Service Overview - Huawei · Auto Scaling Automatic adjustment of computing resources Dynamic...

Elastic Cloud Server

Service Overview

Date 2020-07-06

Contents

1 What Is ECS?............................................................................................................................. 1

2 ECS Advantages........................................................................................................................4

3 ECS Application Scenarios..................................................................................................... 6

4 Notes on Using ECSs............................................................................................................... 8

5 Instances..................................................................................................................................105.1 Overview.................................................................................................................................................................................. 105.2 ECS Lifecycle........................................................................................................................................................................... 105.3 ECS Types................................................................................................................................................................................. 125.4 Kunpeng ECS Specifications ............................................................................................................................................. 135.5 x86 ECS Specifications ........................................................................................................................................................145.6 Kunpeng General Computing-plus ECSs....................................................................................................................... 295.7 General Computing ECSs....................................................................................................................................................315.8 General Computing-plus ECSs.......................................................................................................................................... 335.9 General Computing-Basic ECSs........................................................................................................................................ 385.10 Memory-optimized ECSs.................................................................................................................................................. 405.11 Large-Memory ECSs...........................................................................................................................................................435.12 Disk-intensive ECSs............................................................................................................................................................ 445.13 Ultra-high I/O ECSs............................................................................................................................................................ 475.14 High-Performance Computing ECSs............................................................................................................................ 505.15 GPU-accelerated ECSs....................................................................................................................................................... 52

6 Images......................................................................................................................................61

7 EVS Disks................................................................................................................................. 63

8 Network...................................................................................................................................64

9 Security.................................................................................................................................... 679.1 User Encryption..................................................................................................................................................................... 679.2 Cloud-Init................................................................................................................................................................................. 69

10 Billing..................................................................................................................................... 71

11 CPU Credits........................................................................................................................... 77

12 Region and AZ......................................................................................................................81

Elastic Cloud ServerService Overview Contents

2020-07-06 ii

13 ECS and Other Services......................................................................................................83

14 Change History.................................................................................................................... 86

Elastic Cloud ServerService Overview Contents

2020-07-06 iii

1 What Is ECS?

An Elastic Cloud Server (ECS) is a basic computing unit that consists of vCPUs,memory, OS, and Elastic Volume Service (EVS) disks. After creating an ECS, youcan use it like using your local computer or physical server.

ECSs support self-service creation, modification, and operation. You can create anECS by specifying its vCPUs, memory, OS, and login authentication. After the ECSis created, you can modify its specifications as required. This ensures a reliable,secure, efficient computing environment.

Why ECS?● Rich specifications: A variety of ECS types are available for different scenario

requirements. There are multiple customizable specifications for each type.● Comprehensive images: Public, private, and shared images can be flexibly

selected to request for ECSs.● Differentiated EVS disks: Common I/O, high I/O, ultra-high I/O, and general

purpose SSD disks are available for all of your service requirements.● Flexible billing modes: Yearly/Monthly and pay-per-use billing modes allow

you to purchase and release resources at any time based on servicefluctuation.

● Reliable data: Scalable, reliable high-throughput virtual block storage is basedon distributed architecture.

● Security protection: The network is isolated and protected using securitygroup rules from viruses and Trojan horses. Security services, such as Anti-DDoS, Web Application Firewall (WAF), and Vulnerability Scan Service (VSS)are included to further enhance ECS security.

● Flexible, easy-to-use: Elastic computing resources are automatically adjustedbased on service requirements and policies to efficiently meet servicerequirements.

● Highly efficient O&M: Multi-choice management via the managementconsole, remote access, and APIs with full rights.

● In-cloud monitoring: Cloud Eye samples monitored metrics in real time,correctly generates resource monitoring alarms, and sends notifications torelated personnel immediately after the alarms are triggered.

● Load balancing: Elastic Load Balance (ELB) automatically distributes accesstraffic to multiple ECSs to balance their service load. It enables higher levels

Elastic Cloud ServerService Overview 1 What Is ECS?

2020-07-06 1

of fault tolerance in your applications and expands application servicecapabilities.

For more details, see 2 ECS Advantages and 3 ECS Application Scenarios.

System Architecture

ECS works with other products and services to provide computing, storage,network, and image installation functions.

● ECSs are deployed in multiple availability zones (AZs) connected with eachother through an internal network. If an AZ becomes faulty, other AZs in thesame region will not be affected.

● With the Virtual Private Cloud (VPC) service, you can build a dedicatednetwork, set the subnet and security group, and allow the VPC tocommunicate with the external network through an EIP (bandwidth supportrequired).

● With the Image Management Service (IMS), you can install images on ECSs,or create ECSs using private images and deploy services quickly.

● The Elastic Volume Service (EVS) provides storage and Volume Backup Service(VBS) provides data backup and recovery functions.

● Cloud Eye is a key measure to ensure ECS performance, reliability, andavailability. Using Cloud Eye, you can determine ECS resource usage.

● Cloud Backup and Recovery (CBR) backs up data for EVS disks and ECSs, anduses snapshot backups to restore the EVS disks and ECSs.

Figure 1-1 System architecture


2020-07-06 2

Access MethodsThe public cloud provides a web-based service management platform. You canaccess ECSs through HTTPS-compliant application programming interfaces (APIs)or the management console. These two access methods differ as follows:● Accessing ECSs through APIs

Use this method if you are required to integrate the ECSs on the public cloudplatform into a third-party system for secondary development. For detailedoperations, see Elastic Cloud Server API Reference.

● Accessing ECSs through the management consoleUse this method if you are not required to integrate ECSs with a third-partysystem.After registering on the public cloud, log in to the management console andclick Elastic Cloud Server under Computing on the homepage.


2020-07-06 3

https://support.huaweicloud.com/intl/en-us/api-ecs/ecs.html

2 ECS Advantages

ECS supports automatic adjustment of computing resources based on servicerequirements and scaling policies. You can customize ECS configurations asneeded, including vCPUs, memory, and bandwidth for secure, flexible, and efficientapplications.

Stability and Reliability● Differentiated EVS disks

Common I/O, high I/O, and ultra-high I/O EVS disks are available for all ofyour service requirements.Common I/O EVS disks: feature secure, reliable, and scalable. They are idealfor applications requiring large capacity, moderate read/write speed, and fewtransactions.High I/O EVS disks: feature high performance, scalability, and reliability. Theyare ideal for applications requiring high performance, high read/write speed,and real-time data storage.Ultra-high I/O EVS disks: feature low latency and high performance. They areideal for intensive read/write applications requiring extremely highperformance and read/write speed, and low latency.

● Reliable dataScalable, reliable high-throughput virtual block storage is based on distributedarchitecture. This ensures that data can be rapidly migrated and restored ifany data replica is unavailable, preventing data loss caused by a singlehardware fault.

● Backup and restoration of ECSs and EVS disksAutomatic backup policies can be preset to back up in-service ECSs and EVSdisks. Additionally, the data of ECSs and EVS disks at a specified time can beautomatically backed up through the management console or API.

Security● Various security services are provided for multi-dimensional protection.

Security services, such as WAF and VSS are available.● Security evaluation

Elastic Cloud ServerService Overview 2 ECS Advantages

2020-07-06 4

Cloud environment security evaluation helps you quickly identify securityvulnerabilities and threats. Security configuration check and recommendationsreduce or eliminate your loss from network viruses or attacks.

● Intelligent process managementIntelligent process management automatically prohibits the execution ofunauthorized programs based on a customized allowlist, thereby ensuring ECSsecurity.

● Vulnerability scanVarious scanning services are provided, including general web vulnerabilityscanning, third-party application vulnerability scanning, port detection, andfingerprint identification.

Competitive Advantage● Professional hardware devices

ECSs are deployed on professional hardware devices that support in-depthvirtualization optimization, relieving you of equipment-room concerns.

● Always available virtualization resourcesScalable, dedicated resources can be obtained from the virtualized resourcepool any time, ensuring reliable, secure, flexible, and efficient applicationenvironments. You can use your ECS like using your local computer.

Auto Scaling● Automatic adjustment of computing resources

Dynamic scaling: AS automatically increases or decreases the number of ECSsin an AS group based on monitored data.Periodic/Scheduled scaling: AS increases or decreases the number of ECSs inan AS group periodically or at a specified time based on service expectationand operation plan.

● Flexible adjustment of ECS configurationsECS specifications and bandwidth can be flexibly adjusted based on servicerequirements.

● Flexible billing modesYearly/Monthly and pay-per-use billing modes allow you to purchase andrelease resources at any time based on service fluctuation.

Elastic Cloud ServerService Overview 2 ECS Advantages

2020-07-06 5

3 ECS Application Scenarios

InternetNo special requirements on CPUs, memory, disk space, or bandwidth; strongsecurity and reliability; application deployment based on one or only a few ECSs tominimize initial investment and maintenance costs, such as website R&D andtesting, and small-scale databases

Use general computing ECSs, which provide a balance of computing, memory, andnetwork resources. This ECS type is appropriate for medium-workload applicationsand meets the cloud service needs of both enterprises and individuals.

E-CommerceLarge amount of memory; capable of processing large volumes of data; fastnetwork and rapid data processing, such as precision marketing, E-Commerce, andmobile apps

Use memory-optimized ECSs, which have a large amount of memory and provideultra-high I/O EVS disks and appropriate bandwidths.

Graphics RenderingHigh-quality graphics and video; large amount of memory, capable of processinglarge volumes of data, and high I/O concurrency; fast network and rapid dataprocessing; high GPU performance, such as graphics rendering and engineeringdrawing

Use GPU-accelerated ECSs, including G1 ECSs, which are based on NVIDIA TeslaM60 hardware virtualization and provide cost-effective graphics acceleration.These ECSs support DirectX and OpenGL, provide computing with up to 1 GB ofGPU memory and 4096 x 2160 resolution.

Data AnalysisCapable of processing large volumes of data; high I/O performance and rapid dataswitching and processing, such as MapReduce and Hadoop

Use disk-intensive ECSs, which are designed for applications requiring sequentialread/write on ultra-large datasets in local storage (such as distributed Hadoop

Elastic Cloud ServerService Overview 3 ECS Application Scenarios

2020-07-06 6

computing) as well as large-scale parallel data processing and log processing.Disk-intensive ECSs are based on HDD and a default network bandwidth of 10GE,providing high PPS and low network latency. They also support up to 24 localdisks, 48 vCPUs, and 384 GB of memory.

High-Performance ComputingHigh computing performance and throughput, such as scientific computing,genetic engineering, games and animation, biopharmaceuticals, and storage

Use high-performance computing ECSs to meet the computing, storage, andrendering needs of high-performance infrastructure services and applications thatrequire a large number of parallel computing resources.

Elastic Cloud ServerService Overview 3 ECS Application Scenarios

2020-07-06 7

4 Notes on Using ECSs

Before using ECSs, read the following notes:

Notes on ECS Application Scenarios● Do not use ECSs for any illegal or violation service, such as gambling, private

service, or cross-border VPN.● Do not use ECSs for fake transactions, such as click farming on e-commerce

websites.● Do not use ECSs to initiate network attacks, such as DDoS attacks, CC attacks,

web attacks, brute force cracking, or spreading of viruses and Trojan horses.● Do not use ECSs for traffic transit.● Do not use ECSs to set up the crawler environment for data crawling.● ECSs can be used for probe, such as scan or penetrate external systems only

after being authorized by the external systems.● Do not deploy any illegal websites or applications on ECSs.

Constraints● Do not uninstall the driver on the ECS hardware.● Do not load external hardware devices, such as encryption locks or USB flash

drives, on the ECS.● Do not change the MAC address of a NIC.● ECSs do not support secondary virtualization.● The authentication of certain software may bind a license to the physical

server hardware. Once the ECS deployed on the physical server is migratedout, the bound license fails due to the change of the physical server.

● If an ECS migrates out of a faulty physical server, the ECS may be stopped orrestarted. For high service availability, deploy services in a cluster or on ECSsworking in active/standby mode, or configure automatic ECS startup upon aphysical server failure or startup.

● Back up data for the ECSs where core services are deployed.● Monitor application metrics on ECSs.● Do not change the default DNS server. If you are required to configure public

DNS, configure public and intranet DNS on your ECS.

Elastic Cloud ServerService Overview 4 Notes on Using ECSs

2020-07-06 8

Notes on Using Windows● Do not stop system processes. Otherwise, blue screen of death (BSOD) may

occur on the ECS, or the ECS may restart.● Ensure that there is at least 2 GB of idle memory. Otherwise, BSOD, frame

freezing, or service running failure may occur.● Do not modify the registry. Otherwise, starting the system may fail. If the

modification is mandatory, back up the registry before modifying it.● Do not modify ECS clock settings. Otherwise, DHCP lease may fail, leading to

the loss of IP addresses.● Do not delete the CloudResetPwdAgent or CloudResetPwdUpdateAgent

process. Otherwise, one-click password reset will be unavailable.● Do not disable virtual memory. Otherwise, system performance may

deteriorate, or system exceptions may occur.● Do not delete the VMTool program. Otherwise, the ECS may fail to run

properly.

Notes on Using Linux● Do not modify the /etc/issue file. Otherwise, the system edition will not be

identified.● Do not delete system directories or files. Otherwise, the system may fail to

start or run.● Do not change the permissions or names of system directories. Otherwise, the

system may fail to start or run.● Upgrade Linux kernel only required. For details, see How Can I Upgrade the

Kernel of a Linux ECS?● Do not delete the CloudResetPwdAgent or CloudResetPwdUpdateAgent

process. Otherwise, one-click password reset will be unavailable.● Do not change the default DNS server /etc/resolv.conf. Otherwise, internal

services, such as software sources and NTP may be unavailable.● Do not modify default intranet configurations, such as IP addresses, subnet

mask, and gateway address of an ECS. Otherwise, network exceptions mayoccur.

Elastic Cloud ServerService Overview 4 Notes on Using ECSs

2020-07-06 9

https://support.huaweicloud.com/intl/en-us/ecs_faq/ecs_faq_0723.html


5 Instances

5.1 OverviewAn ECS is a basic computing unit that consists of vCPUs, memory, OS, and EVSdisks.

After creating an ECS, you can use it like using your local computer or physicalserver, ensuring secure, reliable, and efficient computing. ECSs support self-servicecreation, modification, and operation. You can create an ECS by specifying itsvCPUs, memory, OS, and login authentication. After the ECS is created, you canmodify its specifications as required.

The cloud platform provides multiple ECS types for different computing andstorage capabilities. One ECS type provides various flavors with different vCPU andmemory configurations for you to select.

● For details about ECS types, see 5.3 ECS Types.

● For details about all ECS statuses in a lifecycle, see 5.2 ECS Lifecycle.

● For details about ECS specifications, see 5.5 x86 ECS Specifications .

5.2 ECS LifecycleA lifecycle indicates the ECS statuses recorded from the time when the ECS iscreated through the time when the ECS is deleted or released.

Table 5-1 ECS statuses

Status StatusAttribute

Description

Creating Intermediate The ECS has been created but is not running.

Starting Intermediate The ECS is between the Stopped andRunning states.

Elastic Cloud ServerService Overview 5 Instances

2020-07-06 10


Description

Running Stable The ECS is running properly.An ECS in this state can provide services.

Stopping Intermediate The ECS is between the Running andStopped states.

Stopped Stable The ECS has been properly stopped.An ECS in this state cannot provide services.

Restarting Intermediate The ECS is being restarted.

Resizing Intermediate The ECS has received a resizing request andhas started to resize.

Verifyingresizing

Intermediate The ECS is verifying the modifiedconfiguration.

Deleting Intermediate The ECS is being deleted.If the ECS remains in this state for a longtime, exceptions may have occurred. In sucha case, contact the administrator.

Deleted Intermediate The ECS has been deleted. An ECS in thisstate cannot provide services and will bepromptly cleared from the system.

Faulty Stable An exception has occurred on the ECS.An ECS in this state cannot provide services.Contact the administrator.

ReinstallingOS

Intermediate The ECS has received a request to reinstallthe OS and has begun the reinstallation.

ReinstallingOS failed

Stable The ECS received a request to reinstall theOS, but due to exceptions, the reinstallationfailed.An ECS in this state cannot provide services.Contact the administrator.

Changing OS Intermediate The ECS received a request to change the OSand has begun implementing the changes.

OS Changefailed

Stable The ECS has received a request to changethe OS, but due to exceptions, the changesfailed to be implemented.An ECS in this state cannot provide services.Contact the administrator.

Forciblyrestarting

Intermediate The ECS is being forcibly restarted.


2020-07-06 11


Description

Rolling backresizing

Intermediate The ECS is rolling back resizing.

Frozen Stable The ECS has been stopped by theadministrator because the order has expiredor is overdue.An ECS in this state cannot provide services.The system retains it for a period of time. Ifit is not renewed after the time expires, thesystem will automatically delete the ECS.

5.3 ECS TypesThe public cloud provides the following ECS types for different applicationscenarios:● Kunpeng computing

– Kunpeng general computing-plus● x86 computing

– General computing– General computing-plus– General computing-basic– Memory-optimized– Large-memory– Disk-intensive– Ultra-high I/O– High-performance computing– GPU-accelerated

ECS Flavor Naming Rules

ECS flavors are named using the format "AB.C.D".

The format is defined as follows:

● A specifies the ECS type. For example, s indicates a general computing ECS, ca computing ECS, and m a memory-optimized ECS.

● B specifies the type ID. For example, the 1 in s1 indicates a general computingfirst-generation ECS, and the 2 in s2 indicates a general computing second-generation ECS.

● C specifies the flavor size, such as medium, large, or xlarge.● D specifies the ratio of memory to vCPUs expressed in a digit. For example,

value 4 indicates that the ratio of memory to vCPUs is 4.


2020-07-06 12

Network Bandwidth

The intranet bandwidth and PPS of an ECS are determined based on ECS flavors.

● Assured intranet bandwidth: indicates the assured ECS bandwidth.

● Maximum intranet bandwidth: indicates the maximum ECS bandwidth.

● Maximum intranet PPS: indicates the maximum ECS capabilities intransmitting and receiving packets.

5.4 Kunpeng ECS Specifications

Kunpeng General Computing-plus

Table 5-2 KC1 ECS specifications

Flavor vCPUs Memory(GB)

Maximum/AssuredBandwidth(Gbit/s)

Maximum PPS(10,000)

NICMulti-Queue

Maximum NICs

VirtualizationType

kc1.small.1

1 1 2/0.5 20 1 2 KVM

kc1.large.2

2 4 3/0.8 30 2 2 KVM

kc1.xlarge.2

4 8 5/1.5 50 2 3 KVM

kc1.2xlarge.2

8 16 7/3 80 4 4 KVM

kc1.3xlarge.2

12 24 9/4.5 110 4 5 KVM

kc1.4xlarge.2

16 32 12/6 140 4 6 KVM

kc1.6xlarge.2

24 48 15/8.5 200 8 6 KVM

kc1.8xlarge.2

32 64 18/10 260 8 6 KVM

kc1.12xlarge.2

48 96 25/16 350 16 6 KVM

kc1.15xlarge.2

60 120 30/20 400 16 6 KVM

kc1.large.4

2 8 3/0.8 30 2 2 KVM


2020-07-06 13



Maximum PPS(10,000)

NICMulti-Queue

Maximum NICs

VirtualizationType

kc1.xlarge.4

4 16 5/1.5 50 2 3 KVM

kc1.2xlarge.4

8 32 7/3 80 4 4 KVM

kc1.3xlarge.4

12 48 9/4.5 110 4 5 KVM

kc1.4xlarge.4

16 64 12/6 140 4 6 KVM

kc1.6xlarge.4

24 96 15/8.5 200 8 6 KVM

kc1.8xlarge.4

32 128 18/10 260 8 6 KVM

kc1.12xlarge.4

48 192 25/16 350 16 6 KVM

5.5 x86 ECS Specifications

General Computing

Table 5-3 S6 ECS specifications

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

s6.small.1

1 1 0.8/0.1 10 1 2 KVM

s6.medium.2

1 2 0.8/0.1 10 1 2 KVM

s6.large.2

2 4 1.5/0.2 15 1 2 KVM

s6.xlarge.2

4 8 2/0.35 25 1 2 KVM


2020-07-06 14

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

s6.2xlarge.2

8 16 3/0.75 50 2 2 KVM

s6.medium.4

1 4 0.8/0.1 10 1 2 KVM

s6.large.4

2 8 1.5/0.2 15 1 2 KVM

s6.xlarge.4

4 16 2/0.35 25 1 2 KVM

s6.2xlarge.4

8 32 3/0.75 50 2 2 KVM


Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

s3.small.1

1 1 0.5/0.1 5 1 KVM

s3.medium.2

1 2 0.5/0.1 5 1 KVM

s3.large.2

2 4 0.8/0.2 10 1 KVM

s3.xlarge.2

4 8 1.5/0.4 15 1 KVM

s3.2xlarge.2

8 16 3/0.8 20 2 KVM

s3.4xlarge.2

16 32 4/1.5 30 4 KVM

s3.medium.4

1 4 0.5/0.1 5 1 KVM

s3.large.4

2 8 0.8/0.2 10 1 KVM

s3.xlarge.4

4 16 1.5/0.4 15 1 KVM


2020-07-06 15

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

s3.2xlarge.4

8 32 3/0.8 20 2 KVM

s3.4xlarge.4

16 64 4/1.5 30 4 KVM


Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

s2.small.1

1 1 0.5/0.1 5 1 KVM

s2.medium.2

1 2 0.5/0.1 5 1 KVM

s2.large.2

2 4 0.8/0.2 10 1 KVM

s2.xlarge.2

4 8 1.5/0.4 15 1 KVM

s2.2xlarge.2

8 16 3/0.8 20 2 KVM

s2.4xlarge.2

16 32 4/1.5 30 4 KVM

s2.medium.4

1 4 0.5/0.1 5 1 KVM

s2.large.4

2 8 0.8/0.2 10 1 KVM

s2.xlarge.4

4 16 1.5/0.4 15 1 KVM

s2.2xlarge.4

8 32 3/0.8 20 2 KVM

s2.4xlarge.4

16 64 4/1.5 30 4 KVM


2020-07-06 16

General Computing-plus

Table 5-6 C6s ECS specifications

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c6s.large.2

2 4 1/1 30 2 2 KVM

c6s.xlarge.2

4 8 2/2 60 2 3 KVM

c6s.2xlarge.2

8 16 4/4 120 4 4 KVM

c6s.3xlarge.2

12 24 5.5/5.5 180 4 6 KVM

c6s.4xlarge.2

16 32 7.5/7.5 240 8 8 KVM

c6s.6xlarge.2

24 48 11/11 350 8 8 KVM

c6s.8xlarge.2

32 64 15/15 450 16 8 KVM

c6s.12xlarge.2

48 96 22/22 650 16 8 KVM

c6s.16xlarge.2

64 128 30/30 850 32 8 KVM

Table 5-7 C6 ECS specifications

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c6.large.2

2 4 4/1.2 40 2 2 KVM

c6.xlarge.2

4 8 8/2.4 80 2 3 KVM


2020-07-06 17

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c6.2xlarge.2

8 16 15/4.5 150 4 4 KVM

c6.3xlarge.2

12 24 17/7 200 4 6 KVM

c6.4xlarge.2

16 32 20/9 280 8 8 KVM

c6.6xlarge.2

24 48 25/14 400 8 8 KVM

c6.8xlarge.2

32 64 30/18 550 16 8 KVM

c6.16xlarge.2

64 128 40/36 1000 32 8 KVM

c6.large.4

2 8 4/1.2 40 2 2 KVM

c6.xlarge.4

4 16 8/2.4 80 2 3 KVM

c6.2xlarge.4

8 32 15/4.5 150 4 4 KVM

c6.3xlarge.4

12 48 17/7 200 4 6 KVM

c6.4xlarge.4

16 64 20/9 280 8 8 KVM

c6.6xlarge.4

24 96 25/14 400 8 8 KVM

c6.8xlarge.4

32 128 30/18 550 16 8 KVM

c6.16xlarge.4

64 256 40/36 1000 32 8 KVM


2020-07-06 18

Table 5-8 C3ne ECS specifications

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c3ne.large.2

2 4 4/1.3 40 2 2 KVM

c3ne.xlarge.2

4 8 8/2.5 80 2 3 KVM

c3ne.2xlarge.2

8 16 15/5 150 4 4 KVM

c3ne.4xlarge.2

16 32 20/10 280 8 8 KVM

c3ne.8xlarge.2

32 64 30/20 550 16 8 KVM

c3ne.15xlarge.2

60 128 40/40 1000 32 8 KVM

c3ne.large.4

2 8 4/1.3 40 2 2 KVM

c3ne.xlarge.4

4 16 8/2.5 80 2 3 KVM

c3ne.2xlarge.4

8 32 15/5 150 4 4 KVM

c3ne.4xlarge.4

16 64 20/10 280 8 8 KVM

c3ne.8xlarge.4

32 128 30/20 550 16 8 KVM

c3ne.15xlarge.4

60 256 40/40 1000 32 8 KVM


Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

c3.large.2

2 4 1.5/0.6 30 2 KVM


2020-07-06 19

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

c3.xlarge.2

4 8 3/1 50 2 KVM

c3.2xlarge.2

8 16 5/2 90 4 KVM

c3.3xlarge.2

12 24 7/3 110 4 KVM

c3.4xlarge.2

16 32 10/4 130 4 KVM

c3.6xlarge.2

24 48 12/6 200 8 KVM

c3.8xlarge.2

32 64 15/8 260 8 KVM

c3.15xlarge.2

60 128 17/16 500 16 KVM

c3.large.4

2 8 1.5/0.6 30 2 KVM

c3.xlarge.4

4 16 3/1 50 2 KVM

c3.2xlarge.4

8 32 5/2 90 4 KVM

c3.3xlarge.4

12 48 7/3 110 4 KVM

c3.4xlarge.4

16 64 10/4 130 4 KVM

c3.6xlarge.4

24 96 12/6 200 8 KVM

c3.8xlarge.4

32 128 15/8 260 8 KVM

c3.15xlarge.4

60 256 17/16 500 16 KVM


2020-07-06 20

General Computing-Basic

Table 5-10 T6 ECS specifications

Flavor vCPUs

Memory(GB)

InitialCredits

MaximumCredits

CPUCredits/Hour

CPUBaseline(%)

AverageCPUBaseline(%)

MaximumNICs

VirtualizationType

t6.small.1

1 1 30 144 6 10 10 1 KVM

t6.large.1

2 2 60 576 24 40 20 1 KVM

t6.xlarge.1

4 4 120 1152 48 80 20 2 KVM

t6.2xlarge.1

8 8 120 1728 72 120 15 2 KVM

t6.4xlarge.1

16 16 160 3456 144 240 15 2 KVM

t6.medium.2

1 2 30 144 6 10 10 1 KVM

t6.large.2

2 4 60 576 24 40 20 1 KVM

t6.xlarge.2

4 8 120 1152 48 80 20 2 KVM

t6.2xlarge.2

8 16 120 1728 72 120 15 2 KVM

t6.4xlarge.2

16 32 160 3456 144 240 15 2 KVM

t6.large.4

2 8 60 576 24 40 20 1 KVM

t6.xlarge.4

4 16 120 1152 48 80 20 2 KVM

t6.2xlarge.4

8 32 120 1728 72 120 15 2 KVM


2020-07-06 21

Memory-optimized

Table 5-11 M6 ECS specifications

Flavor vCPUs Memory (GB)


Maximum PPS(10,000)

NICMulti-Queue

VirtualizationType

m6.large.8 2 16 4/1.2 40 2 KVM

m6.xlarge.8

4 32 8/2.4 80 2 KVM

m6.2xlarge.8

8 64 15/4.5 150 4 KVM

m6.3xlarge.8

12 96 17/7 200 4 KVM

m6.4xlarge.8

16 128 20/9 280 8 KVM

m6.6xlarge.8

24 192 25/14 400 8 KVM

m6.8xlarge.8

32 256 30/18 550 16 KVM

m6.16xlarge.8

64 512 40/36 1000 32 KVM

Table 5-12 M3ne ECS specifications

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

m3ne.large.8

2 16 4/1.3 40 2 2 KVM

m3ne.xlarge.8

4 32 8/2.5 80 2 3 KVM

m3ne.2xlarge.8

8 64 15/5 150 4 4 KVM

m3ne.3xlarge.8

12 96 17/8 200 4 6 KVM

m3ne.4xlarge.8

16 128 20/10 280 8 8 KVM


2020-07-06 22

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

m3ne.6xlarge.8

24 192 25/16 400 8 8 KVM

m3ne.8xlarge.8

32 256 30/20 550 16 8 KVM

m3ne.15xlarge.8

60 512 40/40 1000 32 8 KVM


Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

m3.large.8 2 16 1.5/0.6 30 2 KVM

m3.xlarge.8

4 32 3/1.1 50 2 KVM

m3.2xlarge.8

8 64 5/2 90 4 KVM

m3.3xlarge.8

12 96 8/3.5 110 4 KVM

m3.4xlarge.8

16 128 10/4.5 130 4 KVM

m3.6xlarge.8

24 192 12/6.5 200 8 KVM

m3.8xlarge.8

32 256 15/9 260 8 KVM

m3.15xlarge.8

60 512 17/17 500 16 KVM


ECS Type vCPUs Memory(GB)

Flavor VirtualizationType

Memory-optimized 2 16 m2.large.8 KVM


2020-07-06 23



4 32 m2.xlarge.8 KVM

8 64 m2.2xlarge.8 KVM


Large-Memory

Table 5-15 E3 ECS specifications

Flavor vCPUs Memory (GB) VirtualizationType

e3.7xlarge.12 28 348 KVM

Disk-intensive

Table 5-16 D2 ECS specifications

Flavor vCPUs

Memory(GB)


Maximum PPS(10,000)

NICMulti-Queue

LocalDisks(GB)

VirtualizationType

d2.xlarge.8

4 32 3/1 15 2 2 x1800

KVM

d2.2xlarge.8

8 64 5/2 30 2 4 x1800

KVM

d2.4xlarge.8

16 128 8/4 40 4 8 x1800

KVM

d2.6xlarge.8

24 192 10/6 50 6 12 x1800

KVM

d2.8xlarge.8

32 256 13/8 60 8 16 x1800

KVM

d2.12xlarge.8

48 384 13/13 90 8 24 x1800

KVM


2020-07-06 24

Ultra-high I/O

Table 5-17 I3 ECS specifications

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

LocalDisks

MaximumNICs

VirtualizationType

i3.2xlarge.8

8 64 8/3.5 100 4 1 x1600GBNVMe

4 KVM

i3.4xlarge.8

16 128 15/7 160 4 2 x1600GBNVMe

8 KVM

i3.8xlarge.8

32 256 20/14 280 8 4 x1600GBNVMe

8 KVM

i3.12xlarge.8

48 384 25/20 420 8 6 x1600GBNVMe

8 KVM

i3.15xlarge.8

60 512 25/25 500 16 7 x1600GBNVMe

8 KVM

i3.16xlarge.8

64 512 25/25 500 16 8 x1600GBNVMe

8 KVM


2020-07-06 25

High-Performance Computing

Table 5-18 H3 ECS specifications

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

h3.large.2 2 4 2/1 30 2 KVM

h3.xlarge.2 4 8 4/2 60 2 KVM

h3.2xlarge.2

8 16 6/3.5 120 4 KVM

h3.3xlarge.2

12 24 6/5.5 160 4 KVM

h3.4xlarge.2

16 32 12/7.5 200 8 KVM

h3.large.4 2 8 2/1 30 2 KVM

h3.xlarge.4 4 16 4/2 60 2 KVM

h3.2xlarge.4

8 32 6/3.5 120 4 KVM

h3.3xlarge.4

12 48 6/5.5 160 4 KVM

h3.4xlarge.4

16 64 12/7.5 200 8 KVM

Table 5-19 HC2 ECS specifications

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

hc2.large.2 2 4 1.5/0.5 10 1 KVM

hc2.xlarge.2

4 8 3/1 15 1 KVM

hc2.2xlarge.2

8 16 5/2 30 2 KVM

hc2.4xlarge.2

16 32 8/4 40 4 KVM

hc2.large.4 2 8 1.5/0.5 10 1 KVM


2020-07-06 26

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

hc2.xlarge.4

4 16 3/1 15 1 KVM

hc2.2xlarge.4

8 32 5/2 30 2 KVM

hc2.4xlarge.4

16 64 8/4 40 4 KVM

GPU-accelerated

Table 5-20 G5 ECS specifications

Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUMemory(GB)

Virtualization Type

g5.8xlarge.4

32 128 25/15 200 16 1 xV100

16 KVM

Table 5-21 P2v ECS specifications

Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUConnection

GPUMemory(GB)

VirtualizationType

p2v.2xlarge.8

8 64 10/4 50 4 1 xV100

N/A 1 x16

KVM

p2v.4xlarge.8

16 128 15/8 100 8 2 xV100

NVLink

2 x16

KVM


2020-07-06 27

Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUConnection

GPUMemory(GB)

VirtualizationType

p2v.8xlarge.8

32 256 25/15 200 16 4 xV100

NVLink

4 x16

KVM

p2v.16xlarge.8

64 512 30/30 400 32 8 xV100

NVLink

8 x16

KVM

Table 5-22 PI2 ECS specifications

Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUMemory(GB)

LocalDisks

Virtualization Type

pi2.2xlarge.4

8 32 10/4 50 4 1 xT4

1 x16

N/A KVM

pi2.4xlarge.4

16 64 15/8 100 8 2 xT4

2 x16

N/A KVM

pi2.8xlarge.4

32 128 25/15 200 16 4 xT4

4 x16

N/A KVM


Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUMemory(GB)

LocalDisks

VirtualizationType

pi1.2xlarge.4

8 32 5/1.6 40 2 1 xP4

1 x 8GB

N/A KVM


2020-07-06 28

Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUMemory(GB)

LocalDisks

VirtualizationType

pi1.4xlarge.4

16 64 8/3.2 70 4 2 xP4

2 x 8GB

N/A KVM

pi1.8xlarge.4

32 128 10/6.5 140 8 4 xP4

4 x 8GB

N/A KVM

5.6 Kunpeng General Computing-plus ECSs

OverviewKunpeng general computing-plus KC1 ECSs use Kunpeng 920 processors and 25GEhigh-speed intelligent NICs to offer powerful computing and high-performancenetworks, meeting the requirements of governments and Internet enterprises forcost-effective, secure, reliable cloud services.

Specifications

Table 5-24 KC1 ECS specifications



Maximum PPS(10,000)

NICMulti-Queue

Maximum NICs

VirtualizationType

kc1.small.1

1 1 2/0.5 20 1 2 KVM

kc1.large.2

2 4 3/0.8 30 2 2 KVM

kc1.xlarge.2

4 8 5/1.5 50 2 3 KVM

kc1.2xlarge.2

8 16 7/3 80 4 4 KVM

kc1.3xlarge.2

12 24 9/4.5 110 4 5 KVM


2020-07-06 29



Maximum PPS(10,000)

NICMulti-Queue

Maximum NICs

VirtualizationType

kc1.4xlarge.2

16 32 12/6 140 4 6 KVM

kc1.6xlarge.2

24 48 15/8.5 200 8 6 KVM

kc1.8xlarge.2

32 64 18/10 260 8 6 KVM

kc1.12xlarge.2

48 96 25/16 350 16 6 KVM

kc1.15xlarge.2

60 120 30/20 400 16 6 KVM

kc1.large.4

2 8 3/0.8 30 2 2 KVM

kc1.xlarge.4

4 16 5/1.5 50 2 3 KVM

kc1.2xlarge.4

8 32 7/3 80 4 4 KVM

kc1.3xlarge.4

12 48 9/4.5 110 4 5 KVM

kc1.4xlarge.4

16 64 12/6 140 4 6 KVM

kc1.6xlarge.4

24 96 15/8.5 200 8 6 KVM

kc1.8xlarge.4

32 128 18/10 260 8 6 KVM

kc1.12xlarge.4

48 192 25/16 350 16 6 KVM

ScenariosKC1 ECSs are suitable for governments, enterprises, and the financial industry withstrict requirements on security and privacy, for Internet applications with highrequirements on network performance, for big data and HPC requiring a largenumber of vCPUs, and for website setups and e-Commerce requiring cost-effectiveness.


2020-07-06 30

5.7 General Computing ECSs

OverviewGeneral computing ECSs provide a balance of computing, memory, and networkresources and a baseline level of vCPU performance with the ability to burst abovethe baseline. These ECSs are suitable for many applications, such as web servers,enterprise R&D, and small-scale databases.

S6 ECSs are suitable for applications that require moderate performance generallybut occasionally burstable high performance, such as light-workload web servers,enterprise R&D and testing environments, and low- and medium-performancedatabases. S6 ECS performance is neither restricted by vCPU credits nor billed foradditional credits. You can determine the CPU usage and vCPU credits inmonitoring details.

Specifications


Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

s6.small.1

1 1 0.8/0.1 10 1 2 KVM

s6.medium.2

1 2 0.8/0.1 10 1 2 KVM

s6.large.2

2 4 1.5/0.2 15 1 2 KVM

s6.xlarge.2

4 8 2/0.35 25 1 2 KVM

s6.2xlarge.2

8 16 3/0.75 50 2 2 KVM

s6.medium.4

1 4 0.8/0.1 10 1 2 KVM

s6.large.4

2 8 1.5/0.2 15 1 2 KVM

s6.xlarge.4

4 16 2/0.35 25 1 2 KVM

s6.2xlarge.4

8 32 3/0.75 50 2 2 KVM


2020-07-06 31


Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

s3.small.1

1 1 0.5/0.1 5 1 KVM

s3.medium.2

1 2 0.5/0.1 5 1 KVM

s3.large.2

2 4 0.8/0.2 10 1 KVM

s3.xlarge.2

4 8 1.5/0.4 15 1 KVM

s3.2xlarge.2

8 16 3/0.8 20 2 KVM

s3.4xlarge.2

16 32 4/1.5 30 4 KVM

s3.medium.4

1 4 0.5/0.1 5 1 KVM

s3.large.4

2 8 0.8/0.2 10 1 KVM

s3.xlarge.4

4 16 1.5/0.4 15 1 KVM

s3.2xlarge.4

8 32 3/0.8 20 2 KVM

s3.4xlarge.4

16 64 4/1.5 30 4 KVM


Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

s2.small.1

1 1 0.5/0.1 5 1 KVM

s2.medium.2

1 2 0.5/0.1 5 1 KVM

s2.large.2

2 4 0.8/0.2 10 1 KVM


2020-07-06 32

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

s2.xlarge.2

4 8 1.5/0.4 15 1 KVM

s2.2xlarge.2

8 16 3/0.8 20 2 KVM

s2.4xlarge.2

16 32 4/1.5 30 4 KVM

s2.medium.4

1 4 0.5/0.1 5 1 KVM

s2.large.4

2 8 0.8/0.2 10 1 KVM

s2.xlarge.4

4 16 1.5/0.4 15 1 KVM

s2.2xlarge.4

8 32 3/0.8 20 2 KVM

s2.4xlarge.4

16 64 4/1.5 30 4 KVM

Scenarios● Applications

General-purpose ECSs are suitable for applications that have no specialrequirements on CPU performance, memory, disk capacity, or bandwidth, buthave high requirements on security and reliability. They feature low initialinvestment and maintenance costs.

● Application scenarios

Enterprise website deployment, enterprise office environment setup,enterprise R&D and testing activities, web servers, R&D and testingenvironments, and small-scale databases

5.8 General Computing-plus ECSs

Overview

General computing-plus ECSs provide dedicated vCPUs when being compared withgeneral computing ECSs, featuring powerful performance. In addition, the ECSsuse latest-generation network acceleration engines and Data Plane DevelopmentKit (DPDK) to provide higher network performance, meeting requirements indifferent scenarios.


2020-07-06 33

● C6s ECSs use second-generation Intel® Xeon® Scalable processors that featurehigh performance, stability, low latency, and cost-effectiveness. They aresuitable for Internet, gaming, and rendering scenarios, especially those thatrequire high computing and network stability.

● C6 ECSs use second-generation Intel® Xeon® Scalable processors withtechnologies optimized and 25GE high-speed intelligent NICs to offerpowerful and stable computing performance, including ultra-high networkbandwidth and PPS.

● C3ne ECSs provide higher computing and network forwarding capabilitiesthan C3 ECSs. Using Intel® Xeon® Scalable processors and 25GE high-speedintelligent NICs, the C3ne ECSs provide a maximum intranet bandwidth of 40Gbit/s and 10 million PPS for enterprise-grade applications with high networkperformance requirements.

● C3 ECSs are newly released. They use Intel® Xeon® Scalable processors andfeature high and stable computing performance. Equipped with high-performance NICs, the C3 ECSs provide high performance and stability,meeting enterprise-grade application requirements.

Specifications

Table 5-28 C6s ECS specifications

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c6s.large.2

2 4 1/1 30 2 2 KVM

c6s.xlarge.2

4 8 2/2 60 2 3 KVM

c6s.2xlarge.2

8 16 4/4 120 4 4 KVM

c6s.3xlarge.2

12 24 5.5/5.5 180 4 6 KVM

c6s.4xlarge.2

16 32 7.5/7.5 240 8 8 KVM

c6s.6xlarge.2

24 48 11/11 350 8 8 KVM

c6s.8xlarge.2

32 64 15/15 450 16 8 KVM

c6s.12xlarge.2

48 96 22/22 650 16 8 KVM


2020-07-06 34

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c6s.16xlarge.2

64 128 30/30 850 32 8 KVM


Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c6.large.2

2 4 4/1.2 40 2 2 KVM

c6.xlarge.2

4 8 8/2.4 80 2 3 KVM

c6.2xlarge.2

8 16 15/4.5 150 4 4 KVM

c6.3xlarge.2

12 24 17/7 200 4 6 KVM

c6.4xlarge.2

16 32 20/9 280 8 8 KVM

c6.6xlarge.2

24 48 25/14 400 8 8 KVM

c6.8xlarge.2

32 64 30/18 550 16 8 KVM

c6.16xlarge.2

64 128 40/36 1000 32 8 KVM

c6.large.4

2 8 4/1.2 40 2 2 KVM

c6.xlarge.4

4 16 8/2.4 80 2 3 KVM

c6.2xlarge.4

8 32 15/4.5 150 4 4 KVM

c6.3xlarge.4

12 48 17/7 200 4 6 KVM


2020-07-06 35

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c6.4xlarge.4

16 64 20/9 280 8 8 KVM

c6.6xlarge.4

24 96 25/14 400 8 8 KVM

c6.8xlarge.4

32 128 30/18 550 16 8 KVM

c6.16xlarge.4

64 256 40/36 1000 32 8 KVM

Table 5-30 C3ne ECS specifications

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c3ne.large.2

2 4 4/1.3 40 2 2 KVM

c3ne.xlarge.2

4 8 8/2.5 80 2 3 KVM

c3ne.2xlarge.2

8 16 15/5 150 4 4 KVM

c3ne.4xlarge.2

16 32 20/10 280 8 8 KVM

c3ne.8xlarge.2

32 64 30/20 550 16 8 KVM

c3ne.15xlarge.2

60 128 40/40 1000 32 8 KVM

c3ne.large.4

2 8 4/1.3 40 2 2 KVM

c3ne.xlarge.4

4 16 8/2.5 80 2 3 KVM

c3ne.2xlarge.4

8 32 15/5 150 4 4 KVM


2020-07-06 36

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

c3ne.4xlarge.4

16 64 20/10 280 8 8 KVM

c3ne.8xlarge.4

32 128 30/20 550 16 8 KVM

c3ne.15xlarge.4

60 256 40/40 1000 32 8 KVM


Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

c3.large.2

2 4 1.5/0.6 30 2 KVM

c3.xlarge.2

4 8 3/1 50 2 KVM

c3.2xlarge.2

8 16 5/2 90 4 KVM

c3.3xlarge.2

12 24 7/3 110 4 KVM

c3.4xlarge.2

16 32 10/4 130 4 KVM

c3.6xlarge.2

24 48 12/6 200 8 KVM

c3.8xlarge.2

32 64 15/8 260 8 KVM

c3.15xlarge.2

60 128 17/16 500 16 KVM

c3.large.4

2 8 1.5/0.6 30 2 KVM

c3.xlarge.4

4 16 3/1 50 2 KVM


2020-07-06 37

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

c3.2xlarge.4

8 32 5/2 90 4 KVM

c3.3xlarge.4

12 48 7/3 110 4 KVM

c3.4xlarge.4

16 64 10/4 130 4 KVM

c3.6xlarge.4

24 96 12/6 200 8 KVM

c3.8xlarge.4

32 128 15/8 260 8 KVM

c3.15xlarge.4

60 256 17/16 500 16 KVM

Scenarios● C6

Websites and web applications, generalized databases and cache servers, andmedium- and heavy-workload enterprise applications with strict requirementson computing and network performance

● C3Small- and medium-scale databases, cache servers, and search clusters withhigh requirements on stability; enterprise-grade applications of diverse typesand in various scales

5.9 General Computing-Basic ECSs

OverviewGeneral computing-basic ECSs provide a balance of computing, memory, andnetwork resources. Each ECS provides a baseline CPU performance, accumulatingCPU credits when workloads operate below the baseline threshold. They aresuitable for applications requiring burstable performance while keeping costs low.

For details about CPU usage calculations, see 11 CPU Credits.


2020-07-06 38

Specifications

Table 5-32 T6 ECS specifications

Flavor vCPUs

Memory(GB)

InitialCredits

MaximumCredits

CPUCredits/Hour

CPUBaseline(%)

AverageCPUBaseline(%)

MaximumNICs

VirtualizationType

t6.small.1

1 1 30 144 6 10 10 1 KVM

t6.large.1

2 2 60 576 24 40 20 1 KVM

t6.xlarge.1

4 4 120 1152 48 80 20 2 KVM

t6.2xlarge.1

8 8 120 1728 72 120 15 2 KVM

t6.4xlarge.1

16 16 160 3456 144 240 15 2 KVM

t6.medium.2

1 2 30 144 6 10 10 1 KVM

t6.large.2

2 4 60 576 24 40 20 1 KVM

t6.xlarge.2

4 8 120 1152 48 80 20 2 KVM

t6.2xlarge.2

8 16 120 1728 72 120 15 2 KVM

t6.4xlarge.2

16 32 160 3456 144 240 15 2 KVM

t6.large.4

2 8 60 576 24 40 20 1 KVM

t6.xlarge.4

4 16 120 1152 48 80 20 2 KVM

t6.2xlarge.4

8 32 120 1728 72 120 15 2 KVM


2020-07-06 39

5.10 Memory-optimized ECSs

Overview

Memory-optimized ECSs have a large memory size and provide high memoryperformance. They are designed for memory-intensive applications that process alarge amount of data, such as precision advertising, e-commerce big data analysis,and IoV big data analysis.

● M6 ECSs use second-generation Intel® Xeon® Scalable processors withtechnologies optimized to offer powerful and stable computing performance.Using 25GE high-speed intelligent NICs, M6 ECSs provide a maximummemory size of 512 GB based on DDR4 for large-memory applications withhigh requirements on network bandwidth and Packets Per Second (PPS).

● M3ne ECSs are suited for large-memory datasets with high networkperformance requirements. Using Intel® Xeon® Scalable processors andHi1822 high-speed intelligent NICs, the M3ne ECSs provide a maximummemory size of 512 GB based on DDR4 for large-memory applications withhigh requirements on network performance.

● M3 ECSs are developed based on the KVM virtualization platform anddesigned for processing large-scale data sets in the memory. They use Intel®Xeon® Scalable processors, network acceleration engines, and DPDK rapidpacket processing mechanism to provide higher network performance,offering a maximum memory size of 512 GB based on DDR4 for high-memorycomputing applications.

● M2 ECSs use Intel Xeon E5-2690 v4 CPUs and are designed for memory-optimized applications.

Specifications




Maximum PPS(10,000)

NICMulti-Queue

VirtualizationType

m6.large.8 2 16 4/1.2 40 2 KVM

m6.xlarge.8

4 32 8/2.4 80 2 KVM

m6.2xlarge.8

8 64 15/4.5 150 4 KVM

m6.3xlarge.8

12 96 17/7 200 4 KVM

m6.4xlarge.8

16 128 20/9 280 8 KVM


2020-07-06 40



Maximum PPS(10,000)

NICMulti-Queue

VirtualizationType

m6.6xlarge.8

24 192 25/14 400 8 KVM

m6.8xlarge.8

32 256 30/18 550 16 KVM

m6.16xlarge.8

64 512 40/36 1000 32 KVM

Table 5-34 M3ne ECS specifications

Flavor vCPUs

Memory(GB)


MaximumPPS(10,000)

NICMulti-Queue

MaximumNICs

VirtualizationType

m3ne.large.8

2 16 4/1.3 40 2 2 KVM

m3ne.xlarge.8

4 32 8/2.5 80 2 3 KVM

m3ne.2xlarge.8

8 64 15/5 150 4 4 KVM

m3ne.3xlarge.8

12 96 17/8 200 4 6 KVM

m3ne.4xlarge.8

16 128 20/10 280 8 8 KVM

m3ne.6xlarge.8

24 192 25/16 400 8 8 KVM

m3ne.8xlarge.8

32 256 30/20 550 16 8 KVM

m3ne.15xlarge.8

60 512 40/40 1000 32 8 KVM


2020-07-06 41


Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

m3.large.8 2 16 1.5/0.6 30 2 KVM

m3.xlarge.8

4 32 3/1.1 50 2 KVM

m3.2xlarge.8

8 64 5/2 90 4 KVM

m3.3xlarge.8

12 96 8/3.5 110 4 KVM

m3.4xlarge.8

16 128 10/4.5 130 4 KVM

m3.6xlarge.8

24 192 12/6.5 200 8 KVM

m3.8xlarge.8

32 256 15/9 260 8 KVM

m3.15xlarge.8

60 512 17/17 500 16 KVM




Memory-optimized 2 16 m2.large.8 KVM

4 32 m2.xlarge.8 KVM



Notes on Using M2 ECSs● To improve network performance, you can set the NIC MTU of an M2 ECS to

8888.


Memory-optimized ECSs are suitable for applications that process largevolumes of data and require a large amount of memory, rapid data switchingand processing, and low-latency storage resources.


2020-07-06 42

● Application scenariosBig data analysis, precision advertising, e-commerce big data analysis, IoV bigdata analysis, relational databases, NoSQL databases, and memory dataanalysis

5.11 Large-Memory ECSs

OverviewLarge-memory ECSs provide an even larger amount of memory than memory-optimized ECSs. They are used for applications that require a large amount ofmemory, rapid data switching, low latency, and process large volumes of data.Large-memory ECSs provide large memory and high computing, storage, andnetwork performance.

● ApplicationsLarge-memory ECSs are suitable for applications that require a large amountof memory, rapid data switching, and low latency, and process large volumesof data.

● Application scenariosE3 ECSs: OLAP and OLTP applications with hyper-threading enabled, SAPHANA applications (including Business Suite S/4HANA, Business Suite onHANA, and Business Warehouse on HANA), and big data processing engines(Apache Spark)

Specifications

Table 5-37 E3 ECS specifications

Flavor vCPUs Memory (GB) VirtualizationType

e3.7xlarge.12 28 348 KVM

Notes● Large-memory ECSs do not support NIC hot swapping.● E3 ECSs support the following OS that has been verified:

SUSE Enterprise Linux Server 12 SP2 64bit● The primary and extension NICs of a large-memory ECS have specified

application scenarios. For details, see Table 5-38.

Table 5-38 Application scenarios of the NICs of a large-memory ECS

NIC Type Application Scenario Remarks

Primary NIC Vertical layer 3communication

N/A


2020-07-06 43

NIC Type Application Scenario Remarks

ExtensionNIC

Horizontal layer 2communication

To improve network performance,you can set the MTU of anextension NIC to 8888.

● An ECS can be attached with a maximum of 60 disks, including the system

disk. For details about constraints, see Can Multiple Disks Be Attached to anECS?An example is provided as follows:An E3 ECS is to be created. It can be attached with up to 60 disks, where– The number of system disks is 1.– The number of EVS disks is at most 59.

NO TE

An existing large-memory ECS can be attached with a maximum of 40 disks (includingthe system disk). To attach 60 disks, enable advanced disk. For details, see EnablingAdvanced Disk.

5.12 Disk-intensive ECSs

OverviewDisk-intensive ECSs are delivered with local disks for high storage bandwidth andIOPS. In addition, local disks are more cost-effective in massive data storagescenarios. Disk-intensive ECSs have the following features:

● They use local disks to provide high sequential read/write performance andlow latency, improving file read/write performance.

● They provide powerful and stable computing capabilities, ensuring efficientdata processing.

● They provide high intranet performance, including high intranet bandwidthand PPS, meeting requirements for data exchange between ECSs during peakhours.

D2 ECSs are KVM-based. They use local storage for high storage performance andintranet bandwidth.


2020-07-06 44



https://support.huaweicloud.com/intl/en-us/usermanual-ecs/ecs_03_0307.html


Specifications

Table 5-39 D2 ECS specifications

Flavor vCPUs

Memory(GB)


Maximum PPS(10,000)

NICMulti-Queue

LocalDisks(GB)

VirtualizationType

d2.xlarge.8

4 32 3/1 15 2 2 x1800

KVM

d2.2xlarge.8

8 64 5/2 30 2 4 x1800

KVM

d2.4xlarge.8

16 128 8/4 40 4 8 x1800

KVM

d2.6xlarge.8

24 192 10/6 50 6 12 x1800

KVM

d2.8xlarge.8

32 256 13/8 60 8 16 x1800

KVM

d2.12xlarge.8

48 384 13/13 90 8 24 x1800

KVM


Disk-intensive ECSs are suitable for applications that require large volumes ofdata to process, high I/O performance, and rapid data switching andprocessing.

● Application scenarios

MPP data warehouse, distributed MapReduce and Hadoop computing,distributed file systems, network file systems, and log/data processing

Specifications of a Single SAS HDD Disk Attached to a D2 ECS

Table 5-40 Specifications of a single SAS HDD disk attached to a D2 ECS

Metric Performance

Disk capacity 1800 GB

Maximum throughput 230 MB/s

Access latency Millisecond-level


2020-07-06 45

Notes on Using D2 ECSs● D2 ECSs support the following OSs:

– CentOS 6.7/6.8/7.2/7.3/7.4 64bit– SUSE Enterprise Linux Server 11 SP3/SP4 64bit– SUSE Enterprise Linux Server 12 SP1/SP2 64bit– Red Hat Enterprise Linux 6.8/7.3 64bit– Windows Server 2008 R2 Enterprise 64bit– Windows Server 2012 R2 Standard 64bit– Windows Server 2016 Standard 64bit– Debian 8.7/9.0.0 64bit– EulerOS 2.2 64bit– Fedora 25/26 64bit– OpenSUSE 42.2/42.3 64bit

● When the physical host where a D2 ECS is deployed becomes faulty, the ECScannot be migrated.

● To improve network performance, you can set the NIC MTU of a D2 ECS to8888.

● D2 ECSs do not support modifying specifications.● D2 ECSs do not support local disk snapshots or backups.● D2 ECSs do not support automatic recovery.● D2 ECSs can use both local disks and EVS disks to store data. In addition, they

can have EVS disks attached to provide a larger storage size. Use restrictionson the two types of storage media are as follows:– Only an EVS disk, not a local disk, can be used as the system disk of a D2

ECS.– Both EVS disks and local disks can be used as data disks of a D2 ECS.– A D2 ECS can be attached with a maximum of 60 disks (including VBD,

SCSI, and local disks). Among the 60 disks, the maximum number of SCSIdisks is 30, and the maximum number of VBD disks is 24 (including thesystem disk). For details about constraints, see Can Multiple Disks BeAttached to an ECS?

– You are advised to use World Wide Names (WWNs), but not drive letters,in applications to perform operations on local disks to prevent drive letterdrift (low probability) on Linux. Take local disk attachment as anexample:If the local disk WWN is wwn-0x50014ee2b14249f6, run themount /dev/disk/by-id/wwn-0x50014ee2b14249f6 command.

NO TE

How can I view the local disk WWN?1. Log in to the ECS.2. Run the following command:

ll /dev/disk/by-id

● The basic resources, including vCPUs, memory, and image of a stopped D2ECS are still billed. To stop billing such an ECS, delete it.


2020-07-06 46



● The local disk data of a D2 ECS may be lost due to some reasons, such asphysical server breakdown or local disk damage. If your application does notuse the data reliability architecture, you are strongly advised to use EVS disksto build your ECS.

● When a D2 ECS is deleted, its local disk data is automatically deleted. Back upthe data before deleting such an ECS. Deleting local disk data is time-consuming. Therefore, a D2 ECS requires a longer period of time than otherECSs for releasing resources.

● Do not store service data for a long time in local disks. Instead, use EVS disksto store the data. In addition, back up data in a timely manner and use a highavailability architecture.

● You are not allowed to buy additional local disks. The quantity and capacityof your local disks are determined according to the specifications of your ECS.For D2 ECSs, if additional local disks are required, buy them when creating theECSs.

5.13 Ultra-high I/O ECSs

OverviewUltra-high I/O ECSs use high-performance local NVMe SSDs to provide highstorage input/output operations per second (IOPS) and low read/write latency.You can create such ECSs with high-performance local NVMe SSDs attached onthe management console.

Ultra-high I/O ECSs can be used for high-performance relational databases,NoSQL databases (such as Cassandra and MongoDB), and ElasticSearch.

Specifications

Table 5-41 I3 ECS specifications

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

LocalDisks

MaximumNICs

VirtualizationType

i3.2xlarge.8

8 64 8/3.5 100 4 1 x1600GBNVMe

4 KVM


2020-07-06 47

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

LocalDisks

MaximumNICs

VirtualizationType

i3.4xlarge.8

16 128 15/7 160 4 2 x1600GBNVMe

8 KVM

i3.8xlarge.8

32 256 20/14 280 8 4 x1600GBNVMe

8 KVM

i3.12xlarge.8

48 384 25/20 420 8 6 x1600GBNVMe

8 KVM

i3.15xlarge.8

60 512 25/25 500 16 7 x1600GBNVMe

8 KVM

i3.16xlarge.8

64 512 25/25 500 16 8 x1600GBNVMe

8 KVM

Features

Table 5-42 lists the IOPS performance of I3 ECSs.

Table 5-42 I3 ECS IOPS performance

Flavor Maximum IOPS for Random 4 KB Read

i3.2xlarge.8 750,000

i3.4xlarge.8 1,500,000

i3.8xlarge.8 3,000,000


2020-07-06 48

Flavor Maximum IOPS for Random 4 KB Read

i3.12xlarge.8 4,500,000

i3.15xlarge.8 5,250,000

Table 5-43 Specifications of a single NVMe disk attached to an I3 ECS

Metric Performance

Disk capacity 1.6 TB

IOPS for random 4 KB read 750,000

IOPS for random 4 KB write 200,000

Read throughput 2.9 GB/s

Write throughput 1.9 GB/s

Access latency Within microseconds

Notes● For details about the OSs supported by a ultra-high I/O ECS, see OSs

Supported by Different Types of ECSs.

● When the physical host where a ultra-high I/O ECS is deployed becomesfaulty, the ECS cannot be migrated.

● Ultra-high I/O ECSs do not support specifications modification.

● Ultra-high I/O ECSs do not support local disk snapshots or backups.

● Ultra-high I/O ECSs can use both local disks and EVS disks to store data. Inaddition, they can have EVS disks attached to provide a larger storage size.Use restrictions on the two types of storage media are as follows:

– Only an EVS disk, not a local disk, can be used as the system disk of aultra-high I/O ECS.

– Both EVS disks and local disks can be used as data disks of a ultra-highI/O ECS.

– A ultra-high I/O ECS can be attached with a maximum of 60 disks(including VBD, SCSI, and local disks). Among the 60 disks, the maximumnumber of SCSI disks is 30, and the maximum number of VBD disks is 22(including the system disk).

– You are advised to use World Wide Names (WWNs), but not drive letters,in applications to perform operations on local disks to prevent drive letterdrift (low probability) on Linux. Take local disk attachment as anexample:

If the local disk WWN is wwn-0x50014ee2b14249f6, run themount /dev/disk/by-id/wwn-0x50014ee2b14249f6 command.


2020-07-06 49

https://support.huaweicloud.com/intl/en-us/productdesc-ims/ims_01_0007.html

https://support.huaweicloud.com/intl/en-us/productdesc-ims/ims_01_0007.html

NO TE

How can I view the local disk WWN?

1. Log in to the ECS.

2. Run the following command:

ll /dev/disk/by-id

● The local disk data of a ultra-high I/O ECS may be lost due to some reasons,such as physical server breakdown or local disk damage. If the data reliabilityof your application cannot be ensured, you are strongly advised to use EVSdisks to build your ECS.

● After a ultra-high I/O ECS is deleted, the data on local NVMe SSDs isautomatically deleted. Back up the data before deleting such an ECS. Deletinglocal disk data is time-consuming. Therefore, a ultra-high I/O ECS requires alonger period of time than other ECSs for releasing resources.

● The data reliability of local disks depends on the reliability of physical serversand hard disks, which are SPOF-prone. Therefore, you are advised to performdata redundancy at the application layer to ensure data availability. Use EVSdisks to store long-term service data.

● The device name of a local disk attached to an I3 ECS is /dev/nvme0n1or /dev/nvme0n2.

● The basic resources, including vCPUs, memory, and image of a ultra-high I/OECS are still billed after the ECS is stopped. To stop billing such an ECS, deleteit.

5.14 High-Performance Computing ECSs

Overview

H3 ECSs use high-performance Intel® Xeon® Scalable processors. Each vCPUcorresponds to the hyper-thread of an Intel® Xeon® Scalable processor core,providing stable computing capabilities. H3 ECSs are suitable for high-performancecomputing services. In addition, the ECSs use latest-generation networkacceleration engines and DPDK rapid packet processing mechanism to providehigh network performance.

The vCPU/memory ratio of an HC2 ECS is 1:2 or 1:4. Each vCPU corresponds to thehyperthreading of an Intel® Xeon® Scalable processor core. HC2 ECSs can be usedfor high-performance computing services. They provide a large number of parallelcomputing resources and high-performance infrastructure services to meet therequirements of high-performance computing and massive storage and ensurerendering efficiency. HC2 ECSs are frequently used in the following scenarios:

● Computing and storage systems for genetic engineering, games, animations,and biopharmaceuticals

● Public rendering platforms for renderfarms and animation and film bases;other rendering platforms for movies and videos

● High-performance frontend clusters, web servers, high-performance scienceand engineering applications, advertisements, video coding, and distributedanalysis


2020-07-06 50

Specifications

Table 5-44 H3 ECS specifications

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

h3.large.2 2 4 2/1 30 2 KVM

h3.xlarge.2 4 8 4/2 60 2 KVM

h3.2xlarge.2

8 16 6/3.5 120 4 KVM

h3.3xlarge.2

12 24 6/5.5 160 4 KVM

h3.4xlarge.2

16 32 12/7.5 200 8 KVM

h3.large.4 2 8 2/1 30 2 KVM

h3.xlarge.4 4 16 4/2 60 2 KVM

h3.2xlarge.4

8 32 6/3.5 120 4 KVM

h3.3xlarge.4

12 48 6/5.5 160 4 KVM

h3.4xlarge.4

16 64 12/7.5 200 8 KVM

Table 5-45 HC2 ECS specifications

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

hc2.large.2 2 4 1.5/0.5 10 1 KVM

hc2.xlarge.2

4 8 3/1 15 1 KVM

hc2.2xlarge.2

8 16 5/2 30 2 KVM

hc2.4xlarge.2

16 32 8/4 40 4 KVM

hc2.large.4 2 8 1.5/0.5 10 1 KVM


2020-07-06 51

Flavor vCPUs

Memory (GB)


MaximumPPS(10,000)

NICMulti-Queue

VirtualizationType

hc2.xlarge.4

4 16 3/1 15 1 KVM

hc2.2xlarge.4

8 32 5/2 30 2 KVM

hc2.4xlarge.4

16 64 8/4 40 4 KVM

Scenarios● Computing and storage systems for genetic engineering, games, animations,

and biopharmaceuticals● Public rendering platforms for renderfarms and animation and film bases;

other rendering platforms for movies and videos● High-performance frontend clusters, web servers, high-performance science

and engineering applications, advertisements, video coding, and distributedanalysis

● Batch-processed workload, HPC applications, and SAP applications● Computing-intensive services, such as large-scale multiplayer online (MMO)

gaming

5.15 GPU-accelerated ECSsGPU-accelerated ECSs provide outstanding floating-point computing capabilities.They are suitable for applications that require real-time, highly concurrent massivecomputing.

GPU-accelerated ECS Types● G series

– Graphics-accelerated Enhancement G5● P series

– Computing-accelerated P2v– Inference-accelerated PI2– Inference-accelerated PI1

Helpful links:● Images Supported by GPU-accelerated ECSs● Installing a GRID Driver on a GPU-accelerated ECS● Installing a NVIDIA GPU Driver and CUDA Toolkit on a GPU-accelerated

ECS


2020-07-06 52

https://support.huaweicloud.com/intl/en-us/usermanual-ecs/en-us_topic_0149610914.html



Table 5-46 GPU-accelerated ECSs

Classification

ECSType

GPU Application Scenario Remarks

Graphics-accelerated

G5 NVIDIA V100 Cloud desktop, imagerendering, 3Dvisualization, andheavy-workloadgraphics design

Remote login onthe managementconsole isunavailable. Tolog in to such anECS, use VNC orthird-party VDI.

Computing-accelerated

P2v NVIDIA V100NVLink (GPUpassthrough)

Machine learning, deeplearning, inferencetraining, scientificcomputing, seismicanalysis, computingfinance, rendering,multimedia encodingand decoding

None

Inference-accelerated

PI2 NVIDIA T4(GPUpassthrough)


None

Inference-accelerated

PI1 NVIDIA P4(GPUpassthrough)


None

Images Supported by GPU-accelerated ECSs

Table 5-47 Image list

Classification ECS Type Supported Image

Graphics-accelerated G5 Windows Server 2016 Standard 64bitWindows Server 2012 R2 Standard 64bitCentOS 7.5 64bit


2020-07-06 53

Classification ECS Type Supported Image

Computing-accelerated

P2v Windows Server 2016 Standard 64bitWindows Server 2012 R2 Standard 64bitUbuntu 16.04 64bitCentOS 7.4 64bitEulerOS 2.2 64bit

Inference-accelerated PI2 Windows Server 2016 Standard 64bitUbuntu 16.04 64bitCentOS 7.5 64bit

Inference-accelerated PI1 Ubuntu 16.04 64bitUbuntu 14.04 64bitCentOS 7.3 64bit

Graphics-accelerated Enhancement G5

Overview

G5 ECSs use NVIDIA Tesla V100 GPUs and support DirectX, OpenGL, and Vulkan.These ECSs provide 16 GB of GPU memory and up to 4096 x 2160 resolution,meeting requirements on professional graphics processing.

Select your desired GPU-accelerated ECS type and specifications.

Specifications

Table 5-48 G5 ECS specifications

Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUMemory(GB)

Virtualization Type

g5.8xlarge.4

32 128 25/15 200 16 1 xV100

16 KVM

NO TE

A g5.8xlarge.4 ECS exclusively uses a V100 GPU for professional graphics acceleration. Suchan ECS can be used for heavy-workload CPU inference.

G5 ECS Features


2020-07-06 54

● Graphics acceleration APIs– DirectX 12, Direct2D, DirectX Video Acceleration (DXVA)– OpenGL 4.5– Vulkan 1.0

● CUDA* and OpenCL● NVIDIA V100 GPUs● Graphics acceleration applications● Heavy-workload CPU inference● Application flow identical to common ECSs● Automatic scheduling of G5 ECSs to AZs where NVIDIA V100 GPUs are used● A maximum specification of 16 GB of GPU memory and 4096 x 2160

resolution for processing graphics and videos

Supported Common Software

G5 ECSs are used in graphics acceleration scenarios, such as video rendering, clouddesktop, and 3D visualization. If the software relies on GPU DirectX and OpenGLhardware acceleration, use G5 ECSs. G5 ECSs support the following commonlyused graphics processing software:● AutoCAD● 3DS MAX● MAYA● Agisoft PhotoScan● ContextCapture

Notes

● G5 ECSs support the following OSs:– Windows Server 2016 Standard 64bit– Windows Server 2012 R2 Standard 64bit– CentOS 7.5 64bit

● A G5 ECS requires the configuration of a GRID license after the ECS is created.● G5 ECSs created using a public image have had the GRID driver of a specific

version installed by default. However, you need to purchase and configure theGRID license by yourself. Ensure that the GRID driver version meets servicerequirements.

For instructions about how to configure a GRID license, see Installing a GRIDDriver on a GPU-accelerated ECS.

● If a G5 ECS is created using a private image, make sure that the GRID driverhas been installed during the private image creation. If not, install the driverfor graphics acceleration after the ECS is created.

For details, see Installing a GRID Driver on a GPU-accelerated ECS.

Computing-accelerated P2vOverview


2020-07-06 55




Compared with P2 ECSs, P2v ECSs use NVIDIA Tesla V100 GPUs to provideflexibility, high-performance computing, and cost-effectiveness. These ECSs useGPU NVLink for direct communication between GPUs, improving datatransmission efficiency. P2v ECSs provide outstanding general computingcapabilities and have strengths in AI-based deep learning, scientific computing,Computational Fluid Dynamics (CFD), computing finance, seismic analysis,molecular modeling, and genomics.

Specifications

Table 5-49 P2v ECS specifications

Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUConnection

GPUMemory(GB)

VirtualizationType

p2v.2xlarge.8

8 64 10/4 50 4 1 xV100

N/A 1 x16

KVM

p2v.4xlarge.8

16 128 15/8 100 8 2 xV100

NVLink

2 x16

KVM

p2v.8xlarge.8

32 256 25/15 200 16 4 xV100

NVLink

4 x16

KVM

p2v.16xlarge.8

64 512 30/30 400 32 8 xV100

NVLink

8 x16

KVM

P2v ECS Features● Up to eight NVIDIA Tesla V100 GPUs on an ECS● NVIDIA CUDA parallel computing and common deep learning frameworks,

such as TensorFlow, Caffe, PyTorch, and MXNet● 15.7 TFLOPS of single-precision computing and 7.8 TFLOPS of double-

precision computing● NVIDIA Tensor cores with 125 TFLOPS of single- and double-precision

computing for deep learning● Up to 30 GB/s of network bandwidth on a single ECS● 16 GB of HBM2 GPU memory with a bandwidth of 900 GB/s● Comprehensive basic capabilities

Networks are user-defined, subnets can be divided, and network accesspolicies can be configured as needed. Mass storage is used, elastic capacity


2020-07-06 56

expansion as well as backup and restoration are supported to make datamore secure. Auto Scaling allows you to add or reduce the number of ECSsquickly.

● Flexibility

Similar to other types of ECSs, P2v ECSs can be provisioned in a few minutes.

● Excellent supercomputing ecosystem

The supercomputing ecosystem allows you to build up a flexible, high-performance, cost-effective computing platform. A large number of HPCapplications and deep-learning frameworks can run on P2v ECSs.


P2v ECSs are used in computing acceleration scenarios, such as deep learningtraining, inference, scientific computing, molecular modeling, and seismic analysis.If the software is required to support GPU CUDA, use P2v ECSs. P2v ECSs supportthe following commonly used software:

● Common deep learning frameworks, such as TensorFlow, Caffe, PyTorch, andMXNet

● CUDA GPU rendering supported by RedShift for Autodesk 3dsMax and V-Rayfor 3ds Max

● Agisoft PhotoScan

● MapD

Notes

● P2v ECSs support the following OSs:

– Windows Server 2016 Standard 64bit

– Windows Server 2012 R2 Standard 64bit

– Ubuntu Server 16.04 64bit

– CentOS 7.4 64bit

– EulerOS 2.2 64bit

● P2v ECSs created using a public image have had the Tesla driver installed bydefault.

● If a P2v ECS is created using a private image, make sure that the Tesla driverhas been installed during the private image creation. If not, install the driverfor computing acceleration after the ECS is created. For details, see Installinga Tesla Driver and CUDA Toolkit on a GPU-accelerated ECS.

Inference-accelerated PI2

Overview

PI2 ECSs use NVIDIA Tesla T4 GPUs dedicated for real-time AI inference. TheseECSs use the T4 INT8 calculator for up to 130 TOPS of INT8 computing. The PI2ECSs can also be used for light-workload training.

Specifications


2020-07-06 57




Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUMemory(GB)

LocalDisks

Virtualization Type

pi2.2xlarge.4

8 32 10/4 50 4 1 xT4

1 x16

N/A KVM

pi2.4xlarge.4

16 64 15/8 100 8 2 xT4

2 x16

N/A KVM

pi2.8xlarge.4

32 128 25/15 200 16 4 xT4

4 x16

N/A KVM

PI2 ECS Features

● Up to four NVIDIA Tesla T4 GPUs on an ECS● GPU hardware passthrough● Up to 8.1 TFLOPS of single-precision computing on a single GPU● Up to 130 TOPS of INT8 computing on a single GPU● 16 GB of GDDR6 GPU memory with a bandwidth of 300 GB/s on a single GPU● One built-in NVENC and two NVDEC GPUs


PI2 ECSs are used in GPU-based inference computing scenarios, such as imagerecognition, voice recognition, and natural language processing. The PI2 ECSs canalso be used for light-workload training.

PI2 ECSs support the following commonly used software:

● Deep learning frameworks, such as TensorFlow, Caffe, PyTorch, and MXNet

Notes

● The basic resources, including vCPUs, memory, and image of a pay-per-usePI2 ECS of flavor pi2.2xlarge.4, pi2.4xlarge.4, or pi2.8xlarge.4 are not billedafter the ECS is stopped, but the system disk of the ECS is still being billedaccording to the disk capacity. The resources associated with the ECS, such asEVS disks, EIPs, and bandwidth, are separately billed.

NO TE

The resources of a pay-per-use PI2 ECS of flavor pi2.2xlarge.4, pi2.4xlarge.4, orpi2.8xlarge.4 are released after the ECS is stopped. If the backend resources areinsufficient when the ECS is started, starting the ECS may fail. If you want to use suchan ECS for a long period of time, change its billing mode to yearly/monthly or do notstop the ECS.


2020-07-06 58

● PI2 ECSs support the following OSs:– Windows Server 2016 Standard 64bit– Ubuntu Server 16.04 64bit– CentOS 7.5 64bit

● PI2 ECSs support automatic recovery when the hosts accommodating suchECSs become faulty.

● PI2 ECSs created using a public image have had the Tesla driver installed bydefault.

● If a PI2 ECS is created using a private image, make sure that the Tesla driverhas been installed during the private image creation. If not, install the driverfor computing acceleration after the ECS is created. For details, see Installinga Tesla Driver and CUDA Toolkit on a GPU-accelerated ECS.

Inference-accelerated PI1

Overview

PI1 ECSs use NVIDIA Tesla P4 GPUs dedicated for real-time AI inference. Workingwith P4 INT8 calculators, PI1 ECSs have shortened the inference latency by 15times. Working with hardware decoding engines, PI1 ECSs concurrently supportreal-time 35-channel HD video transcoding and inference.

Specifications


Flavor vCPUs

Memory(GB)


Max.PPS(10,000)

NICMulti-Queue

GPUs

GPUMemory(GB)

LocalDisks

VirtualizationType

pi1.2xlarge.4

8 32 5/1.6 40 2 1 xP4

1 x 8GB

N/A KVM

pi1.4xlarge.4

16 64 8/3.2 70 4 2 xP4

2 x 8GB

N/A KVM

pi1.8xlarge.4

32 128 10/6.5 140 8 4 xP4

4 x 8GB

N/A KVM

PI1 ECS Features● Up to four NVIDIA Tesla P4 GPUs on an ECS● GPU hardware passthrough● Up to 5.5 TFLOPS of single-precision computing on a single GPU● Up to 22 TOPS of INT8 computing on a single GPU


2020-07-06 59



● 8 GB of ECC GPU memory with a bandwidth of 192 GB/s on a single GPU● Hardware video encoding and decoding engines embedded in GPUs for

concurrent real-time 35-channel HD video transcoding and inference


PI1 ECSs are used in GPU-based inference computing scenarios, such as imagerecognition, voice recognition, and natural language processing.

PI1 ECSs support the following commonly used software:

● Deep learning frameworks, such as TensorFlow, Caffe, PyTorch, and MXNet

Notes● The basic resources, including vCPUs, memory, and image of a pay-per-use

PI1 ECS of flavor pi1.2xlarge.4, pi1.4xlarge.4, or pi1.8xlarge.4 are not billedafter the ECS is stopped, but the system disk of the ECS is still being billedaccording to the disk capacity. The resources associated with the ECS, such asEVS disks, EIPs, and bandwidth, are separately billed.

NO TICE

The resources of a pay-per-use PI1 ECS of flavor pi1.2xlarge.4, pi1.4xlarge.4, orpi1.8xlarge.4 are released after the ECS is stopped. If the backend resourcesare insufficient when the ECS is started, starting the ECS may fail. If you wantto use such an ECS for a long period of time, change its billing mode toyearly/monthly or do not stop the ECS.

● PI1 ECSs do not support specifications modification.● PI1 ECSs support the following OSs:

– Ubuntu Server 16.04 64bit– Ubuntu Server 14.04 64bit– CentOS 7.3 64bit

● PI1 ECSs support automatic recovery when the hosts accommodating suchECSs become faulty.

● PI1 ECSs created using a public image have had the Tesla driver installed bydefault.

● If a PI1 ECS is created using a private image, make sure that the Tesla driverhas been installed during the private image creation. If not, install the driverfor computing acceleration after the ECS is created. For details, see Installinga NVIDIA GPU Driver and CUDA Toolkit on a GPU-accelerated ECS.


2020-07-06 60



6 Images

What Is Image?An image is an ECS template that contains an OS and may also containproprietary software and application software, such as database software. You canuse images to create ECSs.

Images can be public, private, or shared. Public images are provided by the systemby default, private images are manually created, and shared images are privateimages that are shared by another user. You can use any type of image to createan ECS. You can also create a private image using an existing ECS. This providesyou with a simple way to create ECSs that comply with your service requirements.For example, if you use web services, your image can contain web serverconfigurations, static configurations, and dynamic page code. After you use thisimage to create an ECS, the web server will run.

Image TypesImage Type Description

Public image A public image is a standard, widely used image. It contains anOS and preinstalled public applications and is available to allusers. Public images are highly stable and authorized. You canconfigure the application environment or related software asrequired.Public images support the following OSs: Windows, CentOS,Debian, openSUSE, Fedora, Ubuntu, EulerOS, and CoreOS.For more information about public images, see ManagingPublic Images.

Elastic Cloud ServerService Overview 6 Images

2020-07-06 61

https://support.huaweicloud.com/intl/en-us/usermanual-ims/ims_01_0101.html


Image Type Description

Privateimage

A private image is an image that contains an OS or service data,pre-installed public applications, and the owner's privateapplications. It is available only to the user who created it.A private image can be a system disk image, full-ECS image, ordata disk image.● System disk image: contains an OS and pre-installed

application software for running services. You can use asystem disk image to create ECSs and migrate your servicesto the cloud.

● Data disk image: contains only the owner's service data. Youcan also use a data disk image to create EVS disks andmigrate your service data to the cloud.

● Full-ECS image: contains an OS, pre-installed applicationsoftware, and service data. A full-ECS image is created usingdifferential backups and the creation takes a shorter timethan creating a system or data disk image with the same disksize.

Sharedimage

A shared image is an image shared by another tenant with you.For more about shared images, see Sharing Images.

Marketplaceimage

The Marketplace is a store where you can purchase third-partyimages that have the OS, application environment, and softwarepreinstalled. You can use the images to deploy websites andapplication development environments with a few clicks. Noadditional configuration is required.Marketplace images are provided by service providers who haverich experience in configuring and maintaining cloud servers. Allthe images are strictly tested and approved by HUAWEI CLOUDbefore being published.

Elastic Cloud ServerService Overview 6 Images

2020-07-06 62


7 EVS Disks

What Is Elastic Volume Service?Elastic Volume Service (EVS) offers scalable block storage for ECSs. With highreliability, high performance, and rich specifications, EVS disks can be used fordistributed file systems, development and test environments, data warehouseapplications, and high-performance computing (HPC) scenarios to meet diverseservice requirements.

Disk TypesECSs support the following types of EVS disks for storing data:

● Common I/O: EVS disks of this type deliver a maximum of 2200 IOPS. Thisdisk type is suitable for application scenarios that require large capacity, amedium read/write rate, and fewer transactions, such as enterprise officeapplications and small-scale testing.

● High I/O: EVS disks of this type deliver a maximum of 5000 IOPS and aminimum of 1 ms read/write latency. This disk type is designed to meet theneeds of mainstream high-performance, high-reliability application scenarios,such as enterprise applications, small- and medium-scale development andtesting, and web server logs.

● Ultra-high I/O: EVS disks of this type deliver a maximum of 33,000 IOPS and aminimum of 1 ms read/write latency. This disk type is excellent for ultra-highI/O, ultra-high bandwidth, and read/write-intensive application scenarios, suchas distributed file systems in HPC scenarios or NoSQL/RDS in I/O-intensivescenarios.

EVS disks with different I/O capacities provide different features at different prices.Choose EVS disks based on your requirements. For more information about EVSdisk specifications and performance, see Elastic Volume Service User Guide.

Elastic Cloud ServerService Overview 7 EVS Disks

2020-07-06 63

https://support.huaweicloud.com/intl/en-us/evs/index.html

8 Network

VPC

Virtual Private Cloud (VPC) allows you to create customized virtual networks inyour logically isolated AZ. Such networks are dedicated zones that are logicallyisolated for your ECSs. You can define security groups, virtual private networks(VPNs), IP address segments, and bandwidth for a VPC. This facilitates internalnetwork configuration and management as well as secure and convenientnetwork modification. You can also customize the ECS access rules within asecurity group and between security groups to improve ECS security.

For more information about VPC, see Virtual Private Cloud User Guide.

Subnet

A subnet is a range of IP addresses in your VPC and provides IP addressmanagement and DNS resolution functions for ECSs in it. The IP addresses of allECSs in a subnet belong to the subnet.

Figure 8-1 Subnet

By default, ECSs in all subnets of the same VPC can communicate with each other,while ECSs in different VPCs cannot.

Security Group

A security group is a collection of access control rules for ECSs that have the samesecurity protection requirements and that are mutually trusted within a VPC. After

Elastic Cloud ServerService Overview 8 Network

2020-07-06 64

https://support.huaweicloud.com/intl/en-us/vpc/index.html

a security group is created, you can create different access rules for the securitygroup to protect the ECSs that are added to this security group.

Your account automatically comes with a default security group. The defaultsecurity group allows all outbound data, denies all inbound data, and allows alldata between ECSs in the group. Your ECSs in the security group can communicatewith each other without the need to add rules.

Figure 8-2 Default security group

Table 8-1 describes default security group rules.

Table 8-1 Default security group rules

Direction

Protocol Port/Range

Source/Destination

Description

Outbound

All All Destination:0.0.0.0/0

Allow all outboundtraffic.

Inbound All All Source: ID of thecurrent securitygroup (forexample, sg-xxxxx)

Allow communicationamong ECSs withinthe security groupand deny all inboundtraffic (incoming datapackets).

Inbound TCP 22 Source: 0.0.0.0/0 Allow all IP addressesto access Linux ECSsover SSH.

Inbound TCP 3389 Source: 0.0.0.0/0 Allow all IP addressesto access WindowsECSs over RDP.

EIP

An EIP is a public IP address that can be directly accessed over the Internet. An EIPconsists of the public IP address and public network egress bandwidth. EIPs can bebound to or unbound from ECSs, BMSs, virtual IP addresses, NAT gateways, or


2020-07-06 65

load balancers. Various billing modes are provided to meet diversified servicerequirements.

Each EIP can be used by only one cloud resource at a time.

Figure 8-3 Accessing the Internet using an EIP


2020-07-06 66

9 Security

9.1 User EncryptionUser encryption allows you to use the encryption feature provided on the publiccloud platform to encrypt ECS resources, improving data security. User encryptionincludes image encryption and EVS disk encryption.

Image Encryption

Image encryption supports encrypting private images. When creating an ECS, ifyou select an encrypted image, the system disk of the created ECS is automaticallyencrypted, improving data security.

Use either of the following methods to create an encrypted image:

● Use an existing encrypted ECS.● Use an external image file.

For more information about image encryption, see Encrypting Images.

EVS Disk Encryption

EVS disk encryption supports system disk encryption and data disk encryption.

● When creating an ECS, you can encrypt added data disks.● When creating an ECS, if you select an encrypted image, the system disk of

the created ECS automatically has encryption enabled, and the encryptionmode complies with the image encryption mode.

For more information about EVS disk encryption, see EVS Disk Encryption.

Impact on AS

If you use an encrypted ECS to create an Auto Scaling (AS) configuration, theencryption mode of the created AS configuration complies with the ECS encryptionmode.

Elastic Cloud ServerService Overview 9 Security

2020-07-06 67


https://support.huaweicloud.com/intl/en-us/productdesc-evs/evs_01_0001.html

About KeysThe key required for encryption relies on Data Encryption Workshop (DEW). DEWuses a data encryption key (DEK) to encrypt data and a customer master key(CMK) to encrypt the DEK.

Figure 9-1 Data encryption process

Table 9-1 describes the keys involved in the data encryption process.

Table 9-1 Keys

Name Description Function

DEK An encryption key that is used forencrypting data.

Encrypts specific data.

CMK An encryption key created using DEWfor encrypting DEKs.A CMK can encrypt multiple DEKs.

Supports CMKdisabling andscheduled deletion.

Default CMK A master key automatically generatedby the system when you use DEW forencryption for the first time.The name extension of a default CMKis /default, for example, evs/default.

● Supports viewingdetails of thedefault CMK onthe KMS console.

● Does not supportCMK disabling orscheduled deletion.


2020-07-06 68

NO TE

After disabling a CMK or scheduling the deletion of a CMK takes effect, the EVS diskencrypted using this CMK can still be used until the disk is detached from and thenattached to an ECS again. During this process, the disk fails to be attached to the ECSbecause the CMK cannot be obtained. Therefore, the EVS disk becomes unavailable.

For details about DEW, see Data Encryption Workshop User Guide.

9.2 Cloud-InitCloud-Init is an open-source cloud initialization program, which initializes specifiedcustomized configurations, such as the hostname, key pair, and user data, of anewly created ECS.

Using Cloud-Init to initialize your ECSs will affect your ECS, IMS, and AS services.

Impact on IMS

To ensure that ECSs created using a private image support customizedconfigurations, you must install Cloud-Init or Cloudbase-Init before creating theprivate image.

● For Windows OSs, download and install Cloudbase-Init.

● For Linux OSs, download and install Cloud-Init.

After Cloud-Init or Cloudbase-Init is installed in an image, Cloud-Init orCloudbase-Init automatically configures initial ECS attributes when the ECS iscreated.

For instructions about the installation, see Installing Cloud-Init.

Impact on ECS● When creating an ECS, if the selected image supports Cloud-Init, you can use

user data injection to inject customized configuration, such as ECS loginpassword, for initializing.

● After Cloud-Init is supported, you can view and use metadata to configureand manage running ECSs.

Impact on AS● When creating an AS configuration, you can use user data injection to specify

ECS configurations for initialization. If the AS configuration has taken effect inan AS group, the ECSs newly created in the AS group will automaticallyinitialize their configurations.

● For an existing AS configuration, if its private image does not have Cloud-Initor Cloudbase-Init installed, the login mode of the ECSs created in the ASgroup where the AS configuration takes effect will be affected.

To resolve this issue, see "How Does Cloud-Init Affect the AS Service?" inAuto Scaling User Guide.


2020-07-06 69

https://support.huaweicloud.com/intl/en-us/dew/index.html

https://support.huaweicloud.com/intl/en-us/usermanual-ims/en-us_topic_0030730603.html

https://support.huaweicloud.com/intl/en-us/as_faq/as_faq_0013.html

Notes● When using Cloud-Init, enable DHCP in the VPC to which the ECS belongs.● When using Cloud-Init, ensure that security group rules in the outbound

direction meet the following requirements:– Protocol: TCP– Port Range: 80– Remote End: 169.254.0.0/16

NO TE

If you use the default security group rules in the outbound direction, the precedingrequirements are met, and the metadata can be accessed. Default security group rulesin the outbound direction are as follows:● Protocol: ANY● Port Range: ANY● Remote End: 0.0.0.0/0


2020-07-06 70

10 Billing

Billing ItemsHUAWEI CLOUD ECSs are billed based on ECS specifications and service duration.

Table 10-1 ECS billing

Billing Item Description

ECS Pricing is based on the ECS type, specifications (includingvCPUs and memory), service duration, and the number ofpurchased ECSs.For details, see Elastic Cloud Server Pricing Details.

Image Public images of the community edition, such as Linux, arefree of charge. Other commercial images, such asWindows, are billed.

EVS disk Purchasing EVS is mandatory; by default, the system diskis 40 GB. EVS disks can also be purchased based on aspecified quantity or validity period. (However, it isrecommended that the validity period is the same for boththe purchased EVS disk and its associated ECS.) For furtherdetails, see EVS disk billing.

EIP A public IP address is required only if the ECS accesses theInternet. For details about pricing, see EIP billing.

Bandwidth The bandwidth configuration is mandatory only if the ECSis billed by bandwidth. For further details about pricing,see bandwidth billing.

Pay-per-use billingmode

Pay-per-use ECSs are billed down to the second.

Yearly/Monthlybilling mode

Yearly/Monthly ECSs are billed on a yearly or monthlybasis.

Elastic Cloud ServerService Overview 10 Billing

2020-07-06 71

https://www.huaweicloud.com/intl/en-us/pricing/index.html?tab=detail#/ecs

https://huaweicloud.com/intl/en-us/price_detail.html#/evs_detail

https://huaweicloud.com/intl/en-us/price_detail.html#/eip_detail

https://huaweicloud.com/intl/en-us/price_detail.html#/eip_detail

Billing Item Description

Spot price billingmode

Spot price ECSs are billed based on the market price,which varies according to the changes in supply anddemand.

Billing ModesAn ECS can be billed on a pay-per-use, spot price, or yearly/monthly basis.

Table 10-2 lists the differences between the billing modes.

Table 10-2 Billing modes

Billingmode

Yearly/Monthly Pay-per-use Spot price

Paymentmethod

PrepaidBilled by thepurchase periodspecified in theorder.

PostpaidBilled by serviceduration.

PostpaidBilled by market price,which varies accordingto the changes in supplyand demand. The startprice of the bill is themarket price when theECS was purchased, andthe market price of thehour is used for billing.Spot Pricing

Billingperiod

Billed by thepurchase periodspecified in theorder.

Billed by the second.A bill is generated onthe hour.

Billed by the second. Abill is generated on thehour.


2020-07-06 72


Billingmode


No feesforstoppedECSs

Billed by thepurchase periodspecified in theorder, regardlessof whether theECS is stopped ornot.

● Basic resources,including vCPUs,memory, andimage, of an ECSwithout a localdisk or FPGAattached are notbilled after theECS is stopped.Other resourcesassociated withthe ECS, such asEVS disks and EIP,are still billed.

● An ECS with alocal disk attached(such as disk-intensive or GPU-accelerated ECS)or an FPGAattached is stillbilled after theECS is stopped.The vCPU andmemory resourceswill be retainedfor the ECS.

● Basic resources,including vCPUs,memory, and image,of an ECS without alocal disk or FPGAattached are notbilled after the ECS isstopped. Otherresources associatedwith the ECS, such asEVS disks and EIP, arestill billed.

● An ECS with a localdisk attached (suchas disk-intensive orGPU-accelerated ECS)or an FPGA attachedis still billed after theECS is stopped. ThevCPU and memoryresources will beretained for the ECS.

Changing thebillingmode

Can be changedto pay-per-use.Pay-per-usebilling modetakes effect onlyafter thepurchase periodexpires.ChangingYearly/Monthlyto Pay-per-Use

Can be changed toyearly/monthly.Changing Pay-per-Use to Yearly/Monthly

Cannot be changed toyearly/monthly or pay-per-use.

Modifyingspecifications

Supported Supported Not supported


2020-07-06 73







Billingmode


Applicationscenarios

This cost-effective mode isideal when theduration of ECSusage ispredictable. Theyearly/monthlymode isrecommendedfor long-termusers.

This mode is idealwhen you want moreflexibility and controlon ECS usage.

This mode is ideal forapplications that requireshort-term running butnot stability.

● Yearly/Monthly: HUAWEI CLOUD allows ECSs to be billed on a yearly ormonthly basis. Yearly/Monthly payment provides a larger discount than pay-per-use and is recommended for long-term users. A yearly/monthly ECS isbilled based on the purchase period specified in the order.

● Pay-per-use: allows for flexibility, accurately monitoring usage down to thesecond.An ECS is billed from the time when it is provisioned to the time when it isdeleted.After no fees for stopped ECS is enabled, when a pay-per-use ECS without alocal disk or FPGA attached is stopped, the ECS (including its vCPUs, memory,and image) is not billed. However, the resources associated with the ECS, suchas its EVS disks, EIP, and bandwidth, are still billed. The vCPU and memoryresources of the stopped ECS are released. When the ECS is restarted, thevCPU and memory resources must be requested again. However, if theresources are insufficient, the startup may fail. In such a case, wait severalminutes before attempting another restart or modify the ECS specifications.An ECS with a local disk attached (such as disk-intensive or GPU-acceleratedECS) or an FPGA attached is still billed after the ECS is stopped. The vCPU andmemory resources will be retained for the ECS.

● Spot pricePrice: Spot price ECSs are billed based on the market price, which variesaccording to the changes in supply and demand. The quotation you set duringECS purchasing is not used as a billing basis. A higher quotation makes itmore likely for you to purchase such an ECS. A spot price ECS can be usedonly when the market price is lower than your quotation and inventoryresources are sufficient. When the market price exceeds your quotation, theECS will be reclaimed.Billing period: A spot price ECS is billed by the second. A bill is generated onthe hour. The start price of the bill is the market price when the ECS waspurchased, and the market price of the hour is used for billing.Associated services: Spot prices apply only to vCPUs and memory. The OS,system disk, data disk, bandwidth, and IP address are billed following thebilling rules for these items in pay-per-use billing mode. A system disk is


2020-07-06 74

created and released with the ECS to which the system disk is attached. Adata disk must be manually deleted.Coupons: Only cash coupons and stored value cards are applicable.

Billing Involved in Configuration Modifications● Changing the billing mode

The pay-per-use method of payment can be changed to the yearly/monthlymethod of payment and vice versa. If the billing mode is changed, a neworder is then generated for you. When adjusting from pay-per-use to yearly/monthly, the new billing mode takes effect immediately after you pay for theorder. However, when changing from yearly/monthly to pay-per-use, the newbilling mode takes effect only after the yearly/monthly payment period haselapsed. A spot price ECS cannot be changed to a pay-per-use or yearly/monthly ECS.

Figure 10-1 Changing the billing mode

● Modifying ECS SpecificationsThe specifications of a pay-per-use or yearly/monthly ECS can be modified, forexample, its vCPU and memory specifications can be modified. But a spotprice ECS does not support specifications modification.Notes– Vouchers will not be refunded if the specifications of the ECS purchased

with those vouchers are downgraded.– If ECS specifications are upgraded, the price difference between the

original and new specifications must be returned according to the in-service duration.

– ECS specifications (vCPU or memory) degrade deteriorates the ECSperformance.

– The price difference must be reimbursed if a downgraded ECS needs tobe upgraded back to its original specifications.


2020-07-06 75

Helpful Links● What Are the Differences Between Yearly/Monthly and Pay-per-Use

Payments?● Is a Pay-per-Use ECS Billed After Being Stopped?● How Do I Change the Billing Mode of an ECS from Yearly/Monthly to Pay-

Per-Use?● How Can ECS Billing Be Stopped?● FAQs About Spot Price ECSs


2020-07-06 76



https://support.huaweicloud.com/intl/en-us/ecs_faq/en-us_topic_0018124776.html





11 CPU Credits

ConceptCPU credits measure computing, storage, and network resource usage of an ECS.ECSs use CPU credits to ensure baseline performance, preventing issues caused byCPU overcommitment.

CPU-credit-based ECSs are suitable for the applications requiring baseline level ofvCPU performance generally and burstable performance in case of traffic bursts.

General computing-basic ECSs run based on CPU credits. For more details, see 5.9General Computing-Basic ECSs.

Working RulesAfter a CPU-credit-based ECS is created, the cloud platform automaticallyallocates initial CPU credits to the ECS for its burstable performance.

After the ECS runs, its credits are accrued or spent. When the actual computingperformance of the ECS is higher than the baseline CPU performance, the CPUcredits are spent to meet the performance requirements. When the actualcomputing performance is lower than the baseline CPU performance, the CPUcredits are accrued until the CPU credit balance limit is reached.

NO TE

● CPU credits can be accrued. However, after the credits reach the CPU credit balancelimit, any new credits that are earned will be discarded.

● Initial credits are not counted in the CPU credit balance limit.● When an ECS starts to spend CPU credits, it preferentially uses the initial CPU credits.● One CPU credit is equal to one vCPU running at 100% usage for one minute.● When the actual computing performance is higher than the baseline performance, the

accrued credits are spent until they are used up. Then, the actual computingperformance cannot exceed the baseline performance.

Elastic Cloud ServerService Overview 11 CPU Credits

2020-07-06 77

Related Terms

Table 11-1 Terms related to CPU credits (taking a T6 ECS as an example)

Term Description Example

Initial CPUcredits

After a T6 ECS is created, thecloud platform automaticallyallocates CPU credits to this ECS.These credits are initial CPUcredits. Initial CPU credits areallocated only after an ECS iscreated.

After a t6.large.1 is created,it has 60 initial CPU credits.

CPU creditbalance limit

When the actual computingperformance is lower than thebaseline CPU performance, theCPU credits are accrued. Theaccrued credits will not expire ona running ECS. When the creditsreach the maximum valueallowed, which is specified by theCPU credit balance limit, any newcredits that are earned will bediscarded. The CPU creditbalance limit varies depending onECS flavors.

The CPU credit balance limitfor a t6.large.1 ECS is 576.When its accrued CPUcredits reach 576, no morecredits will be accrued.When its accrued CPUcredits are smaller than 576,the CPU credits can beaccrued again.

CPU creditearn rate(credits/hour)

The number of CPU creditsearned by an ECS per hour, whichcorresponds to CPU baseline.One CPU credit is equal to onevCPU running at 100% usage forone minute.

The CPU credit earn rate ofa t6.large.1 ECS is 24,indicating that a t6.large.1ECS can earn 24 CPU creditsper hour.

CPU baseline(%)

When the number of CPU creditsthat an ECS spends per minute isthe same as the number of CPUcredits that the ECS earns perminute, the ECS runs at the CPUbaseline.

The CPU baseline of at6.large.1 ECS is 40%. Whenthe actual computingperformance of a t6.large.1ECS reaches 40%, thenumber of credits spent bythe ECS per minute is thesame as the number ofcredits earned by the ECSper minute.


2020-07-06 78


Average CPUbaseline (%)

When an ECS runs at CPUbaseline, the computingperformance of each vCPU is theaverage CPU baseline, which iscalculated using the followingformula:Average CPU baseline = CPUbaseline/Number of vCPUs

The CPU baseline of at6.large.1 ECS is 40%, andthe ECS has two vCPUs.Then, the average CPUbaseline is 20%.

Spent CPUcredits

When the actual computingperformance of an ECS is higherthan the baseline CPUperformance, the CPU credits arespent to meet the performancerequirements.One CPU credit is spent for onevCPU running at 100% usage forone minute.The formula for calculating theCPU credits spent per minute isas follows:Number of CPU credits spentper minute = 1 CPU credit xActual computing performance

When a t6.large.1 ECS runsat the computingperformance of 20% for oneminute, the ECS spends 0.2CPU credits.


2020-07-06 79


Accrued CPUcredits

● When the actual computingperformance of an ECS is lessthan the baseline CPUperformance, the number ofCPU credits spent per minuteis smaller than the number ofCPU credits earned perminute. Therefore, theremaining CPU credits areaccrued until the CPU creditbalance limit is reached.

● When the actual computingperformance is higher thanthe baseline CPUperformance, the number ofCPU credits spent per minuteis greater than the number ofCPU credits earned perminute. In such a case, theECS spends accrued CPUcredits (initial CPU creditspreferentially used) to complywith burstable CPUperformance.

The formula for calculating thenumber of CPU credits accruedper minute is as follows:Number of CPU credits accruedper minute = 1 CPU credit x(CPU baseline – Actualcomputing performance)

The CPU baseline of at6.large.1 ECS is 40%. Whenthe actual computingperformance of the ECS is10%, the ECS accrues 0.3CPU credits per minute.

Impact of CPU Credits After an ECS Is StoppedThe change of CPU credits varies depending on the ECS billing mode and networktype.

Table 11-2 Billing modes and CPU credits

Billing Mode CPU Credit Change After an ECS Is Stopped

Yearly/Monthly The existing CPU credits are retained and accrued untilthe CPU credit balance limit is reached.

Pay-per-use The existing CPU credits are retained but not accrued.

Spot price The existing CPU credits are retained but not accrued.


2020-07-06 80

12 Region and AZ

Concept

A region and availability zone (AZ) identify the location of a data center. You cancreate resources in a specific region and AZ.

● Regions are divided based on geographical location and network latency.Public services, such as Elastic Cloud Server (ECS), Elastic Volume Service(EVS), Object Storage Service (OBS), Virtual Private Cloud (VPC), Elastic IP(EIP), and Image Management Service (IMS), are shared within the sameregion. Regions are classified into universal regions and dedicated regions. Auniversal region provides universal cloud services for common tenants. Adedicated region provides specific services for specific tenants.

● An AZ contains one or more physical data centers. Each AZ has independentcooling, fire extinguishing, moisture-proof, and electricity facilities. Within anAZ, computing, network, storage, and other resources are logically dividedinto multiple clusters. AZs within a region are interconnected using high-speed optical fibers to support cross-AZ high-availability systems.

Figure 12-1 shows the relationship between regions and AZs.

Figure 12-1 Regions and AZs

HUAWEI CLOUD provides services in many regions around the world. Select aregion and AZ based on requirements. For more information, see HUAWEI CLOUDGlobal Regions.

Elastic Cloud ServerService Overview 12 Region and AZ

2020-07-06 81

https://www.huaweicloud.com/intl/en-us/global/

https://www.huaweicloud.com/intl/en-us/global/

Selecting a RegionWhen selecting a region, consider the following factors:

● LocationIt is recommended that you select the closest region for low network latencyand quick access. Regions within the Chinese mainland provide the sameinfrastructure, BGP network quality, as well as resource operations andconfigurations. Therefore, if your target users are on the Chinese mainland,you do not need to consider the network latency differences when selecting aregion.– If your target users are in Asia Pacific (excluding the Chinese mainland),

select the AP-Hong Kong, AP-Bangkok, or AP-Singapore region.– If your target users are in Africa, select the AF-Johannesburg region.– If your target users are in Europe, select the EU-Paris region.– If your target users are in Latin America, select the LA-Santiago region.

NO TE

The LA-Santiago region is located in Chile.

● Resource priceResource prices may vary in different regions. For details, see Product PricingDetails.

Selecting an AZWhen deploying resources, consider your applications' requirements on disasterrecovery (DR) and network latency.

● For high DR capability, deploy resources in different AZs within the sameregion.

● For low network latency, deploy resources in the same AZ.

Regions and EndpointsBefore you use an API to call resources, specify its region and endpoint. For moredetails, see Regions and Endpoints.

Elastic Cloud ServerService Overview 12 Region and AZ

2020-07-06 82

https://www.huaweicloud.com/intl/en-us/pricing/index.html

https://www.huaweicloud.com/intl/en-us/pricing/index.html

https://developer.huaweicloud.com/intl/en-us/endpoint

13 ECS and Other Services

Figure 13-1 shows the relationships between ECS and other services.

Figure 13-1 Relationships between ECS and other services

Elastic Cloud ServerService Overview 13 ECS and Other Services

2020-07-06 83

ECS-related Services

Table 13-1 ECS-related services

Service Function Related Operation

Auto Scaling (AS) Automatically adjusts ECSservice resources based onthe configured AS policies.This improves resourceusage and reducesresource costs.

● Using an Existing ECS toCreate an ASConfiguration

● Using a NewSpecifications Templateto Create an ASConfiguration

Load Balancer Automatically distributestraffic to multiple ECSs.This enhances systemservice and fault tolerancecapabilities.

● Backend Server

Elastic VolumeService (EVS)

Enables you to attach EVSdisks to an ECS andexpand their capacity.

● Attaching a Non-SharedDisk

● Attaching a Shared Disk

Virtual PrivateCloud (VPC)

Enables you to configureinternal networks andchange networkconfigurations bycustomizing securitygroups, VPNs, IP addresssegments, and bandwidth.This simplifies networkmanagement. You can alsocustomize the ECS accessrules within a securitygroup and betweensecurity groups to improveECS security.

● Assigning an EIP andBinding It to an ECS

● Adding a Security GroupRule

ImageManagementService (IMS)

Enables you to create ECSsusing images. Thisimproves the efficiency ofECS creation. You can alsouse an existing ECS tocreate a private image andexport the data of the ECSsystem disk or data disks.

● Creating a Data DiskImage Using an ECS DataDisk

● Creating a Full-ECS ImageUsing an ECS


2020-07-06 84

https://support.huaweicloud.com/intl/en-us/usermanual-as/as_02_0102.html







https://support.huaweicloud.com/intl/en-us/usermanual-elb/en-us_topic_0052569729.html

https://support.huaweicloud.com/intl/en-us/qs-evs/evs_01_0036.html



https://support.huaweicloud.com/intl/en-us/usermanual-vpc/en-us_topic_0013748738.html









Service Function Related Operation

DedicatedComputingCluster (DCC)

To physically isolate yourECS, apply for a DCCbefore creating the ECS.After you obtain the DCCand set a region for it,your ECS is automaticallyallocated to the DCC.

● Enabling a DeC● Applying for DCC

Resources

Cloud Eye Allows you to check thestatus of monitored serviceobjects after you haveobtained an ECS. This canbe done without requiringadditional plug-ins beinstalled.

● Basic ECS Metrics● ECS Metrics Under OS

Monitoring (with AgentInstalled)

Data EncryptionWorkshop (DEW)

The encryption featurerelies on DEW. You can usean encrypted image or EVSdisks when creating anECS. In such a case, youare required to use the keyprovided by DEW toimprove data security.

● EVS Disk Encryption● Encrypting Images● Creating a Key Pair

Cloud TraceService (CTS)

Records ECS-relatedoperations for later query,audit, and backtrack.

● Key Operations on ECS

Cloud Backup andRecovery (CBR)

Backs up EVS disks andECSs for restoration.

Purchasing a Server BackupVault

Tag ManagementService (TMS)

Identifies ECSs to facilitateclassification and search.

● Adding Tags● Searching for Resources

by Tag


2020-07-06 85

https://support.huaweicloud.com/intl/en-us/usermanual-dcc/dcc_03_0002.html







https://support.huaweicloud.com/intl/en-us/productdesc-evs/evs_01_0001.html


https://support.huaweicloud.com/intl/en-us/usermanual-dew/dew_01_0034.html

https://support.huaweicloud.com/intl/en-us/usermanual-cts/cts_03_0024.html

https://support.huaweicloud.com/intl/en-us/qs-cbr/cbr_02_0003.html

https://support.huaweicloud.com/intl/en-us/qs-cbr/cbr_02_0003.html




14 Change History

Released On Description

2020-07-06 This issue is the sixteenth official release.● Added C6s ECSs in 5.8 General Computing-plus ECSs.

2020-06-30 This issue is the fifteenth official release.● Added C3ne ECSs in 5.8 General Computing-plus ECSs.● Added M3ne ECSs in 5.10 Memory-optimized ECSs.

2020-06-29 This issue is the fourteenth official release.● Added PI2 ECSs in 5.15 GPU-accelerated ECSs.

2020-03-23 This issue is the thirteenth official release.● Added i3.16xlarge.8 in 5.13 Ultra-high I/O ECSs.● Added 5.6 Kunpeng General Computing-plus ECSs.● Added 10 Billing.● Added G5 ECSs in 5.15 GPU-accelerated ECSs.

2020-01-20 This issue is the twelfth official release.● Added S6 ECSs in 5.7 General Computing ECSs.● Added C6 ECSs in 5.8 General Computing-plus ECSs.● Added M6 ECSs in 5.10 Memory-optimized ECSs.● Added 5.5 x86 ECS Specifications .

2019-11-15 This issue is the eleventh official release.● Changed Intel Xeon Cascade Lake CPUs to second-

generation Intel® Xeon® Scalable processors.● Changed Intel Xeon Skylake CPUs to Intel® Xeon® Scalable

processors.

2019-10-28 This issue is the tenth official release.● Added 5.15 GPU-accelerated ECSs.

Elastic Cloud ServerService Overview 14 Change History

2020-07-06 86

Released On Description

2019-10-24 This issue is the ninth official release.● Added 5.9 General Computing-Basic ECSs.● Added 11 CPU Credits.

2019-10-16 This issue is the eighth official release.● Added 5.13 Ultra-high I/O ECSs.● Moved "Spot Price ECSs" and "Reserved Instances" to User

Guide.

2019-09-06 This issue is the seventh official release.● Added spot price ECSs.● Modified 12 Region and AZ.

2019-07-12 This issue is the sixth official release.● Added 2 ECS Advantages.● Added 3 ECS Application Scenarios.● Added "Why ECS" in 1 What Is ECS?● Modified the ECS architecture in 1 What Is ECS?● Optimized the document structure.● Deleted API statuses in 5.2 ECS Lifecycle.

2019-06-24 This issue is the fifth official release.● Added 4 Notes on Using ECSs.● Modified 12 Region and AZ.

2019-05-07 This issue is the fourth official release.Modified the following content:● Added the newly released c3.6xlarge.2, c3.8xlarge.2, and

c3.15xlarge.2 flavors in 5.8 General Computing-plus ECSs.● Added the newly released d2.2xlarge.8, d2.4xlarge.8,

d2.6xlarge.8, d2.8xlarge.8, and d2.12xlarge.8 flavors in 5.12Disk-intensive ECSs.

2019-04-09 This issue is the third official release.Modified the following content:● Added the newly released m3.15xlarge.8 flavor in 5.10

Memory-optimized ECSs.

2019-03-04 This issue is the second official release.● Added reserved instances.● Modified 13 ECS and Other Services.

2018-11-19 This issue is the first official release.

Elastic Cloud ServerService Overview 14 Change History

2020-07-06 87

Service Overview - Huawei · Auto Scaling Automatic adjustment of computing resources Dynamic...

Documents

Transcript of Service Overview - Huawei · Auto Scaling Automatic adjustment of computing resources Dynamic...