Each specification of elastic cloud servers is provided by different hardware devices, and comes with different performance and pricing. Based on the business scenario and configurations such as vCPU, memory, network performance, and storage throughput, eSurfing Cloud provides a variety of instance types for cloud servers. Each type includes multiple instance specifications.
The instance specification name format is x<Specification Type>x<Generation>.xxxx<vCPU Count>.x<Memory Ratio to vCPU>.
<Specification Type>: Such as s, m, c, k, h, f, ip, d, etc., representing the type of cloud server specification. Where, s stands for general-purpose instances, m for memory-optimized instances, c for Computing-plus instances, k represents Kunpeng (a brand of ARM-based processors), h represents Hygon (a Chinese CPU manufacturer), f represents Phytium (another brand of ARM-based processors), ip represents instances with ultra-high I/O local cloud server, and d represents instances with enhanced disk performance.
<Generation>: Such as 2, 6, etc., represents different hardware specifications, with larger numbers generally having higher performance.
<vCPU Count>: Such as small, medium, large, xlarge, etc., represents the number of vCPU cores. Where, "medium" corresponds to 1 vCPU, "large" to 2 vCPUs, "xlarge" to 4 vCPUs, and "small" specifically refers to a small specification of 1 vCPU with 1G of memory.
<Memory Ratio to vCPU>: Such as 1, 2, 4, 8, etc., represents the ratio of memory in GB to the number of vCPUs.
For example, s2.2xlarge.1 indicates a general-purpose, 2nd generation cloud server with 8 vCPU cores and 8GB of memory.
It is recommended that you first understand the product specifications, performance, and usage restrictions, and then make a selection according to your actual needs. After you have completed the selection process and started using the cloud server, it is recommended that you verify whether the chosen instance specification is appropriate based on the performance monitoring information over a period of time. For example: If monitoring reveals that the instance CPU usage is consistently low, it is recommended that you log in to the instance to check if the memory usage is high. If the memory usage is high, you can adjust to a specification family that has a more suitable ratio of processor to memory resources.
The specifications and indicators are described as follows:
Metric Name | Description |
Network Packets Received and Transmitted | The network packet transmission capability refers to the sum of the maximum capabilities for both outgoing and incoming directions. For the testing method of network packet transmission PPS performance, please refer to the network performance testing methods. |
Network Bandwidth | The network bandwidth capability refers to the maximum capacity that can be achieved by combining both the outgoing and incoming directions. For the testing method of network bandwidth performance, please refer to theNetwork Performance Testing Methods. |
Base Bandwidth | The highest speed at which the elastic cloud server can stably transmit data over the network. |
Maximum Bandwidth | The maximum internal network bandwidth that an elastic cloud server can achieve within a certain period depends on the network resource usage of other virtual machines on the physical machine where the elastic cloud server is located. |
Number of Connections | The maximum number of client requests that an elastic cloud server can accommodate at the same time, including both actively and passively established connections. |
NIC Multi-Queue | The maximum number of NIC queues supported by the primary NIC, covering both the receive and transmit directions. By enabling the multi-queue feature of the NIC, you can distribute the NIC interrupts to different CPUs for processing within the Elastic Cloud Server, to meet the demands of the NIC, thereby enhancing the network PPS (Packets Per Second) and bandwidth performance. |
ENI | The maximum number of elastic NICs supported by elastic cloud servers. An elastic NIC is a virtual network interface within a Virtual Private Cloud (VPC), used to connect ECS (Elastic Compute Service) with a private network. Including the primary and secondary NICs, the primary NIC cannot be created independently nor can it be detached from the instance. |
Local disk | The number of local disks * The capacity of each individual disk. Local disks are the local hard disk devices on the physical machine where the cloud server is located. Local disks provide local storage access capabilities for cloud servers, offering the advantages of low latency, high random IOPS, high throughput, and cost-effectiveness. |
GPU | GPU Quantity*GPU Type. A GPU, or Graphics Processing Unit, is responsible for tasks such as graphic rendering, image processing, and computational acceleration. GPUs have more processing units and higher parallel processing capabilities compared to CPUs, making them suitable for scenarios like large-scale parallel computing. |