Instance vouchers are another payment method for computing instances. After prepaying the vouchers monthly/yearly at a price close to the price for yearly/monthly package pay-in-advance billing, the vouchers can be used for fee deduction for computing instances (virtual machine instance, container instance, and POD instance) billed by configuration.
Instance vouchers decouple the payment and resource creation, which determines the characteristics of instances that can be deducted, not limited to a specific instance or certain instances. Therefore, it can meet not only the economic considerations when long-term resource use is scheduled, but also the needs of maintaining flexibility in resource creation and deletion.
From the perspective of applicable instance types, they can be divided into Cloud virtual machine instance voucher, Container instance voucher and POD instance voucher:
From the perspective of whether the resources are reserved, can be divided into ** no resource reservation type** and ** resource reservation type** two categories (reserved type for the public test phase, if you need to buy, please contact customer service)
No resource reservation type | Resource reservation type | |
---|---|---|
Purchase method | Specify geographic region, buy by specification family and computing power 1 | Specify geographic region and available area, buy by specification and quantity |
How to use | Configurable per-configuration billing instances with matching 2 in any available zone in the same locale | Configurable per-configuration billing instances with matching specifications in the same locale in the same available zone can be used |
Resource reservation | Without reservation, if resources are in short supply, there may be situations where instances with sold-out specifications cannot be created, resulting in a waste of coupons. | With reservations, resources are reserved according to the specifications and quantities purchased to ensure that no matter how the resources are sold, at least the coupons will be sold. Instances of the specifications and quantities purchased with coupons can be created |
Instance vouchers are a new way to purchase resources used in conjunction with per-configuration billing instances. Several purchase options are compared as follows:
Notification:
- If you are currently deploying a business based on K8S or using elastic scaling services, and the instances in the cluster need to be frequently deleted-created for resizing due to policies, it is recommended that you evaluate the uptime of the virtual machines in the cluster and consider using instance vouchers to get more price discounts;
- If you are currently using annual monthly instances and have no need to remove the resizing, it is not recommended that you resize to the instance voucher + per-configuration billing instance model. Because the instance voucher does not currently support resource reservation, in case of resource constraint, the instance may not be created successfully again after deletion due to insufficient resources.
Note:
* Generic Shared, Compute Optimized Shared, Bare Metal and GPU Virtualized specifications do not support the purchase of instance vouchers at this time;
* Since the specifications of instances are sold in different regions, there are also regional differences in the specification families supported by the instances vouchers, please refer to the purchase page for the specific specification families available in each region;
- If you want to reserve a certain amount of resources by purchasing these vouchers, you can either purchase the instances first or purchase the vouchers first;
- Please make sure to check the specifications of the built instances and pay attention to the sold-out status of the planned instances before purchasing the vouchers.
The resource reserved instance vouchers can be purchased after specifying the region, availability zone, specifications, and quantity. The price of a single reserved instance is the same as the price of the purchased instance, and most specifications in the host/native container are available for sale.
No resource reservation type instance voucher, after specifying the locale, applicable product and specification family, is purchased by computing power 1. The specification families currently supported for sale and the specifications that do not support the use of instance voucher deduction are listed below:
On-Sale Specification FamilySpecifications that do not support creditingMainframe available for purchaseContainers/PODs available for purchasePurchase unit conversions | |||||
---|---|---|---|---|---|
General Standards | g.n3 | -- | √ | × | 1 computing power = 1 core |
g.n2 | -- | √ | √ | ||
g.n1 | g.n1.xlarge_m | √ | √ | ||
Computational Optimization - Standard | -- | √ | × | ||
c.n2 | -- | √ | √ | ||
c.n1 | c.n1.medium,c.n1.xlarge_m,c.n1.2xlarge_s, c.n1.2xlarge_m,c.n1.4xlarge_m |
√ | √ | ||
Computational optimization-intensive | -- | √ | × | ||
Memory Optimization - Standard | -- | √ | × | ||
m.n2 | -- | √ | √ | ||
m.n1 | m.n1.medium | √ | √ | ||
High Frequency Computing - General | -- | √ | √ | ||
h.g1 | -- | √ | × | ||
Storage Optimization - IO | s.i3.22xlarge | √ | × | ||
s.i1 | -- | √ | × | ||
Storage Optimization - Big Data | s.d2 | s.i3.22xlarge | √ | × | |
GPU-Standard | p.n1p40 | -- | √ | × | 1 computing power = 1 GPU card |
p.n1p40h | -- | √ | × | ||
p.n1v100 | -- | √ | × |
Non-GPU specs (1 compute power = 1 core) | GPU Specifications (1 computing power = 1 card) | |||
---|---|---|---|---|
Reserved type | No set-aside type | No set-aside type | ||
Virtual Machines | 500 computing power/geography | 500 computing power/geography | 100 computing power/territory | 100 computing power/territory |
Native Container | 500 computing power/geography | 500 computing power/geography | (No such specifications available at this time) | |
POD | 500 computing power/geography | 500 computing power/geography |
To upgrade your quota please submit a work order.
Notes
- Computational power: A metric that reflects the computing power of a single-instance or multi-instance cluster within the same specification family. Compute power = ∑ spec size x number (spec size is a measure of spec compute power within the same spec family, CPU spec is usually equal to the number of cores, GPU spec is usually equal to the number of GPU cards), e.g. if you have 2 instances of 2C8G and 2 instances of 4C16G under your generic standard second generation spec family g.n2, the compute power of 4 instances under that spec family is 2 x 2 + 4 x 2 = 12; & lt;br>
- There are individual specifications in the no-resource-reserved instance voucher that are not covered by the credit, as detailed above: Type of specifications supported for purchase;<br
- Calculation volume: The number of computational tasks that can be performed by a certain computational power in a certain time period, = computational power * duration (seconds). For instance, if one g.n2.large instance runs for one hour, its computation volume is 1×2×3600.
我们的产品专家为您找到最合适的产品/解决⽅案
1v1线上咨询获取售前专业咨询
专业产品顾问,随时随地沟通