Overview
QoS limits disk I/O and network traffic to reduce resource contention. Actual performance also depends on storage, networks, and workloads. QoS does not guarantee that latency or packet loss is eliminated.
Principle
- Disk bandwidth limits I/O throughput in MB/s, GB/s, or TB/s. Disk IOPS limits reads and writes per second. Network bandwidth limits upstream and downstream traffic in Kbps, Mbps, or Gbps.
- If a QoS limit is not set for a metric, that metric is not rate-limited by QoS.
Considerations
- If you do not set a QoS for an instance/disk offering as an admin (the QoS is not set by default), the QoS you can set for a VM or volume as a normal account/project manager/project admin/ project member is not limited.
- If you set a QoS for an instance/disk offering as an admin, the QoS you can set for a VM or volume as a normal account/project manager/project admin cannot exceed the QoS of the instance/disk offering.
- If you set a QoS for an instance/disk offering and use the offering to create a VM instance or volume, the VM/volume QoS is consistent with the offering QoS by default. The deletion of the offering or modifications on the offering QoS does not influence the QoS of the VM/volume you have created before.
- If you cancel or modify a QoS as an admin, the QoS range that a normal account/project admin/project manager/project member can set is limited by the original QoS value.
- If you change an instance offering for a VM instance, the VM root volume QoS and NIC QoS is limited by the QoS of this new instance offering.
Preparations
Before you begin with this tutorial, make sure that you have installed a ZStack Cloud of the latest version and deployed necessary resources.
For detailed information, see Installation and Deployment in User Guide.
This tutorial described how to add and use QoS in details.
If you want to set QoS for backup jobs, install a Backup Service license in addition to the base license.
ZStack Cloud based on a c72 ISO does not support Backup Service.
If you want to set QoS for V2V migration jobs, install a Migration Service License in addition to the base license.
Scenario Practices
Set QoS for Basic Resources
Basic resources refer to root volumes, data volumes, NICs, virtual IPs, and VPC vRouters. You can set QoS for these basic resources.
Set QoS for Root Volume
This topic describes how to set QoS for a VM root volume and check the setting.
Before you begin this practice, make sure that you have deployed a ZStack Cloud of the latest version.
To set QoS for a VM root volume, follow these steps:
- Create an instance offering.
- Create a VM instance.
- Set QoS for the VM root volume.
- Check the root volume QoS.
Create an Instance Offering
On the main menu of ZStack Cloud, choose Resource Center > Resource Pool > Compute Configuration > Instance Offering. On the Instance Offering page, click Create Instance Offering. Then, the Instance Offering page is displayed. On the displayed page, you can set QoS for the instance offering you are creating through the following parameters:
Set Disk IOPS: Optional. You can set the disk maximum reads/writes per second (IOPS).
If you disable this parameter, the disk IOPS is unlimited. If you need to limit the disk IOPS, enable this parameter and choose either of the following methods:
By Total IOPS: Set a maximum reads & writes IOPS.
By Read/Write IOPS: Set a maximum read IOPS and a maximum write IOPS respectively.
- Read IOPS: Set a maximum read IOPS.
- Write IOPS: Set a maximum write IOPS.
Set Disk Bandwidth: Optional. Set the maximum I/O bandwidth. Unit: MB/s, GB/s, and TB/s.
If you disable this parameter, the disk I/O bandwidth is unlimited. If you need to limit the disk I/O bandwidth, enable this parameter and choose either of the following methods:
By Total Speed: Set a maximum read & write speed.
By Read/Write Speed: Set a maximum read speed and a maximum write speed respectively.
- Read Speed: Set a maximum read speed.
- Write Speed: Set a maximum write speed.
Create a VM Instance
On the main menu of ZStack Cloud, choose Resource Center > Resource Pool > Virtual Resource > VM Instance. On the VM Instance page, click Create VM Instance and choose Standard Creation. Create a VM instance from the instance offering you created before. The VM instance QoS is consistent with the QoS you set for the instance offering.
Modify VM root volume QoS.
On the Configuration Info tab of the VM details page, select the volume whose type is
Rootand click Set Volume QoS.
Check VM Root Volume QoS
Assume that the disk bandwidth is 50 MB/s in this practice. Run fio commands to check the QoS of the VM root volume.
Confirm that
/var/tmpbelongs to the root volume under test rather than a memory filesystem or another volume. Otherwise, choose a directory on the root volume. Use a dedicated test file with at least 4 GiB of free space. This command tests read performance. Screenshot values are examples and vary by environment.Use the VM instance created above. Remove the root volume bandwidth and IOPS limits before recording the unlimited baseline.
Run the following fio command to pressurize the disk I/O for 80 seconds. The example IOPS of the VM root volume is approximately 3198.
fio -filename=/var/tmp/qos-test -direct=1 -iodepth 1 -thread -rw=read -ioengine=psync -bs=16k -size=4G -numjobs=10 -runtime=80 -time_based=1 -group_reporting -name=mytest
Set the root volume read or total bandwidth limit to 5 MB/s and leave IOPS unlimited. Run the same fio command to pressurize the disk I/O for 80 seconds. The example root volume IOPS is approximately 299.
As proved in the step b and c, the QoS you set takes effect and the root volume read bandwidth decreases to approximately 5 MB/s. With a fixed 16 KiB block size, IOPS also decreases. Bandwidth and IOPS are different metrics. You can also see the root volume read bandwidth trend in the Monitoring Data part on the VM details page.
Now, you have set a QoS for a VM root volume successfully. For more action information of VM root volume, see the VM-instance-relevant chapters in User Guide.
Set QoS for NIC
This topic describes how to set a QoS for an NIC and check the setting.
Before you begin the practice, make sure that you have deployed a ZStack Cloud of the latest version.
To set a QoS for an NIC, follow these steps:
- Create an instance offering.
- Create a VM instance.
- Modify the NIC QoS.
- Check the NIC QoS.
Create an instance offering.
On the main menu of ZStack Cloud, choose Resource Center > Resource Pool > Compute Configuration > Instance Offering. On the Instance Offering page, click Create Instance Offering.
Enable Set Network Bandwidth and configure the following parameters as needed. Supported units are Kbps, Mbps, and Gbps. If this option is disabled, the network bandwidth is unlimited.
- Upstream Bandwidth: Optional. Set an upper limit for the network bandwidth used by the VM instance to upload data.
- Downstream Bandwidth: Optional. Set an upper limit for the network bandwidth used by the VM instance to download data.
Create a VM instance.
On the main menu of ZStack Cloud, choose Resource Center > Resource Pool > Virtual Resource > VM Instance. On the VM Instance page, click Create VM Instance and choose Standard Creation. Select the instance offering created in the previous step. The VM instance inherits the network QoS settings of this instance offering.
Modify the NIC QoS.
On the Configuration Info tab of the VM details page, you can modify the VM NIC QoS.
Check the NIC QoS.
Assume that the network bandwidth is 500 Mbps in this practice. Use the VM instance you create here as a sender and another VM instance or host as a receiver. Run iperf3 commands to check the NIC QoS.
Replace RECEIVER_IP with the server IP address. Allow TCP port 5201 between the endpoints, use the same path for both tests, and remove any lower competing QoS limit. Convert units before comparing terminal results in Mbits/sec with monitoring data in MB/s: 1 MB/s equals 8 Mbps.
Use the VM instance created above. Remove the NIC rate limit before recording the unlimited baseline.
Run
iperf3 -son the receiver andiperf3 -c RECEIVER_IP -i 1 -t 1000on the sender to query the bandwidth speed.
Set both NIC bandwidth limits to 50 Mbps and repeat the same command.
Run
iperf3 -c RECEIVER_IP -i 1 -t 1000 -Ron the client to test the reverse direction. Record an unlimited baseline, then apply the same limit and compare the results. By default, the client sends data; with-R, the server sends data. The sender and receiver lines describe the two endpoints of one direction.Comparing the unlimited and limited tests shows that the average bandwidth in each tested direction approaches the configured limit. You can also see the bandwidth trend in the Monitoring Data on the VM details page.
Now, you have set a QoS for an NIC successfully. For more information about NIC, see VM-instance-relevant chapters in User Guide.
Set QoS for Data Volume
This topic describes how to set a QoS for a data volume and check the data volume QoS.
Before you begin the practice, make sure that you have deployed ZStack Cloud of the latest version.
To set a QoS for a data volume, follow these steps:
- Create a disk offering.
- Create a data volume.
- Modify the data volume QoS.
- Check the data volume QoS.
Create a Disk Offering
On the main menu of ZStack Cloud, choose Resource Center > Resource Pool > Compute Configuration > Disk Offering. On the Disk Offering page, click Create Disk Offering. You can set a QoS for the disk offering you are creating through the following parameter:
Set Disk Bandwidth: Optional. Set an upper limit for the disk I/O bandwidth.
If you disable this parameter, the disk I/O bandwidth is unlimited. If you need to limit the disk I/O bandwidth, enable this parameter and choose either of the following method:
By Total Speed: Set a maximum read & write speed.
By Read/Write Speed: Set a maximum read speed and a maximum write speed respectively.
- Read Speed: Optional. Set a maximum read speed.
- Write Speed: Optional. Set a maximum write speed.
Create a data volume.
On the main menu of ZStack Cloud, choose Resource Center > Resource Pool > Virtual Resource > Volume. On the Volume page, click Create Volume. On the Create Volume page, you can create a data volume based on the disk offering you created before and attach it to a VM instance. The data volume QoS is consistent with the QoS you set for the disk offering.
Modify the data volume QoS.
You can modify the data volume QoS through one of the following methods:
On the volume action list, click Set Volume QoS and set a new QoS for the volume.
On the volume action list, click Attach and attach the volume to a VM instance. Then, you can modify the data volume QoS on the Configuration Info tab of the VM details page.
Check the data volume QoS.
Assume that the disk bandwidth is 50 MB/s in this practice. Run fio commands to check the data volume QoS.
Attach and mount the data volume at
/mnt/qos-databefore testing. Confirm that this directory belongs to the data volume under test and replace it with the actual mount path to avoid testing the root volume. Use a dedicated test file with at least 4 GiB of free space. This command tests read performance. Screenshot values are examples and vary by environment.Use the data volume created above. Remove its bandwidth and IOPS limits before recording the unlimited baseline.
Run the following fio command to pressurize the disk I/O. the example data volume IOPS is approximately 3198.
fio -filename=/mnt/qos-data/qos-test -direct=1 -iodepth 1 -thread -rw=read -ioengine=psync -bs=16k -size=4G -numjobs=10 -runtime=80 -time_based=1 -group_reporting -name=mytest
Set the data volume read or total bandwidth limit to 5 MB/s and leave IOPS unlimited. Run the fio command to pressurize the disk I/O for 80 seconds. The data volume example IOPS is approximately 299.
As proved in step b and c, the QoS you set takes effect and the data volume read bandwidth decreases to approximately 5 MB/s. With a fixed 16 KiB block size, IOPS also decreases. Bandwidth and IOPS are different metrics. You can also see the volume read bandwidth trend if you attach this data volume to a VM instance.
Now, you have set a QoS for a data volume successfully. For more information about data volumes, see volume-relevant chapters in User Guide.
Set Individual QoS for Virtual IP
You can set an individual QoS for public/flat network VIPs. In addition, public network VIP can use a shared bandwidth to enjoy a QoS service. This topic describes how to set an individual QoS for a virtual IP (VIP) and check the VIP QoS.
Before you begin the practice, make sure that you have deployed ZStack Cloud of the latest version.
To set a QoS for a VIP, follow these steps:
- Create a VIP.
- Add/delete a VIP QoS.
- Check the VIP QoS.
Create a VIP.
On the main menu of ZStack Cloud, choose Resource Center > Network Service > Basic Network Service > VIP. On the Custom page, click Create VIP. Then, the Create VIP page is displayed. You can set a QoS for the VIP you are creating through following parameters:
You can limit the network bandwidth of public VIPs and flat network VIPs. You can set the QoS for a custom VIP when you create the VIP or after the VIP is created.
Public Network VIP: Supports 3 options: Unlimited, Add QoS, and Bind Shared Bandwidth. Default: Unlimited.
Unlimited: Does not limit the VIP network bandwidth.
Add QoS: Adds a QoS for this VIP individually. Set the following parameters:
Port: Optional. Specify a port to set QoS rule for the port.
- Enter an integer that ranges from 1 to 65535.
- If left blank, the QoS rule takes effect on ports from 1 to 65535.
Upstream Bandwidth: Optional. Set the upstream bandwidth for a VIP.
- Enter an integer. Unit: Kbps, Mbps, and Gbps.
- If left blank, the upstream bandwidth is not limited.
Downstream Bandwidth: Optional. Set the downstream bandwidth for a VIP.
- Enter an integer. Unit: Kbps, Mbps, and Gbps.
- If left blank, the downstream bandwidth is not limited.
Bind Shared Bandwidth: Binds the VIP to a shared bandwidth to make it share bandwidth resources and central speed limit with other VIPs.
Flat Network VIP: Supports 2 options: Unlimited and Add QoS. Default: Unlimited.
Unlimited: Does not limit the VIP network bandwidth.
Add QoS: Adds a QoS for this VIP individually. Set the following parameters:
Port: Optional. Specify a port to set QoS rule for the port.
- Enter an integer that ranges from 1 to 65535.
- If left blank, the QoS rule takes effect on ports from 1 to 65535.
Upstream Bandwidth: Optional. Set the upstream bandwidth for a VIP.
- Enter an integer. Unit: Kbps, Mbps, and Gbps.
- If left blank, the upstream bandwidth is not limited.
Downstream Bandwidth: Optional. Set the downstream bandwidth for a VIP.
- Enter an integer. Unit: Kbps, Mbps, and Gbps.
- If left blank, the downstream bandwidth is not limited.
In this scenario, use the option Add QoS to set an individual QoS for the VIP.
Add/delete a VIP QoS.
You cannot modify an existing VIP QoS, but you can add/delete a QoS for a VIP on the QoS tab of the VIP details page. Both custom and system VIPs support this action.
Check the VIP QoS
Assume that the network bandwidth is 500 Mbps in this practice. Use the VIP to provide an EIP service and attach it to a VM instance. Take this VM instance as a sender and another VM instance/host as a receiver. Run iperf3 commands to check the VIP QoS.
Replace RECEIVER_IP with the server IP address. Allow TCP port 5201 between the endpoints, use the same path for both tests, and remove any lower competing QoS limit. Convert units before comparing terminal results in Mbits/sec with monitoring data in MB/s: 1 MB/s equals 8 Mbps. Test traffic must traverse the EIP bound to this VIP. A port-specific QoS rule must cover the test traffic.
Create a VIP on ZStack Cloud to provide an EIP service. Attach the EIP to a VM instance. The VIP QoS is not limited.
Run
iperf3 -son the VM instance/host as the receiver andiperf3 -c RECEIVER_IP -i 1 -t 1000on the VM instance as the sender to check the forward-test bandwidth.
Set both VIP bandwidth limits to 50 Mbps and repeat the same command.
Run
iperf3 -c RECEIVER_IP -i 1 -t 1000 -Ron the client to test the reverse direction. Record an unlimited baseline, then apply the same limit and compare the results. By default, the client sends data; with-R, the server sends data. The sender and receiver lines describe the two endpoints of one direction.
Now, you have set an individual QoS for a VIP successfully. For more information about VIP, see VIP-relevant chapters in User Guide.
Set QoS for VPC vRouter
This topic describes how to set a QoS for a VPC vRouter and check the VPC vRouter QoS.
VPC vRouter QoS limits the I/O bandwidth of flows passing through the VPC vRouter interface. The upstream bandwidth is flows going into the VPC vRouter. The downstream bandwidth is flows going out the VPC vRouter.
Before you begin this practice, make sure that you have deployed a ZStack Cloud of the latest version.
To set a QoS for a VPC vRouter, follow these steps:
- Create a VPC vRouter.
- Create a VPC network.
- Create a VM instance.
- Modify the VPC vRouter QoS.
- Check the VPC vRouter QoS.
Create a VPC vRouter.
On the main menu of ZStack Cloud, choose Resource Center > vRouter > VPC vRouter. On the VPC vRouter page, click Create VPC vRouter. On the Create VPC vRouter page, set necessary parameters and create a VPC vRouter. For detailed information, see VPC-vRouter-relevant chapters in User Guide.
Create a VPC network.
On the main menu of ZStack Cloud, choose Resource Center > Network Resource > L3 Network Resources > VPC Network. On the VPC Network page, click Create VPC Network. Then, the Create VPC Network page is displayed. Use the VPC vRouter you created before to create a VPC network. For detailed information, see VPC-relevant chapters in User Guide.
Create a VM instance.
On the main menu of ZStack Cloud, choose Resource Center > Resource Pool > Virtual Resource > VM Instance. On the VM Instance page, click Create VM Instance. Then, the Create VM Instance page is displayed. On this page, choose Standard Creation and use the VPC network you created before to create a VM instance.
Modify the VPC vRouter QoS.
On the Network tab of the VPC vRouter details page, you can modify QoS of the associated VPC network and public network.
Check the VPC vRouter QoS.
Assume that the network bandwidth is 500 Mbps in this practice. Take the VM instance you create in this practice as the sender and another VM instance on the public network as a receiver. Run iperf3 commands to check the VPC vRouter QoS.
Replace RECEIVER_IP with the server IP address. Allow TCP port 5201 between the endpoints, use the same path for both tests, and remove any lower competing QoS limit. Convert units before comparing terminal results in Mbits/sec with monitoring data in MB/s: 1 MB/s equals 8 Mbps. Test traffic must traverse the VPC vRouter interface under test. Upstream and downstream are defined relative to that interface.
Create a VM instance based on the public network on ZStack Cloud. The QoS of the instance offering the VM based on is not limited. Make sure that the sender and receiver can intercommunicate with each other.
Run
iperf3 -son the receiver andiperf3 -c RECEIVER_IP -i 1 -t 1000on the sender to check the bandwidth speed.
Set both bandwidth limits of the tested VPC vRouter interface to 50 Mbps and repeat the same command.
Run
iperf3 -c RECEIVER_IP -i 1 -t 1000 -Ron the client to test the reverse direction. Record an unlimited baseline, then apply the same limit and compare the results. By default, the client sends data; with-R, the server sends data. The sender and receiver lines describe the two endpoints of one direction.As proved in step b and c, the VPC vRouter QoS takes effect and the average bandwidth in each tested direction approaches the configured limit.
Now, you have set a QoS for a VPC vRouter successfully.
Set QoS for Backup Job
This topic describes how to set a QoS for a backup job and check the backup job QoS.
Before you begin this practice, installed the latest h76c ZStack Cloud and deployed necessary resources.
You need a Backup Service License in addition to a base license.
c72 ZStack Cloud does not support Backup Service.
To set a QoS for a backup job, follow these steps:
- Create a backup job.
- Modify the backup job QoS.
- Check the backup job QoS.
Create a backup job.
On the main menu of ZStack Cloud, choose Platform O&M > Backup Management > Backup Job. On the Backup Job page, click VM/Volume > Create Backup Job. Then, the Create Backup Job page is displayed. On this page, you can set QoS for the backup job you are creating through following parameters:
Set QoS: Optional. You can set network QoS and disk QoS for the backup job.
- Downstream Bandwidth: Set the maximum downstream bandwidth. Unit: Kbps, Mbps, and Gbps. Valid values: 8 Kbps to 30 Gbps. If you do not specify the downstream bandwidth, this bandwidth is unlimited.
- Upstream Bandwidth: Set the maximum upstream bandwidth. Unit: Kbps, Mbps, and Gbps. Valid values: 8 Kbps to 30 Gbps. If you do not specify the upstream bandwidth, this bandwidth is unlimited.
- Disk Read Speed: Set the maximum read speed. Unit: MB/s and GB/s. Valid values: 1MB/s to 100GB/s. If you do not specify the disk read speed, this speed is unlimited.
- Disk Write Speed: Set the maximum write speed. Unit: MB/s and GB/s. Valid values: 1MB/s to 100GB/s. If you do not specify the disk write speed, this speed is unlimited.
We recommend that you set a QoS based on your physical network environments and the concurrent number of implemented backup jobs.
Modify the backup job QoS
On the details page of the backup job, you can click Edit QoS to modify the downstream/upstream bandwidth and disk read/write speed. The modification takes effect when the job automatically creates a backup for the next time.
Check the backup job QoS
Assume that the network bandwidth is 500 Mbps in this practice. Prepare two same volumes, volume1 and volume2, to create backup jobs. Run iftop commands to check QoS of the backup jobs.
Create backup job1 based on volume1. The QoS of backup job1 is unlimited. Use iftop to monitor the network bandwidth speed.
Create backup job 2 based on volume 2. The QoS of backup job 2 is 40 Mbps. Use iftop to monitor the network bandwidth speed.
As proved in step a and b, the QoS you set for backup job2 takes effect and the backup job bandwidth reduces to a threshold level. Choose Backup Management > Backup Service > Local Backup Data > Volume. Click the volume 1 and 2 respectively and enter corresponding backup details pages. On details pages, you can see the time consumptions of these 2 backup jobs.
Now, you have set a QoS for a backup job. For more information, see Backup Service Tutorial.
Set QoS for V2V Conversion Host
This topic describes how to set a QoS for a V2V conversion host and check the setting.
Before you begin this practice, make sure that you have deployed a latest ZStack Cloud.
Take over KVM to ZStack Cloud.
To use Migration Service, install a Migration Service license in addition to the base license.
To set a QoS for a V2V conversion host, follow these steps:
- Add a V2V conversion host.
- Modify the QoS of the V2V conversion host.
- Check the QoS of the V2V conversion host.
Add a V2V conversion host.
On the main menu of ZStack Cloud, choose Platform O&M > Migration Service > V2V Conversion Host. On the V2V Conversion Host page, click Add V2V Conversion Host. Then, the Add V2V Conversion Host page is displayed. On this page, you can set a QoS for the V2V conversion host you are adding through following parameters:
- Upstream Bandwidth: Optional. Set a maximum upstream bandwidth. This parameter limits the speed at which the V2V conversion host uploads data to the target primary storage. Units: Mbps and Gbps. Valid values: 1 Mbps to 32 Gbps.
- Downstream Bandwidth: Optional. Set a maximum downstream bandwidth. This parameter limits the speed at which the V2V conversion host downloads data from the source primary storage. Units: Mbps and Gbps. Valid values: 1 Mbps to 32 Gbps.
Modify the QoS of the V2V migration host.
You can modify the QoS of the V2V migration host on its details page.
Check the QoS of the V2V conversion host.
Assume that the network bandwidth is 500 Mbps in this practice. Use iftop commands to check the QoS of the V2V conversion host.
Add a V2V conversion host to ZStack Cloud. The QoS of the V2V conversion host is unlimited.
Create a V2V job and set the QoS of the V2V conversion host to 100 Mbps. Run
iftop -i ifb0on the V2V conversion host to query the bandwidth speed.
As proved in the step a and b, the QoS you set takes effect and the network bandwidth reduces to the threshold level. You can see the time consumed by V2V jobs on the V2V Migration page.
Monitor the interface carrying the migration traffic while the job runs. The interface
ifb0is used in this example; replace it with the corresponding interface in your environment.
Now, you have set a QoS for a V2V jobs successfully. For more information, see V2V Migration Tutorial.
Glossary
Instance: An instance is a virtual machine or server that runs the images of operating systems in Cloud, such as VM instance and elastic baremetal instance.
VM Instance: A VM instance is a virtual machine instance running on a host. A VM instance has its own IP address and can access public networks and run application services.
Volume: A volume provides storage space for a VM instance. Volumes are categorized into root volumes and data volumes.
Root Volume: A root volume provides support for the system operations of a VM instance.
Data Volume: A data volume provides extended storage space for a VM instance.
Image: An image is a template file used to create a VM instance or volume. Images are categorized into system images and volume images.
Instance Offering: An instance offering defines the number of vCPU cores, memory size, network bandwidth, and other configuration settings of VM instances.
Disk Offering: A disk offering defines the capacity and other configuration settings of volumes.
GPU Specification: A GPU specification defines the frame per second (FPS), video memory, resolution, and other configuration settings of a physical or virtual GPU. GPU specifications are categorized into physical GPU specifications and virtual GPU specifications.
vNUMA Configuration: vNUMA uses CPU pinning to passthrough the topology of associated host physical NUMA (pNUMA) nodes to a VM instance, generating a topology of virtual NUMA (vNUMA) nodes for the VM instance. This topology enables a vCPU on a vNUMA node to primarily access the local memory and thus improves VM performance.
NUMA (Non-Uniform Memory Access): Non-uniform memory access (NUMA) is a computer memory design where the memory access time depends on the memory location relative to the CPU. Under NUMA, a processor can access its own local memory faster than non-local memory and thus improves VM performance.
pNUMA Node (physical NUMA Node): A pNUMA node (physical NUMA node) is a host NUMA node predefined based on the host NUMA architecture. It is used to manage the CPUs and memory of the host.
pNUMA Topology (physical NUMA Topology): A pNUMA topology (physical NUMA topology) is the topology of the host NUMA nodes predefined by the CPU vendor based on the host NUMA architecture.
vNUMA Node (virtual NUMA Node): A vNUMA node (virtual NUMA node) is generated by passing-through associated pNUMA nodes via CPU pinning. It is used to manage the CPUs and memory of a VM instance.
vNUMA Topology (virtual NUMA Topology): A vNUMA topology (virtual NUMA topology) is the topology of VM NUMA nodes generated by passing-through associated pNUMA nodes via CPU pinning.
Local Memory: Local memory is the memory that a CPU (pCPU or vCPU) accesses through the Uncore iMC (Integrated Memory Controller) of the same NUMA (pNUMA or vNUMA) node. Compared with accessing non-local memory, accessing local memory has lower latencies.
CPU Pinning: CPU pinning assigns the virtual CPUs (vCPUs) of a VM instance to specific physical CPUs (pCPUs) of the host, which improves VM performance.
EmulatorPin Configuration: EmulatorPin assigns all other threads than virtual CPU (vCPU) threads and IO threads of a VM instance to physical CPUs (pCPUs) of the host so that these threads run on assigned pCPUs.
Auto-Scaling Group: An auto-scaling group is a group of VM instances that are used for the same scenarios. An auto-scaling group can automatically scale out or in based on application workloads or health status of VM instances in the group.
Snapshot: A snapshot is a point-in-time capture of data status in a volume.
Affinity Group: A VM scheduling policy is a resource orchestration policy based on which VM instances are assigned hosts to achieve the high performance and high availability of businesses.
Zone: A zone is a logical group of resources such as clusters, L2 networks, and primary storage. Zone is the largest resource scope defined in the Cloud.
Cluster: A cluster is a logical group of hosts (compute nodes).
Host: A host provides compute, network, and storage resources for VM instances.
Primary Storage: A primary storage is one or more servers that store volume files of VM instances. These files include root volume snapshots, data volume snapshots, image caches, root volumes, and data volumes.
Image Storage: An image storage is a storage server that stores VM image templates, including ISO image files.
iSCSI Storage: iSCSI storage is an SAN storage that uses the iSCSI protocol for data transmission. You can add an iSCSI SAN block as a Shared Block primary storage or pass through the block to a VM instance.
FC Storage: FC storage is an SAN storage that uses the FC technology for data transmission. You can add an FC SAN block as a Shared Block primary storage or pass through the block to a VM instance.
NVMe Storage: A type of storage implemented via the NVMe-oF (NVMe over fabrics) protocol. You can add a block device configured from an NVMe storage as SharedBlock primary storage.
L2 Network: An L2 network is a layer 2 broadcast domain used for layer 2 isolation. Generally, L2 networks are identified by names of devices on the physical network.
VXLAN Pool: A VXLAN pool is a collection of VXLAN networks established based on VXLAN Tunnel Endpoints (VTEPs). The VNI of each VXLAN network in a VXLAN pool must be unique.
L3 Network: An L3 network includes IP ranges, gateway, DNS, and other network configurations that are used by VM instances.
Public Network: Generally, a public network is a logical network that is connected to the Internet. However, in an environment that has no access to the Internet, you can also create a public network.
Flat Network: A flat network is connected to the network where the host is located and has direct access to the Internet. VM instances in a flat network can access public networks by using elastic IP addresses.
VPC Network: A VPC network is a private network where VM instances can be created. A VM instance in a VPC network can access the Internet through a VPC vRouter.
Management Network: A management network is used to manage physical resources in the Cloud. For example, you can create a management network to manage access to hosts, primary storage, image storage, and VPC vRouters.
Flow Network: A flow network is a dedicated network for port mirror transmission. You can use a flow network to transmit the mirrors of data packets of NIC ports to the target ports.
VPC vRouter: A VPC vRouter is a dedicated VM instance that provides multiple network services.
VPC vRouter HA Group: A VPC vRouter HA group consists of two VPC vRouters. Either VPC vRouter can be a primary or secondary VPC vRouter for the group. If the primary VPC vRouter does not work as expected, the VPC vRouter becomes the secondary VPC vRouter in the group to ensure high availability of business.
vRouter Image: A vRouter image encapsulates network services and can be used to create VPC vRouters.
Dedicated-Performance LB Image: A dedicated-performance load balancer (LB) image encapsulates dedicated-performance load-balancing services and can be used to create load balancer instances. However, a dedicated-performance load balancer image cannot be used to create VM instances.
vRouter Offering: A vRouter offering defines the number of vCPU cores, memory size, image, management network, and public network configuration settings of VPC vRouters. You can use a vRouter offering to create VPC vRouters that can provide network services for public networks and VPC networks.
LB Instance Offering: A load balancer (LB) instance offering defines the CPU, memory, image, and management network configuration settings used to create LB instances. LB instances provide load balancing services for the public network, flat network, and VPC network.
SDN Controller: The SDN controller is the core of the SDN architecture, responsible for centralized management and control of network devices.
SDN Cluster: A cluster of dedicated VM instances designed to provide highly available SDN capabilities.
SDN Instance: A dedicated VM instance designed to provide SDN network capabilities.
SDN Image: An SDN image encapsulates an SDN software and can be used to create SDN instances.
SDN Instance Offering: An SDN instance offering defines the CPU, memory, SDN image, and management network configuration used for creating SDN instances.
Security Group: A security group provides security control services for VM NICs. It filters the ingress or egress TCP, UDP, and ICMP packets of VM NICs based on the specified security rules.
VIP: In bridged network environments, a virtual IP address (VIP) provides network services such as serving as an elastic IP address (EIP), port forwarding, load balancing, IPsec tunneling. When a VIP provides the preceding network services, packets are sent to the VIP and then routed to the destination network where VM instances are located.
EIP: An elastic IP address (EIP) functions based on the NAT technology. IP addresses in a private network are translated into an EIP that is in another network. This way, private networks can be accessed from other networks by using EIPs.
Port Forwarding: Port forwarding functions based on the layer-3 forwarding service of VPC vRouters. This service forwards traffic flows of the specified IP addresses and ports in a public network to specified ports of VM instances by using the specified protocol. If your public IP addresses are insufficient, you can configure port forwarding for multiple VM instances by using one public IP address and port.
Load Balancer: A load balancer distributes traffic flows of a virtual IP address to backend servers. It automatically inspects the availability of backend servers and isolates unavailable servers during traffic distribution. This way, the load balancer improves the availability and service capability of your business.
Listener: A listener monitors the frontend requests of a load balancer and distributes the requests to a backend server based on the specified policy. In addition, the listener performs health checks on backend servers.
Forwarding Rule: A forwarding rule forwards the requests from different domain names or URLs to different backend server groups.
Backend Server Group: A backend server group is a group of backend servers that handles requests distributed by load balancers. It is the basic unit for traffic distribution by load balancer instances.
Backend Server: A backend server handles requests distributed by a load balancer. You can add a VM instance on the Cloud or a server on a third-party cloud as a backend server.
Frontend Network: A frontend network is a type of network that is associated with a load balancer. Requests from the network are distributed by the load balancer to backend servers based on a specified policy.
Backend Network: A backend network is a type of network that is associated with a load balancer. Requests from frontend networks are distributed by the load balancer to servers in the backend network.
Load Balancer Instance: A load balancer instance is a custom VM instance used to provide load balancing services.
Certificate: If you select HTTPS for a listener, associate it with a certificate to make the listener take effect. You can upload either a certificate or certificate chain.
Firewall: A firewall is an access control policy that monitors ingress and egress traffic of VPC vRouters and decides whether to allow or block specific traffic based on the associated rule sets and rules.
Firewall Rule Set: A firewall rule set is a set of rules that a firewall uses to defend against network attacks. You need to associate a rule set with the egress or ingress flow direction of VPC vRouter NICs to make the rule set take effect.
Firewall Rule: A firewall rule is an access control entry associated with the egress or ingress flow direction of VPC vRouter NICs to defend against network attacks. A firewall rule includes rule priority, match condition, and behavior.
Rule Template: A rule template is a template that you can select when you add rules to a rule set or a firewall.
IP/Port Set: An IP or port set is a set of IP addresses or ports that you can select when you add rules to a rule set or a firewall.
IPsec Tunnel: An IPSec tunnel encrypts and verifies IP packets that transmit over a virtual private network (VPN) from one site to another.
OSPF Area: An Open Shortest Path First (OSPF) area is divided from an autonomous system based on the OSPF protocol. This simplifies the hierarchical management of vRouters.
NetFlow: A NetFlow monitors the ingress and egress traffic of the NICs of VPC vRouters. The supported versions of data flows are V5 and V9.
Port Mirroring: Port mirroring mirrors the traffic data of VM NICs and sends the traffic data to the target ports. This allows for the analysis of data packets of ports and simplifies the monitoring and management of data traffic and makes it easier to locate network errors and exceptions.
Route Table: A route table contains information about various routes that you configure. Route entries in a route table must include the destination network, next hop, and route priority.
CloudFormation: CloudFormation is a service that simplifies the management of cloud resources and automates deployment and O&S. You can create a stack template to configure cloud resources and their dependencies. This way, resources can be automatically configured and deployed in batches. CloudFormation provides easy management of the lifecycle of cloud resources and integrates automatic O&S into API and SDK.
Resource Stack: A resource stack is a stack of resources that are configured by using a stack template. The resources in the stack have dependencies with each other. You can manage resources in the stack by managing the resource stack.
Stack Template: A stack template is a UTF8-encoded file based on which you can create resource stacks. The stack template defines the resources that you want, the dependencies between the resources, and the configuration settings of the resources. When you use a stack template to create a resource stack, CloudFormation parses the template and the resources are automatically created and configured.
Sample Template: A sample template is a commonly used resource stack. You can use a sample template provide by the Cloud to create resource stacks.
Designer: A designer is a CloudFormation tool that allows you to orchestrate cloud resources. You can drag and drop resources on a canvas and use lines to establish dependencies between the resources.
Baremetal Cluster: A baremetal cluster consists of baremetal chassis. You can manage baremetal chassis by managing a baremetal cluster where the chassis reside.
Deployment Server: A deployment server is a server that provides PXE service and console proxy service for baremetal chassis.
Baremetal Chassis: A baremetal chassis is used to create a baremetal instance and is identified based on the BMC interface and IPMI configuration setting.
Preconfigured Template: A preconfigured template is used to create a preconfigured file that allows for unattended batch installation of an operating system for baremetal instances.
Baremetal Instance: A baremetal instance is an instantiated baremetal chassis.
Elastic Baremetal Management: Elastic Baremetal Management provides dedicated physical servers for your applications to ensure high performance and stability. In addition, this feature allows elastic scaling. You can apply for and scale resources based on your needs.
Provision Network: A provision network is a dedicated network for PXE boot and image downloads while creating elastic baremetal instances in a gateway proxy cluster.
Elastic Baremetal Cluster: Provides a separated cluster to manage baremetal nodes.
Gateway Node: A gateway node is a node where the ingress and egress traffic of the Cloud and elastic baremetal instances in gateway proxy clusters is forwarded.
Baremetal Node: A baremetal node is used to create a baremetal instance and is identified based on the BMC interface and IPMI configuration setting.
Elastic Baremetal Instance: An elastic baremetal instance has the same performance as physical servers and allows elastic scaling. You can apply for and scale resources based on your needs.
Elastic Baremetal Offering: An elastic baremetal offering defines the number of vCPU cores, memory size, CPU architecture, CPU model, and other configuration settings of elastic baremetal instances.
vCenter: The Cloud allows you to take over vCenter and manage resources on the vCenter.
VM Instance: A VM instance is an ESXi virtual machine instance running on a host. A VM instance has its own IP address to access public networks and can run application services.
Network: A vCenter network defines the network settings of VM instances on vCenter, such as IP range, gateway, DNS, and network services.
Volume: A volume provides storage space for a VM instance on vCenter. A volume attached to a VM instance can be used as a root volume or data volume. A root volume provides support for the system operations of a VM instance. A data volume provides extended storage space for a VM instance.
Image: An image is a template file used to create a VM instance or volume on vCenter. Images are categorized into system images and volume images.
Event Message: Event Message displays event alarm messages of vCenter that is took over by the Cloud. This feature allows you to locate errors and exceptions efficiently.
Network Topology: A network topology visualizes the network architecture of the Cloud. It allows for efficient planning, management, and improvement of network architecture. Network topologies can be categorized into global topologies and custom topologies.
Performance Analysis: Performance Analysis displays the performance metrics of key resources monitored externally or internally in the Cloud. You can view the performance analysis or export the analysis report as needed to improve the O&M efficiency.
Capacity Management: Capacity Management visualizes the capacities and usages of key resources in the Cloud. You can use this feature to improve O&S efficiency.
MN Monitoring: Management Node (MN) monitoring allows you to view the health status of each management node when you use multiple management nodes to achieve high availability.
Alarm: An alarm is used to monitor the status of time-series data and events and respond to the status change. Alarms can be categorized into resource alarm, event alarm, and extended alarm.
One-Click Alarm: A one-click alarm integrates multiple metrics of a resource. You can create one-click alarms for multiple resources to monitor these resources.
Alarm Template: An alarm template is a template of alarm rules. If you associate an alarm template with a resource group, an alarm is created to monitor the resources in the group.
Resource Group: A resource group consists of resources grouped based on your business needs. If you associate an alarm template with a resource group, the alarm rules specified by the template take effect on all the resources in the group.
Message Template: A message template specifies the text template of a resource alarm message or event alarm message sent to an SNS system.
Message Source: A message source is used to take over extended alarm messages. If you configure alarms for message sources, extended alarm messages can be sent to various endpoints.
Endpoint: An endpoint is a method that users obtain subscribed messages. Endpoints are categorized into system endpoints, email, DingTalk, HTTP application, short message service, and Microsoft Teams.
Alarm Message: An alarm message is a message sent the time when an alarm is triggered.
Current Task: A current task is an ongoing operation performed in the Cloud. You can perform centralized management over ongoing operations.
Operation Log: An operation log is a chronological record of operations on the specified objects and their operation results.
Audit: Audit monitors and records all activities on the Cloud. You can use this feature to implement operation tracking, cybersecurity classified protection compliance, security analysis, troubleshooting, and automatic O&M.
Log Collection: Allows you to collect with one click the log data from the Cloud and various nodes on the Cloud generated in the specified time period and download the log data.
One-Click Inspection: Comprehensively inspects the health status of key resources and services of the Cloud and scores their healthiness based on the inspection results. In addition, the one-click inspection service provides O&M suggestions and inspection reports.
Backup Management: Backup management integrates multiple disaster recovery technologies such as incremental backup and full backup that are suitable for multiple business scenarios. You can implement local backup and remote backup based on your business needs.
Backup Job: You can create a backup job to back up local VM instances, volumes, or databases to a specified storage server on a regular basis.
Local Backup Data: Local backup data of VM instances, volumes, and databases is stored in the local backup server.
Local Backup Server: A local backup server is located at the local data center and is used to store local backup data.
Remote Backup Server: A remote backup server is located at a remote data center or a public cloud and is used to store remote backup data.
Continuous Data Protection (CDP): Continuous Data Protection (CDP) provides second-level and fine-grained continuous backups for important business systems in VM instances, allowing users to restore VM data to a specific time state, and retrieve files without restoring the system.
CDP Task: You can create a CDP task to continuously back up your VM data to a specified backup server to achieve continuous data protection and recovery.
CDP Data: The backup data generated from continuous data protection on VM instances is stored in local backup servers.
Recovery Point: A recovery point is a data point generated during continuous data protection. A recovery point corresponds to a data record within the recovery point interval specified by the user.
Locked Recovery Point: You can lock or unlock a recovery point as needed. After a recovery point is locked, data of the recovery point will not be automatically cleared or deleted.
Recovery Task: A recovery task helps you quickly restore data by specifying a CDP task and recovery point, and allows you to view the recovery progress and logs in a more friendly way.
Cryptography Security Compliance: The Cryptography Security Compliance service provides applications with cloud security capabilities based on commercial cryptography, meeting the requirements of commercial cryptography application security assessments.
HSM Pool: An HSM pool is a logical group of hardware security modules (HSMs) and is used to provide unified cryptography services such as signature validation and encryption.
HSM: A hardware security module (HSM) is a dedicated device that encrypts, decrypts, and authenticates information by using the cryptographic technology.
Platform Cryptography Security Compliance: Enables the Cloud to meet the requirements of Cryptography Security Compliance through the cryptography capabilities provided by HSM pools.
Certificate Login: Authenticates the identity of a user by using a UKey device.
Data Protection: Protects important data on the Cloud to ensure the data confidentiality and integrity.
Scheduled Job: A scheduled job defines that a specific action be implemented at a specified time based on a scheduler.
Scheduler: A scheduler is used to schedule jobs. It is suitable for business scenarios that last for a long time.
Tag: A tag is used to mark resources. You can use a tag to search for and aggregate resources.
Migration Service: The Cloud provides V2V migration service that allows you to migrate VM instances and data from other virtualized platform to the current cloud platform.
ZMigrate Migration Service: A migration service installed from Application Market that migrates VM instances and their data from VMware environments to the current cloud platform.
V2V Migration: V2V Migration allows you to migrate VM instances from the VMware or KVM platform to the current cloud platform.
V2V Conversion Host: A V2V conversion host is a host in the destination cluster that you need to specify during V2V migration to cache VM instances and data when you implement V2V migration. After the VM instances and data are cached in the V2Vconversion host, they are migrated to the destination primary storage.
User: A user is a natural person that constructs the most basic unit in Tenant Management.
User Group: A user group is a collection of natural persons or a collection of project members. You can use a user group to grant permissions.
Role: A role is a collection of permissions that can be granted to users. A user that assumes a role can call API operations based on the permissions specified by the role. Roles are categorized into platform roles and project roles.
Single Sign-On: The Single Sign-On service provided by the Cloud. It supports seamless access to SSO systems. Through the service, related users can directly log in to the Cloud and manage cloud resources.
Project: A project is a task that needs to be accomplished by specific personnel at a specified time. In Tenant Management, you can plan resources at the project granularity and allocate an independent resource pool to a project. The word Tenant in Tenant Management mainly refers to projects. A project is a tenant.
Project Member: A project member is a member in a project who is granted permissions on specific project resources and can use the resources to accomplish tasks. Project members include the project admin, project managers, and normal project members.
Process Management: Process management is part of ticket management that manages the processes related to the resources of projects. Processes can be categorized into default processes and custom processes.
My Approvals: In the Cloud, only the administrator and project administrators are granted approval permissions. the administrator and project administrators can approve or reject a ticket. If a ticket is approved, resources are automatically deployed and allocated to the specified project.
Bills: A bill is the expense of resources totaled at a specified time period. Billing is accurate to the second. Bills can be categorized into project bills, department bills, and account bills.
Pricing List: A pricing list is a list of unit prices of different resources. The unit price of a resource is set based on the specification and usage time of the resource.
Console Proxy: Console proxy allows you to log in to a VM instance by using the IP address of a proxy.
AccessKey Management: An AccessKey pair is a security credential that one party authorizes another party to call API operations and access its resources in the Cloud. AccessKey pairs shall be kept confidential.
IP Allowlist/Blocklist: An IP allowlist or blocklist identifies and filters IP addresses that access the Cloud. You can create an IP allowlist or blocklist to improve access control of the Cloud.
Application Center: Application Market allows you to add applications to the Cloud and then access the applications with one click. It extends the functionality of the Cloud. You can add default applications through the built-in installation package or add more applications through URLs.
Sub-Account Management: A sub-account can be created by the admin or synced from an SSO authentication system and is managed by the admin. Resources created under a sub-account are managed by the sub-account.
Theme and Appearance: You can customize the theme and appearance of the Cloud.
Email Server: If you select Email as the endpoint of an alarm, you need to set an email server. Then alarm messages are sent to the email server.
Log Server: A log server is used to collect management node logs or the platform operation logs. You can add a log server to the cloud and use the collected logs for operation trace or troubleshooting. This makes your O&M more efficient.
Global Setting: Global Setting allows you to configure settings that take effect on the whole platform.
Scenario Template: Scenario Template provides multiple templates that encapsulate scenario-based global settings. You can apply a template globally with one click based on your business needs. This improves your O&M efficiency.
HA Policy: HA Policy is a mechanism that ensures sustained and stable running of the business if VM instances are unexpectedly stopped or are errored because of errors occurring to compute, network, or storage resources associated with the VM instances. By enabling this feature, you can customize VM HA policies to ensure your business continuity and stability.
Time Management: Manages the Cloud system time and allows you to configure time servers for the Cloud. After you configure NTP time servers for the Cloud, the clock of the time servers is synced with all nodes of the Cloud.
GPU Device: A GPU device is a powerful microprocessor with high computational capabilities. You can use a GPU device to handle intricate graphics rendering and parallel computing jobs, thus improving the efficiency of businesses such as graphic production, video processing, and machine learning.
Script Library: The script library stores and manages script files centrally. By executing scripts on VM instances, you can complete complex O&M operations and automated jobs.
XML Hook: An XML Hook is a script that can flexibly insert or modify parameters in XML files of VM instances. By attaching an XML Hook to a VM instance, you can customize VM configurations and enable specialized functionalities.
Container Service: A simple and user-friendly container management service, providing features like GPU management & scheduling, multi-tenancy, multi-cluster, quota configuration, CI/CD. and microservice. The service reduces the container using complexity and aligns well with traditional user’s habits, helping you easily manage and deploy your container cluster, and enjoy the benefits of cloud-native technologies in a quick and convenient way.
Advanced Monitoring Server: An advanced monitoring server is a dedicated VM instance used to receive advanced monitoring data of load balancers and other resources.
Advanced Monitoring Server Image: An advanced monitoring server image encapsulates the advanced monitoring service and can be used to create advanced monitoring server.
Advanced Monitoring Server Offering: An advanced monitoring server offering defines the CPU cores, memory size, image, management network, and public network configurations of advanced monitoring server. You can use an advanced monitoring server offering to create advanced monitoring servers.
Plugin Management: You can package extended resources or tools into standardized plugins for quick installation and integration, expanding the Cloud capabilities.
Region Management: A region is a self-contained cloud environment with independent management node(s), networks, hardware, and cloud resources. ZStack IAM enabled user synchronization and SSO across multiple regions.
