are powered by an all-new optimized virtualization platform and designed to deliver fast and stable performance for workloads that process large data sets in memory. They are ideal for applications that require high in-memory computing performance.
Features
A new-generation Star Lake two-way server developed by Tencent Cloud, with the 3rd generation Intel® Xeon® scalable processor
Intel® Xeon® Ice Lake processor, with a CPU clock rate of 2.7 GHz and a turbo frequency of 3.3 GHz
Intel® SGX encrypted computing, with a single instance containing up to 428 GB encrypted memory, ensuring the confidentiality and security of critical data and codes
The processor-to-memory ratio is 1: 8. The encrypted memory accounts for about 50% of the memory
Lowest price per GB of memory among instance types
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
These instances are ideal for:
Data sharing and computing, protecting confidential data shared between different users or vendors
Blockchain applications, enhancing the privacy and security of transactions and key storage, etc.
Scenarios with high security and reliability requirements, such as finance, government agencies, medical care, etc.
Confidential computing scenarios and data encryption applications
Requirements
M6ce instances are pay-as-you-go.
Supports only VPC.
See below for available M6ce specifications:
Note:
If you have used hardware-bound keys to encrypt data in Intel SGX Enclave (such as SGX Sealing), you will not be able to decrypt the corresponding data after the host changes in the following scenarios. It is recommended to back up your data at the application layer to ensure application reliability. Scenarios that may cause host changes include host failure, adjusting instance specifications, using the shutdown without charge feature, and modifying instance placement groups.
For more information on how to use security-enhanced memory M6ce instances, please see Building an SGX Confidential Computing Environment.
Specifications | vCPU | Memory (GB) | Encrypted memory (GB) | Network packets transmitted and received (pps) (out + in) | Number of connections | Number of Queues | Private Network Bandwidth Capability (Gbps) (Outbound + Inbound) | Clock Rate | Remarks |
M6ce.MEDIUM16 | 2 | 16 | 8 | 300,000 | 250,000 | 2 | 2 | 2.7/3.3GHz | - |
M6ce.LARGE32 | 4 | 32 | 16 | 0.6 million | 0.5 million | 4 | 4 | 2.7/3.3GHz | - |
M6ce.2XLARGE64 | 8 | 64 | 32 | 1.2 million | 1 million | 8 | 7 | 2.7/3.3GHz | - |
M6ce.4XLARGE128 | 16 | 128 | 64 | 2.5 million | 2,000,000 | 16 | 13 | 2.7/3.3GHz | - |
M6ce.8XLARGE256 | 32 | 256 | 128 | 5,000,000 | 4,000,000 | 32 | 26 | 2.7/3.3GHz | - |
M6ce.31XMEDIUM428 | 62 | 428 | 214 | 10 million | 8 million | 48 | 50 | 2.7/3.3GHz | - |
M6ce.31XLARGE856 | 124 | 856 | 428 | 19 million | 8 million | 48 | 100 | 2.7/3.3GHz | - |
MEM optimized M6p
As the latest generation of memory optimized instances, M6p instances are powered by an all-new optimized virtualization platform and designed to deliver fast and stable performance for workloads that process large data sets in memory. They are ideal for applications that require high in-memory computing performance.
Features
A new-generation Star Lake two-way server developed by Tencent Cloud, with the 3rd generation Intel® Xeon® scalable processor
Intel® Xeon® Ice Lake processor, with a CPU clock rate of 2.7 GHz and a turbo frequency of 3.3 GHz
Paired with Intel® 2nd Gen Optane™ Persistent Memory (BPS), offering a cost-effective memory solution with high performance.
Lowest price per GB of memory among instance types
Use Cases
These instances are ideal for:
Applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching
External Hadoop clusters or Redis for fields such as computational genomics
Large memory application scenarios such as Hadoop or Redis clusters
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Requirements
M6p instances are pay-as-you-go.
The network of M6p instances must be based on a VPC.
M6p instances do not support configuration adjustments and the "No Charge when Shut Down" feature.
See below for available M6p specifications:
Note:
The reliability of data in persistent memory depends on the reliability of the physical server and the persistent memory device, posing a risk of data loss (e.g., during host downtime). It is recommended that you store data with long-term storage reliability requirements on cloud disks.
For more information on how to use memory-optimized M6p instances, see Configuring Persistent Memory for M6p Instances.
Specifications | vCPU | Memory (GB) | Persistent Memory (GB) | Network packets transmitted and received (pps) (out + in) | Number of connections | Number of Queues | Private Network Bandwidth Capability (Gbps) (Outbound + Inbound) | Clock Rate | Remarks |
M6p.LARGE16 | 4 | 16 | 62 | 0.6 million | 0.5 million | 4 | 4 | 2.7GHz | - |
M6p.2XLARGE32 | 8 | 32 | 124 | 1.2 million | 1 million | 8 | 7 | 2.7GHz | - |
M6p.4XLARGE53 | 16 | 53 | 248 | 2.5 million | 2,000,000 | 16 | 13 | 2.7GHz | - |
M6p.8XLARGE106 | 32 | 106 | 496 | 5,000,000 | 4,000,000 | 32 | 26 | 2.7GHz | - |
M6p.31XMEDIUM212 | 62 | 212 | 992 | 10 million | 8 million | 48 | 50 | 2.7GHz | - |
M6p.31XLARGE424 | 124 | 424 | 1984 | 19 million | 16,000,000 | 48 | 100 | 2.7GHz | - |
Memory Optimized M5
As the latest generation of memory optimized instances, M5 instances are powered by an all-new optimized virtualization platform and designed to deliver fast and stable performance for workloads that process large data sets in memory. They are ideal for applications that require high in-memory computing performance.
Features
Intel® Xeon® Cascade Lake or Intel® Xeon® Cooper Lake processor, with a CPU clock rate of 2.5 GHz and turbo frequency of 3.1 GHz
New Intel Advanced Vector Extension (AVX-512) instruction set
CPU/RAM ratio: 1:8
Lowest price per GB of memory among instance types
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
These instances are ideal for:
Applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching
External Hadoop clusters or Redis for fields such as computational genomics
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Requirements
M5 instances are pay-as-you-go.
M5 instances can only be launched in a VPC.
Configuration purchase is available for M5 instances. Please see the instance specifications below.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
M5.SMALL8 | 1 | 8 | 250,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
M5.MEDIUM16 | 2 | 16 | 300,000 | 250,000 | 2 | 1.5 | 2.5GHz | - |
M5.LARGE32 | 4 | 32 | 0.5 million | 250,000 | 2 | 1.5 | 2.5GHz | - |
M5.2XLARGE64 | 8 | 64 | 8 million | 250,000 | 2 | 3.0 | 2.5GHz | - |
M5.3XLARGE96 | 12 | 96 | 1 million | 250,000 | 3 | 5.0 | 2.5GHz | - |
M5.4XLARGE128 | 16 | 128 | 1.5 million | 300,000 | 4 | 6.0 | 2.5GHz | - |
M5.8XLARGE256 | 32 | 256 | 2.5 million | 0.6 million | 8 | 12.0 | 2.5GHz | - |
M5.16XLARGE512 | 64 | 512 | 5,000,000 | 1.2 million | 16 | 23.0 | 2.5GHz | - |
MEM optimized MA2
As the latest generation of memory optimized instances, MA2 instances are powered by an all-new optimized virtualization platform and designed to deliver fast and stable performance for workloads that process large data sets in memory. They are ideal for applications that require high in-memory computing performance.
Features
Provides high performance with high reliability, security, and stability based on the Star Lake servers developed by Tencent Cloud
Equipped with AMD EPYC™ ROME processor, with a base frequency of 2.6GHz and a turbo frequency of 3.3GHz.
CPU/RAM ratio: 1:8
For instances with the same memory size, the lowest price is offered.
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
These instances are ideal for:
Applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching
External Hadoop clusters or Redis for fields such as computational genomics
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Requirements
MA2 instances are pay-as-you-go.
The network of MA2 instances must be based on a VPC.
See below for available MA2 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
MA2.SMALL8 | 1 | 8 | 250,000 | 250,000 | 1 | 1.5 | 2.6GHz | - |
MA2.LARGE32 | 4 | 32 | 0.5 million | 250,000 | 4 | 1.5 | 2.6GHz | - |
MA2.2XLARGE64 | 8 | 64 | 0.7 million | 250,000 | 8 | 1.5 | 2.6GHz | - |
MA2.4XLARGE128 | 16 | 128 | 1 million | 300,000 | 16 | 3 | 2.6GHz | - |
MA2.8XLARGE256 | 32 | 256 | 1.4 million | 0.7 million | 32 | 5 | 2.6GHz | - |
MA2.12XLARGE384 | 48 | 384 | 2,100,000 | 1 million | 48 | 7 | 2.6GHz | - |
MEM optimized M4
M4 instances are designed to deliver fast performance for workloads that process large data sets in memory. They are ideal for applications that require high in-memory computing performance.
As a new generation of memory optimized instances, M4 instances come with new Xeon® Skylake processors with the latest DDR4 memory. They are network optimized by default, offering a private network throughput up to 6,000,000 pps, and support for ultra-high private network bandwidth.
Use Cases
These instances are ideal for:
Applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching
External Hadoop clusters or Redis for fields such as computational genomics
Scenarios that require sending and receiving massive network packets, such as on-screen video comments, live video broadcasting, and gaming
Features
2.4 GHz Intel® Xeon® Skylake 6148 processor, offering stable computing performance
New Intel Advanced Vector Extension (AVX-512) instruction set
The latest generation of 6-channel DDR4 memory with a bandwidth of 2,666 MT/s
Higher specification, M4.16XLARGE512, offering 64 vCPUs and 512 GB memory
1:8 or 1:12 processor to memory ratio
Lowest price per GB of memory among instance types
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Requirements
M4 instances are pay-as-you-go.
Supports only VPC.
See below for available M4 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
M4.SMALL8 | 1 | 8 | 250,000 | 1 | 1.5 | 2.4GHz | - |
M4.MEDIUM16 | 2 | 16 | 300,000 | 2 | 1.5 | 2.4GHz | - |
M4.LARGE32 | 4 | 32 | 0.5 million | 2 | 1.5 | 2.4GHz | - |
M4.2XLARGE64 | 8 | 64 | 8 million | 2 | 3.0 | 2.4GHz | - |
M4.3XLARGE96 | 12 | 96 | 1 million | 3 | 4.0 | 2.4GHz | - |
M4.3XLARGE144 | 12 | 144 | 1 million | 3 | 4.0 | 2.4GHz | - |
M4.4XLARGE128 | 16 | 128 | 1.5 million | 4 | 6.0 | 2.4GHz | - |
M4.4XLARGE192 | 16 | 192 | 1.5 million | 4 | 6.0 | 2.4GHz | - |
M4.8XLARGE256 | 32 | 256 | 2.5 million | 8 | 11.0 | 2.4GHz | - |
M4.8XLARGE384 | 32 | 384 | 2.5 million | 8 | 11.0 | 2.4GHz | - |
M4.16XLARGE512 | 64 | 512 | 5,000,000 | 16 | 22.0 | 2.4GHz | - |
M4.18XLARGE648 | 72 | 648 | 6 million | 16 | 24.0 | 2.4GHz | - |
MEM optimized M3
M3 instances are designed to deliver fast performance for workloads that process large data sets in memory. They are ideal for applications that require high in-memory computing performance.
M3 instances are powered by new Intel® Xeon® Skylake processors with the latest DDR4 memory, and support up to 10 Gbps of private network bandwidth.
Use Cases
These instances are ideal for:
Applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching
External Hadoop clusters or Redis for fields such as computational genomics
Features
2.5 GHz Intel Xeon® Skylake 6133 processors with stable computing performance
The latest generation of 6-channel DDR4 memory with a bandwidth of 2,666 MT/s
Higher specification, M3.16XLARGE512, offering 64 vCPUs and 512 GB memory
1:8 or 1:12 processor to memory ratio
Lowest price per GB of memory among instance types
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Requirements
M3 instances are pay-as-you-go.
Supports both classic network and VPC.
See below for available M3 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
M3.SMALL8 | 1 | 8 | 300,000 | 250,000 | 1 | 1.5 | 2.5GHz | - |
M3.MEDIUM16 | 2 | 16 | 400,000 | 250,000 | 2 | 1.5 | 2.5GHz | - |
M3.LARGE32 | 4 | 32 | 8 million | 300,000 | 4 | 1.5 | 2.5GHz | - |
M3.2XLARGE64 | 8 | 64 | 1.2 million | 0.6 million | 8 | 1.5 | 2.5GHz | - |
M3.3XLARGE96 | 12 | 96 | 1.2 million | 900,000 | 12 | 1.5 | 2.5GHz | - |
M3.3XLARGE144 | 12 | 144 | 1.2 million | 900,000 | 12 | 1.5 | 2.5GHz | - |
M3.4XLARGE128 | 16 | 128 | 1.2 million | 1.2 million | 16 | 2.0 | 2.5GHz | - |
M3.4XLARGE192 | 16 | 192 | 1.2 million | 1.2 million | 16 | 2.0 | 2.5GHz | - |
M3.8XLARGE256 | 32 | 256 | 1.2 million | 2.4 million | 16 | 4.0 | 2.5GHz | - |
M3.8XLARGE384 | 32 | 384 | 1.2 million | 2.4 million | 16 | 4.0 | 2.5GHz | - |
M3.16XLARGE512 | 64 | 512 | 1.2 million | 4.8 million | 16 | 8.0 | 2.5GHz | - |
MEM optimized M2
Memory-optimized M2 instances are designed to deliver fast performance for workloads that process large in-memory data sets. Characterized by their large memory capacity, they are the optimal choice for high-memory computing applications.
Use Cases
These instances are ideal for:
Applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching
External Hadoop clusters or Redis for fields such as computational genomics
Features
2.4 GHz Intel® Xeon® E5-2680v4 processor with DDR4 memory
Memory of up to 448 GB is available. The M2.14XLARGE448 model offers 56 vCPUs and 448 GB memory
CPU/RAM ratio: 1:8
Lowest price per GB of memory among instance types
Requirements
M2 instances are pay-as-you-go, and can also be used as production instances of memory optimized host HM20 in CDHs.
Supports only VPC.
See below for available M2 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
M2.SMALL8 | 1 | 8 | 200,000 | 250,000 | 1 | 1.5 | 2.4GHz | - |
M2.MEDIUM16 | 2 | 16 | 250,000 | 250,000 | 2 | 1.5 | 2.4GHz | - |
M2.LARGE32 | 4 | 32 | 0.45 million | 350,000 | 4 | 1.5 | 2.4GHz | - |
M2.2XLARGE64 | 8 | 64 | 0.5 million | 0.7 million | 8 | 1.5 | 2.4GHz | - |
M2.3XLARGE96 | 12 | 96 | 0.5 million | 1 million | 8 | 2.5 | 2.4GHz | - |
M2.4XLARGE128 | 16 | 128 | 0.5 million | 1,300,000 | 8 | 3.0 | 2.4GHz | - |
M2.6XLARGE192 | 24 | 192 | 0.7 million | 2,000,000 | 8 | 4.5 | 2.4GHz | - |
M2.8XLARGE256 | 32 | 256 | 0.7 million | 2,600,000 | 8 | 6.0 | 2.4GHz | - |
M2.12XLARGE384 | 48 | 384 | 0.7 million | 3.9 million | 8 | 9.0 | 2.4GHz | - |
M2.14XLARGE448 | 56 | 448 | 0.7 million | 4.5 million | 8 | 10.0 | 2.4GHz | - |
MEM optimized M1
M1 instances come with approximately 1:8 CPU to memory ratio. They are well suited for applications that require memory-intensive operations, searches, and computations, such as high-performance databases and distributed in-memory caching.
Features
2.3 GHz Intel Xeon® E5-2670 v3 processor with DDR3 memory, providing more instance sizes with stronger computing capacities
Golden CPU to memory ratio for memory-intensive applications, satisfying the needs for large-scale business deployment
Use Cases
These instances are ideal for:
Applications that require extensive memory operations, searches, and computing, such as high-performance databases and distributed memory caching
External Hadoop clusters or Redis for fields such as computational genomics
Requirements
M1 instances are pay-as-you-go.
Supports both classic network and VPC.
In the instance specification list, a "-" indicates that there is no fixed commitment for network packet transmission and processor frequency performance for that specification. If you have a strong demand for consistent performance, it is recommended to choose models with performance commitments.
See below for available M1 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
M1.SMALL8 | 1 | 8 | - | 250,000 | 1 | 1.5 | 2.3GHz | - |
M1.MEDIUM16 | 2 | 16 | - | 250,000 | 2 | 1.5 | 2.3GHz | - |
M1.LARGE32 | 4 | 32 | - | 350,000 | 4 | 1.5 | 2.3GHz | - |
M1.2XLARGE64 | 8 | 64 | - | 0.7 million | 8 | 2.0 | 2.3GHz | - |
M1.3XLARGE96 | 12 | 96 | - | 1 million | 8 | 2.5 | 2.3GHz | - |
M1.4XLARGE128 | 16 | 128 | - | 1.4 million | 8 | 3.5 | 2.3GHz | - |
M1.6XLARGE192 | 24 | 192 | - | 2,000,000 | 8 | 5.0 | 2.3GHz | - |
M1.8XLARGE256 | 32 | 256 | - | 2,700,000 | 8 | 7.0 | 2.3GHz | - |
M1.12XLARGE368 | 48 | 368 | - | 4,000,000 | 8 | 10.0 | 2.3GHz | - |
Note:
In the M1 instance specification list, a "-" in the network packet transmission and reception column indicates that there is no fixed commitment for network packet transmission and reception for that specification. If you have a strong demand for consistent performance, it is recommended to choose a model with a performance commitment.
High I/O Family
This family features high random IOPS, high throughput and low latency, making it suitable for I/O-intensive applications that require high disk read/write performance and low latency, such as high-performance databases.
Note:
The data disks of IT3 and IT5 instances are local storage, which are at risk of data loss (e.g., during host downtime). If your application cannot ensure data reliability, it is not recommended to use IT3 or IT5 instances. We strongly advise using instances that allow you to choose cloud disks as data disks.
High IO IT5
High I/O IT5 instances are the latest generation of high I/O instances designed specifically for I/O-intensive workloads. Based on NVMe SSD instance storage, they provide low-latency, ultra-high IOPS, and high-throughput storage resources, offering up to 2.05 million IOPS and up to 11GB/s throughput capacity at a relatively low cost. They are suitable for high-performance relational databases, Elasticsearch, and other I/O-intensive businesses.
Use Cases
High-performance databases, NoSQL databases (e.g. MongoDB), and clustered databases
I/O intensive applications that require low latency, such as online transaction processing (OLTP) systems, and Elasticsearch
Features
2.5 GHz Intel® Xeon® Cascade Lake processor, offering stable computing performance
The latest generation of 6-channel DDR4 memory
Supports up to 23 Gbps of private network bandwidth to meet extremely high private network transmission requirements
NVMe SSD is used for instance storage, offering low latency and ultra-high IOPS
Up to 650,000 random read IOPS (block size of 4 KB), and up to 2.8 GB/s sequential read throughput (block size of 128 KB) for a single disk
Up to 2.05 million random read IOPS (block size of 4 KB), and up to 11 GB/s sequential read throughput (block size of 128 KB) for a CVM
Requirements
If the CVM instance crashes, we will notify you and perform fixes.
IT5 instances without monitoring components installed may result in the platform being unable to perform more detailed monitoring. If an instance encounters a failure, proper notifications may not be sent, posing a high risk. Please refer to Installing CVM Monitoring Components to complete the installation of monitoring components.
IT5 instances are pay-as-you-go.
Supports only VPC.
IT5 instances do not support configuration adjustment and the "No Charge when Shut Down" feature.
See below for available IT5 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
IT5.4XLARGE64 | 16 | 64 | 1.5 million | 1.5 million | 4 | 6 | 2.5GHz | 1 x 3570 GB local NVMe SSD |
IT5.8XLARGE128 | 32 | 128 | 2.5 million | 300K | 8 | 12 | 2.5GHz | 2 x 3570 GB local NVMe SSD |
IT5.16XLARGE256 | 64 | 256 | 5,000,000 | 6 million | 16 | 23 | 2.5GHz | 4 x 3570 GB local NVMe SSD |
High I/O IT3
High I/O IT3 instances are the latest generation of high I/O instances designed specifically for I/O-intensive workloads. Based on NVMe SSD instance storage, they provide low-latency, extremely high IOPS, and high-throughput storage resources, offering up to 1.8 million IOPS and up to 11 GB/s throughput capacity at a relatively low cost. They are suitable for high-performance relational databases, Elasticsearch, and other I/O-intensive businesses.
IT3 instances feature inexpensive and intensive storage. They can provide each CPU core with higher IOPS and larger private network bandwidth.
Use Cases
High-performance databases, NoSQL databases (e.g. MongoDB), and clustered databases
I/O intensive applications that require low latency, such as online transaction processing (OLTP) systems, and Elasticsearch
Features
2.5 GHz Intel Xeon® Skylake 6133 processors with stable computing performance
The latest generation of 6-channel DDR4 memory
Supports up to 23 Gbps of private network bandwidth to meet extremely high private network transmission requirements
NVMe SSD is used for instance storage, offering low latency and ultra-high IOPS
Up to 620,000 random read IOPS (block size of 4 KB), and up to 2.9 GB/s sequential read throughput (block size of 128 KB) for a single disk
Up to 1.8 million random read IOPS (block size of 4 KB), and up to 11 GB/s sequential read throughput (block size of 128 KB) for a CVM
Requirements
If the CVM instance crashes, we will notify you and perform fixes.
IT3 instances without monitoring components installed may result in the platform being unable to perform more detailed monitoring. If an instance encounters a failure, proper notifications may not be sent, posing a high risk. Please refer to Installing CVM Monitoring Components to complete the installation of monitoring components.
IT3 instances are pay-as-you-go.
Supports only VPC.
IT3 instances do not support configuration adjustment and the no charges when shut down feature.
See below for available IT3 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
IT3.4XLARGE64 | 16 | 64 | 1.5 million | 4 | 6 | 2.5GHz | 1 × 3720 GB local NVMe SSD |
IT3.8XLARGE128 | 32 | 128 | 2.5 million | 8 | 12 | 2.5GHz | 2 × 3720 GB local NVMe SSDs |
IT3.16XLARGE256 | 64 | 256 | 5,000,000 | 16 | 23 | 2.5GHz | 4 × 3720 GB local NVMe SSD |
Big Data Family
The big data family is equipped with massive storage resources, features high throughput, and is suitable for throughput-intensive applications such as Hadoop distributed computing, massive log processing, distributed file systems, and large data warehouses.
Note:
D3 and D2 instances use local disk as the data disk, which may lose data (e.g., when the host crashes). If your application cannot guarantee data reliability, we recommend you choose an instance that can use cloud disks as the data disk.
D3
As the latest generation of big data instances, D3 instances are equipped with massive storage resources, and carry up to 94 TB SATA HDD local storage, making them suitable for throughput-intensive services, such as Hadoop distributed computing and parallel data processing.
Use Cases
Distributed computing services such as Hadoop MapReduce/HDFS/Hive/HBase
Workloads such as Elasticsearch, log processing, and large data warehouse
Customers in the Internet, finance, and industries that require big data computing and storage analysis, as well as workloads that require massive data storage and computing
Features
2.5 GHz Intel® Xeon® Cascade Lake processor with DDR4 memory
Instances carry up to 24 local disks of 4 TB and up to 94 TB of HDD local storage.
At least 190 MB/s sequential read throughput performance for a single disk
Millisecond read/write latency
Processor to memory ratio 1:4, tailored for the big data scenarios
Requirements
If the CVM instance crashes, we will notify you and perform fixes.
D3 instances without monitoring components installed may result in the platform being unable to perform more detailed monitoring. If an instance encounters a failure, proper notifications may not be sent, posing a high risk. Please refer to Installing CVM Monitoring Components to complete the installation of monitoring components.
Big Data D3 instances are pay-as-you-go.
Supports only VPC.
The configuration of D3 instances cannot be changed.
See below for available D3 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
D3.2XLARGE32 | 8 | 32 | 8 million | 250,000 | 2 | 4.0 | 2.5GHz | 4 x 3720 GB local SATA HDDs |
D3.4XLARGE64 | 16 | 64 | 1.5 million | 300,000 | 4 | 7.0 | 2.5GHz | 8 x 3720 GB local SATA HDDs |
D3.8XLARGE128 | 32 | 128 | 2.5 million | 0.6 million | 8 | 14.0 | 2.5GHz | 12 x 3720 GB local SATA HDDs |
D3.16XLARGE256 | 64 | 256 | 5,000,000 | 1.2 million | 12 | 27.0 | 2.5GHz | 24 x 3720 GB local SATA HDDs |
D3.21XLARGE320 | 84 | 320 | 6 million | 1,600,000 | 16 | 32.0 | 2.5GHz | 24 x 3720 GB local SATA HDDs |
Big Data D2
D2 instances are equipped with massive storage resources, and carry up to 144 TB SATA HDD local storage, making them suitable for throughput-intensive services, such as Hadoop distributed computing and parallel data processing.
Use Cases
Distributed computing services such as Hadoop MapReduce/HDFS/Hive/HBase
Workloads such as Elasticsearch, log processing, and large data warehouse
Customers in the Internet, finance, and industries that require big data computing and storage analysis, as well as workloads that require massive data storage and computing
Features
2.4 GHz Intel® Xeon® Skylake 6148 processor with DDR4 memory
Instances carry up to 12 local disks of 12 TB and up to 144 TB of HDD local storage
For a single disk, the sequential read throughput is 220+ MB/s and sequential write throughput is 220+ MB/s (128 KB block size and depth of 32)
Overall instance throughput can reach up to 2.8 GB/s (128 KB block size and depth of 32)
Read/write latency is as low as 2-5 ms
Higher specification, D2.19XLARGE320, offering 76 vCPUs and 320 GB of memory
Local storage with a unit price as low as 1/10 of S2, making its total cost close to that of a self-built Hadoop clusters in IDC
Processor to memory ratio 1:4, tailored for the big data scenarios
Requirements
If the CVM instance crashes, we will notify you and perform fixes.
Big Data D2 instances are pay-as-you-go.
Supports only VPC.
The configuration of D2 instances cannot be changed.
See below for available D2 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
D2.2XLARGE32 | 8 | 32 | 8 million | 2 | 3.0 | 2.4GHz | Equipped with one 11176 GB SATA HDD local disk |
D2.4XLARGE64 | 16 | 64 | 1.5 million | 4 | 6.0 | 2.4GHz | Equipped with two 11176 GB SATA HDD local disks |
D2.6XLARGE96 | 24 | 96 | 2,000,000 | 6 | 8.0 | 2.4GHz | Equipped with three 11176 GB SATA HDD local disks |
D2.8XLARGE128 | 32 | 128 | 2.5 million | 8 | 11.0 | 2.4GHz | Equipped with four 11176 GB SATA HDD local disks |
D2.16XLARGE256 | 64 | 256 | 5,000,000 | 16 | 22.0 | 2.4GHz | Equipped with eight 11176 GB SATA HDD local disks |
D2.19XLARGE320 | 76 | 320 | 6 million | 16 | 25.0 | 2.4GHz | Equipped with twelve 11176 GB SATA HDD local disks |
Computing Instance Family
Compute-optimized instances offer CPU frequencies of up to 3.8GHz, providing the highest single-core computing performance. They are suitable for compute-intensive applications such as batch processing, high-performance computing, and large-scale game servers.
C6 for computing tasks
The C6 Compute instances are the latest generation of compute instances, providing stable, reliable, and high-performance computing services based on a new-generation virtualization platform. They offer greater bandwidth and lower latency. With the highest base frequency processors and the best price-performance ratio in CVM, they are the ideal choice for applications with high computational performance and high concurrent read/write requirements that are constrained by computing limitations.
C6 instances come with new Intel® Xeon® processors, and support up to 100 Gbps of private network bandwidth.
Use Cases
These instances are ideal for:
Batch processing workloads and high performance computing (HPC)
High-traffic Web frontend server
Other compute-intensive services such as massively multiplayer online (MMO) game servers
Features
A new-generation Star Lake two-way server developed by Tencent Cloud, with the 3rd generation Intel® Xeon® scalable processor
Intel® Xeon® Ice Lake processor, with a CPU clock rate of 3.2 GHz and a turbo frequency of 3.5 GHz
Offers various processor-to-memory ratios, such as 1:2 and 1:4.
Up to 100 Gbps private network bandwidth for ultra high packet throughput to meet extremely high private network transfer requirements
Requirements
C6 instances are pay-as-you-go.
We recommend you use the instance together with the TencentOS Server operating system to enjoy the best application performance.
C6 instances are available for purchase with various configurations. Please refer to the instance specifications below. Ensure that the C6 instance size you choose meets the minimum CPU and memory requirements for your operating system and applications. In many use cases, operating systems with resource-intensive graphical user interfaces (such as Windows) may require larger instance sizes. As your workload's demand for memory and CPU increases over time, you can scale up to higher configurations or choose other instance types.
Instances can achieve up to 100 Gbps network bandwidth, depending on the support of the instance's operating system kernel version and runtime environment. When the PPS exceeds 10 million and the bandwidth is greater than 50 Gbps, the kernel protocol stack may cause significant network performance degradation. In this case, the netperf test bandwidth value may not meet expectations. You can use the DPDK method to mask the differences in the cloud server kernel protocol stack and obtain the true network performance of the instance. For testing methods, please refer to High-Throughput Network Performance Testing.
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
C6.LARGE8 | 4 | 8 | 8 million | 0.6 million | 4 | 5 | 3.2GHz | - |
C6.LARGE16 | 4 | 16 | 8 million | 0.6 million | 4 | 5 | 3.2GHz | - |
C6.2XLARGE16 | 8 | 16 | 1,600,000 | 1,300,000 | 8 | 9 | 3.2GHz | - |
C6.2XLARGE32 | 8 | 32 | 1,600,000 | 1,300,000 | 8 | 9 | 3.2GHz | - |
C6.4XLARGE32 | 16 | 32 | 3,300,000 | 2,700,000 | 16 | 18 | 3.2GHz | - |
C6.4XLARGE64 | 16 | 64 | 3,300,000 | 2,700,000 | 16 | 18 | 3.2GHz | - |
C6.8XLARGE128 | 32 | 128 | 6,600,000 | 5,500,000 | 32 | 35 | 3.2GHz | - |
C6.23MEDIUM216 | 46 | 216 | 9.5 million | 8 million | 46 | 50 | 3.2GHz | - |
C6.16XLARGE256 | 64 | 256 | 13,000,000 | 11,000,000 | 48 | 70 | 3.2GHz | - |
C6.20XLARGE320 | 80 | 320 | 16,000,000 | 13,900,000 | 48 | 87 | 3.2GHz | - |
C6.23XLARGE432 | 92 | 432 | 19 million | 16,000,000 | 48 | 100 | 3.2GHz | - |
C5 for computing tasks
Compute-optimized C5 instances are the latest generation of compute instances, providing stable and reliable high-performance computing services based on a cutting-edge virtualization platform. They offer greater bandwidth and lower latency. With the highest base frequency processors and the best price-performance ratio in CVM, they are the ideal choice for compute-intensive and high-concurrency read-write applications that are constrained by computing limitations.
C5 instances come with the latest Xeon® Cooper Lake processors.
Features
Intel® Xeon® Cooper Lake processor, with a base frequency of 3.4 GHz and turbo frequency of 3.8 GHz
New Intel Brain Floating Point 16-bit (bfloat16) instruction set
1:2 or 1:4 processor to memory ratio
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Use Cases
These instances are ideal for:
Batch processing workloads and high performance computing (HPC)
High-traffic Web frontend server
Other compute-intensive services such as massively multiplayer online (MMO) game servers
Requirements
C5 instances are pay-as-you-go.
The network of C5 instances must be based on a VPC.
See below for available C5 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
C5.LARGE8 | 4 | 8 | 0.5 million | 250,000 | 4 | 1.5 | 3.4GHz | - |
C5.LARGE16 | 4 | 16 | 0.5 million | 250,000 | 4 | 1.5 | 3.4GHz | - |
C5.2XLARGE16 | 8 | 16 | 8 million | 250,000 | 8 | 3.0 | 3.4GHz | - |
C5.2XLARGE32 | 8 | 32 | 8 million | 250,000 | 8 | 3.0 | 3.4GHz | - |
C5.4XLARGE32 | 16 | 32 | 1.5 million | 300,000 | 8 | 6.0 | 3.4GHz | - |
C5.4XLARGE64 | 16 | 64 | 1.5 million | 300,000 | 8 | 6.0 | 3.4GHz | - |
C5.8XLARGE64 | 32 | 64 | 2.5 million | 0.6 million | 16 | 12.0 | 3.4GHz | - |
C5.8XLARGE128 | 32 | 128 | 2.5 million | 0.6 million | 16 | 12.0 | 3.4GHz | - |
C5.12XLARGE96 | 48 | 96 | 4,000,000 | 900,000 | 24 | 18.0 | 3.4GHz | - |
C5.12XLARGE192 | 48 | 192 | 4,000,000 | 900,000 | 24 | 18.0 | 3.4GHz | - |
C5.13XLARGE184 | 52 | 184 | 4,000,000 | 1 million | 32 | 19.0 | 3.4GHz | - |
C5.16XLARGE256 | 64 | 256 | 5,000,000 | 1.2 million | 32 | 24.0 | 3.4GHz | - |
C5.26XLARGE368 | 104 | 368 | 6 million | 2,000,000 | 32 | 36.0 | 3.4GHz | - |
Compute Optimized C4
Compute-optimized C4 instances are the next-generation compute instances with up to 25 Gbps internal network bandwidth, offering greater bandwidth and lower latency. They provide the highest base frequency processors and the best price-performance ratio in CVM, making them an ideal choice for applications with high computational performance and high concurrent read/write operations that are compute-bound.
C4 instances are equipped with the latest Xeon® Cascade Lake processors, and support up to 25 Gbps of private network bandwidth.
Use Cases
These instances are ideal for:
Batch processing workloads and high performance computing (HPC)
High-traffic Web frontend server
Other compute-intensive services such as massively multiplayer online (MMO) game servers
Features
3.2 GHz Intel® Xeon® Cascade Lake processor, with a turbo frequency 3.7 GHz
New Intel Advanced Vector Extension (AVX-512) instruction set
Supports up to 25 Gbps of private network bandwidth to meet extremely high private network transmission requirements
1:2 or 1:4 processor to memory ratio
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Requirements
Supports pay-as-you-go.
C4 instances can be launched in classic networks and VPCs.
See below for available C4 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
C4.LARGE8 | 4 | 8 | 0.6 million | 350,000 | 4 | 2.0 | 3.2GHz | - |
C4.LARGE16 | 4 | 16 | 0.6 million | 350,000 | 4 | 2.0 | 3.2GHz | - |
C4.2XLARGE16 | 8 | 16 | 1 million | 0.7 million | 8 | 4.0 | 3.2GHz | - |
C4.2XLARGE32 | 8 | 32 | 1 million | 0.7 million | 8 | 4.0 | 3.2GHz | - |
C4.4XLARGE64 | 16 | 64 | 1 million | 1,300,000 | 16 | 7.0 | 3.2GHz | - |
C4.8XLARGE174 | 32 | 174 | 1 million | 2.5 million | 16 | 13.0 | 3.2GHz | - |
C4.16XLARGE348 | 64 | 348 | 1 million | 5,000,000 | 16 | 25.0 | 3.2GHz | - |
Compute Network-optimized CN3
The CN3 Compute Optimized instances are the next-generation compute-optimized instances, with a maximum internal network bandwidth of up to 25 Gbps, offering greater bandwidth and lower latency. They provide the highest base clock frequency processors and the best price-performance ratio in CVM, making them an ideal choice for compute-intensive and high-concurrency read/write applications with computational constraints.
CN3 instances are equipped with the latest Xeon® Skylake processors, and support up to 25 Gbps of private network bandwidth, which is 2.5 times greater than Compute C3.
Use Cases
These instances are ideal for:
Batch processing workloads and high performance computing (HPC)
High-traffic Web frontend server
Other compute-intensive services such as massively multiplayer online (MMO) game servers
Features
3.2 GHz Intel® Xeon® Skylake 6146 processor
New Intel Advanced Vector Extension (AVX-512) instruction set
Supports up to 25 Gbps of private network bandwidth to meet extremely high private network transmission requirements
The latest generation of 6-channel DDR4 memory with a bandwidth of 2,666 MT/s
1:2 or 1:4 processor to memory ratio
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Requirements
CN3 instances are pay-as-you-go.
Supports both classic network and VPC.
See below for available CN3 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
CN3.LARGE8 | 4 | 8 | 0.6 million | 350,000 | 4 | 3.0 | 3.2GHz | - |
CN3.LARGE16 | 4 | 16 | 0.6 million | 350,000 | 4 | 3.0 | 3.2GHz | - |
CN3.2XLARGE16 | 8 | 16 | 1 million | 0.7 million | 8 | 5.0 | 3.2GHz | - |
CN3.2XLARGE32 | 8 | 32 | 1 million | 0.7 million | 8 | 5.0 | 3.2GHz | - |
CN3.4XLARGE32 | 16 | 32 | 1 million | 1.4 million | 16 | 9.0 | 3.2GHz | - |
CN3.4XLARGE64 | 16 | 64 | 1 million | 1.4 million | 16 | 9.0 | 3.2GHz | - |
CN3.8XLARGE64 | 32 | 64 | 1 million | 2,700,000 | 16 | 17.0 | 3.2GHz | - |
CN3.8XLARGE128 | 32 | 128 | 1 million | 270 | 16 | 17.0 | 3.2GHz | - |
CN3.12XLARGE160 | 48 | 160 | 1 million | 4,000,000 | 16 | 25.0 | 3.2GHz | - |
C3 for computing tasks
C3 instances have the most powerful processors and the best cost performance among CVMs. They are ideal for applications that require high computing performance and high-concurrence reads/writes.
C3 instances are equipped with new Xeon® Skylake processors with the latest DDR4 memory, offering larger private network bandwidth.
Use Cases
These instances are ideal for:
Batch processing workloads and high performance computing (HPC)
High-traffic Web frontend server
Other compute-intensive services such as massively multiplayer online (MMO) game servers
Features
3.2 GHz Intel® Xeon® Skylake 6146 processor
New Intel Advanced Vector Extension (AVX-512) instruction set
The latest generation of 6-channel DDR4 memory with a bandwidth of 2,666 MT/s
1:2 or 1:4 processor to memory ratio
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Requirements
C3 instances are pay-as-you-go.
Supports both classic network and VPC.
See below for available C3 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
C3.LARGE8 | 4 | 8 | 0.6 million | 350,000 | 2 | 2.5 | 3.2GHz | - |
C3.LARGE16 | 4 | 16 | 0.6 million | 350,000 | 2 | 2.5 | 3.2GHz | - |
C3.2XLARGE16 | 8 | 16 | 1 million | 0.7 million | 2 | 3.0 | 3.2GHz | - |
C3.2XLARGE32 | 8 | 32 | 1 million | 0.7 million | 2 | 3.0 | 3.2GHz | - |
C3.4XLARGE32 | 16 | 32 | 1 million | 1.4 million | 4 | 4.5 | 3.2GHz | - |
C3.4XLARGE64 | 16 | 64 | 1 million | 1.4 million | 4 | 4.5 | 3.2GHz | - |
C3.8XLARGE64 | 32 | 64 | 1 million | 2,700,000 | 8 | 8.0 | 3.2GHz | - |
C3.8XLARGE128 | 32 | 128 | 1 million | 2,700,000 | 8 | 8.0 | 3.2GHz | - |
C2 for computing tasks
C2 instances are optimized with high-performance processors and deliver high cost performance. They are ideal for applications that require high computing performance and high-concurrence reads/writes.
Use Cases
These instances are ideal for:
Batch processing workloads
High-traffic Web servers, and massively multiplayer online (MMO) game servers
High-performance computing (HPC) and other compute-intensive applications.
Features
3.2 GHz Intel® Xeon® E5-2667v4 processor with a turbo frequency of up to 3.6 GHz and DDR4 memory
1:2 or 1:4 processor to memory ratio
The network performance of an instance depends on its specification. A higher specification means a greater network forwarding performance and a higher private network bandwidth cap.
Requirements
C2 instances are pay-as-you-go, and can also be used as production instances of computing host HC20 in CDHs.
Supports both classic network and VPC.
See below for available C2 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of connections | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
C2.LARGE8 | 4 | 8 | 550,000 | 400,000 | 4 | 2.5 | 3.2GHz | - |
C2.LARGE16 | 4 | 16 | 550,000 | 400,000 | 4 | 2.5 | 3.2GHz | - |
C2.LARGE32 | 4 | 32 | 550,000 | 400,000 | 4 | 2.5 | 3.2GHz | - |
C2.2XLARGE16 | 8 | 16 | 550,000 | 8 million | 8 | 3.5 | 3.2GHz | - |
C2.2XLARGE32 | 8 | 32 | 550,000 | 8 million | 8 | 3.5 | 3.2GHz | - |
C2.4XLARGE32 | 16 | 32 | 850,000 | 1.5 million | 8 | 6.0 | 3.2GHz | - |
C2.4XLARGE64 | 16 | 64 | 850,000 | 1.5 million | 8 | 6.0 | 3.2GHz | - |
C2.8XLARGE96 | 32 | 96 | 850,000 | 300K | 8 | 10.0 | 3.2GHz | - |
Heterogeneous Computing Instance Family
This family is equipped with heterogeneous hardware such as GPU and FPGA to deliver real-time, fast parallel computing and floating-point computing capabilities. It is suitable for high-performance applications such as deep learning, scientific computing, video encoding/decoding, and graphics workstations.
AMD GPU series instances utilize AMD FirePro S7150, offering outstanding graphics processing capabilities. They are the top choice for application scenarios such as remote desktops, 3D rendering, and cloud gaming. AMD GPU series instances >
NVIDIA GPU series instances utilize NVIDIA Tesla series GPUs, including the previous generation M40, the current mainstream P4/P40, and the latest generation V100. They offer outstanding general computing capabilities, making them the top choice for deep learning training/inference, scientific computing, and other application scenarios. NVIDIA GPU series instances >
Batch-based Instance Family
With the lowest per core-hour cost, this family is suitable for compute-intensive applications that frequently deploy ultra-large amounts of computing nodes for short periods of time. Examples include rendering, genetic analysis, and solid-state pharmaceutical research.
Note:
This instance type is currently only available to whitelisted users. Contact your account manager if you wish to purchase these instance type.
Batch-based Computing BC1
Batch Compute BC1 instances are cost-effective instances billed by core-hour, with billing accuracy down to the second and prices as low as 0.1 CNY/core-hour. They offer flexibility with immediate availability and termination upon completion, supporting various specifications to meet the demands of compute-intensive users requiring frequent short-term usage of large-scale compute nodes, such as rendering, gene analysis, and crystallography.
Use Cases
Video/Film rendering
Genetic analysis, solid-state pharmaceutical research etc.
HPC compute-intensive applications such as weather forecasting and astronomy
Features
Cost-effective with the lowest price per hour among all instances with the same specification
CPU/RAM ratio: 1:4
Requirements
Batch-based computing BC1 instances are pay-as-you-go.
BC1 instances can be launched in classic networks and VPCs.
See below for available BC1 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
BC1.LARGE16 | 4 | 16 | - | 4 | 1.5 | - | - |
BC1.3XLARGE48 | 12 | 48 | - | 8 | 2.5 | - | - |
BC1.6XLARGE96 | 24 | 96 | - | 8 | 5.0 | - | - |
Note:
In the BC1 instance specification list, a "-" in the network packet transmission and CPU frequency columns indicates that there is no fixed commitment to network packet transmission and processor frequency performance for that specification. If you have a strong demand for consistent performance, it is recommended to choose a model with performance commitments.
Batch-based Computing BS1
Batch computing BS1 instances are cost-effective instances billed by core-hour, with billing accuracy down to the second and prices as low as 0.09 CNY/core-hour. They offer flexibility with immediate availability and termination upon completion, supporting various specifications to meet the demands of compute-intensive users requiring frequent short-term usage of large-scale computing nodes, such as rendering, gene analysis, and crystallography.
Features
Cost-effective with the lowest price per hour among all instances with the same specification
CPU/RAM ratio: 1:2
Use Cases
Video/Film rendering
Genetic analysis, solid-state pharmaceutical research etc.
HPC compute-intensive applications such as weather forecasting and astronomy
Requirements
Batch-based computing BS1 instances are pay-as-you-go.
BS1 instances can be launched in classic networks and VPCs.
See below for available BS1 specifications:
Specifications | vCPU | Memory (GB) | Throughput (pps) (received and sent packets) | Number of Queues | Private network bandwidth (Gbps) (outbound and inbound) | Clock Rate | Remarks |
BS1.LARGE8 | 4 | 8 | - | 4 | 1.5 | - | - |
BS1.3XLARGE24 | 12 | 24 | - | 8 | 2.5 | - | - |
BS1.6XLARGE48 | 24 | 48 | - | 8 | 5.0 | - | - |
Note:
In the BS1 Instance Specification list, a "-" in the Network Packets and CPU Frequency columns indicates that there is no fixed commitment to network packet transmission and processor frequency performance for that specification. If you have a strong demand for consistent performance, it is recommended to choose a model with performance commitments.
Cloud Bare Metal
Tencent Cloud Bare Metal (CBM) is a bare metal cloud service that combines the elasticity of cloud servers with the performance of physical machines. It seamlessly integrates with Tencent Cloud's full range of products (such as networking, storage, and databases) to provide you with high-performance, securely isolated physical server clusters in the cloud. Your business applications can directly access the processor and memory of the bare metal cloud server without any virtualization overhead. With this service, you can elastically scale the number of physical servers based on your business needs, reducing the time to acquire physical servers to minutes. You can entrust capacity management and operations to Tencent Cloud's professional team, allowing you to focus on business innovation.
HPC Cluster
High-performance computing clusters use bare-metal cloud servers as nodes, interconnected via RDMA, providing high-bandwidth and ultra-low-latency network services. These clusters can meet the parallel computing requirements of large-scale high-performance computing, artificial intelligence, and big data recommendation applications.
For instance specifications, please refer to High-Performance Computing Cluster Instance Specifications.