Billing

HUAWEI CLOUD DMS for Kafka supports pay-per-use and yearly/monthly billing. For details, see Pricing Details.

Billing Items

HUAWEI CLOUD DMS for Kafka is billed based on Kafka instance specifications and storage space. There are two types of specifications. The specifications available in the CN East-Shanghai1, CN East-Shanghai2, CN North-Beijing4, and CN South-Guangzhou regions are different from those in other regions.

Table 1 DMS for Kafka billing

Billing Item

Description

Instance

  • Kafka instances in regions other than CN East-Shanghai1, CN East-Shanghai2, CN North-Beijing4, and CN South-Guangzhou are billed based on their bandwidth. To ensure stable service running, you are advised to choose a bandwidth 30% higher than the actual throughput expected for read or write traffic, whichever is higher. You should also consider other parameters described in Table 2.

    For example, if the maximum write traffic is 70 MB/s and the maximum read traffic is 200 MB/s, the bandwidth to be chosen should be at least 200 MB/s x (1 + 30%) = 260 MB/s.

  • Kafka instances in regions CN East-Shanghai1, CN East-Shanghai2, CN North-Beijing4, and CN South-Guangzhou are billed based on the ECS flavor and the number of brokers. When purchasing an instance, select appropriate ECS flavors and the number of brokers based on service evaluation. Table 3 lists the performance per broker.
  • Kafka instances can be billed on a yearly/monthly or pay-per-use (hourly) basis.
  • If you enable the dumping function, there will be additional charges.

    For example:

    • If you enable dumping for a 100 MB/s instance (3 brokers), 2 more brokers with the same specifications will be created for dumping and you need to pay for them.
    • If you enable dumping for a c6.4u8g.cluster instance, 2 more brokers will be created with the same flavor for dumping and you need to pay for them.

Storage space

  • Kafka instances are also billed based on the storage space. In the CN East-Shanghai1 region, you can choose the common I/O or ultra-high I/O disk type. In other regions, you can choose the common I/O, high I/O, or ultra-high I/O disk type to meet your service requirements.

    You can specify the number of replicas. For example, if the required disk size to store message data is 500 GB and there are three replicas, the disk capacity should be at least: 500 GB x 3 = 1500 GB.

  • Table 3 describes the storage space options in 100 GB increments.
  • The storage space can be billed on a yearly/monthly or pay-per-use (hourly) basis.
Table 2 Specifications of Kafka instances in regions other than CN East-Shanghai1, CN East-Shanghai2, CN North-Beijing4, and CN South-Guangzhou

Specification (Bandwidth)

Maximum Partitions

Maximum Consumer Groups

Maximum Client Connections

Storage Space

100 MB/s

300

60

3000

600 GB–90,000 GB

300 MB/s

900

300

10,000

1200 GB–90,000 GB

600 MB/s

1800

600

20,000

2400 GB–90,000 GB

1200 MB/s

1800

600

20,000

4800 GB–90,000 GB

Table 3 Specifications of Kafka instances in regions CN East-Shanghai1, CN East-Shanghai2, CN North-Beijing4, and CN South-Guangzhou

Flavor

Brokers

Maximum TPS per Broker

Maximum Partitions per Broker

Maximum Consumer Groups per Broker

Maximum Client Connections per Broker

Storage Space

c6.2u4g.cluster

3–30

30,000

250

20

1000

600 GB–300,000 GB

c6.4u8g.cluster

3–30

100,000

500

100

2000

1200 GB–600,000 GB

c6.8u16g.cluster

3–30

150,000

1000

150

3000

2400 GB–900,000 GB

c6.12u24g.cluster

3–30

200,000

1500

200

4000

3600 GB–900,000 GB

c6.16u32g.cluster

3–30

250,000

2000

200

5000

4800 GB–900,000 GB

Mapping Between Old and New Flavors

Table 4 compares the old and new Kafka instance flavors.

Table 4 Mapping between old and new Kafka instance flavors

Old Flavor

Maximum Partitions (Old)

TPS (Old)

New Flavor

Maximum Partitions (New)

TPS (New)

100 MB/s (c6.large.2 x 3)

300

100,000

c6.2u4g.cluster x 3

750

100,000

300 MB/s (c6.xlarge.2 x 3)

900

300,000

c6.4u8g.cluster x 3

1500

300,000

600 MB/s (c6.2xlarge.2 x 4)

1800

600,000

c6.8u16g.cluster x 4

4000

600,000

1200 MB/s (c6.2xlarge.2 x 8)

1800

1.2 million

c6.8u16g.cluster x 8

8000

1.2 million

Assume that you have four 8 vCPUs | 16 GB VMs, and require three replicas and 10,000 to 15,000 partitions. The write traffic of each VM is 720 MB/s, the read traffic is 7.2 GB/s, and the replication traffic is 90 MB/s x 2. The flavor and cost of your Kafka instance are calculated as follows:

Total write traffic: 720 MB/s x 3 x 4 = 1080 MB/s

Total read traffic: 7.2 GB/s x 4 = 3686 MB/s

Total replication traffic: 90 MB/s x 2 x 4 = 720 MB/s

Total traffic: 1080 MB/s + 3686 MB/s + 720 MB/s = 5486 MB/s

To ensure stable services, the total traffic should be only 80% of the bandwidth you configure for your instance, so the actual total traffic should be 5486 MB/s/80%=6857.5 MB/s.

If the instance uses an ultra-high I/O disk, a singe broker can support a maximum of 350 MB/s traffic. Therefore, the number of required brokers is 20 (6857.5/350). The flavor of the Kafka instance should be c6.8u16g.cluster x 20. A c6.8u16g.cluster instance supports a maximum of 1000 partitions per broker, so c6.8u16g.cluster x 20 can meet your requirement for 10,000 to 15,000 partitions.

The minimum storage space of a single broker is 800 GB, so the total storage space is 800 GB x 20 = 16,000 GB.

Kafka instance's total price: ¥48,600 (instance price) + ¥16,000 (storage space price) = ¥64,600/month.

Billing Modes

Two billing modes are available, allowing you to pay less by using more.

  • Yearly/Monthly: Provides a larger discount than pay-per-use mode and is recommended for long-term users.
  • Pay-per-use (hourly) mode: More flexible, enabling you to start and stop services anytime. You pay only for what you use. The minimum time unit is one hour. Less than an hour is recorded as an hour.

Changing Configurations

  • You can change the bandwidth for a Kafka instance. You will then be billed based on the new specifications immediately after the change.
  • You can change the number of brokers for a Kafka instance. You will then be billed based on the new specifications immediately after the change.
  • You can also change the storage space of Kafka queues. You will be billed based on the new storage space immediately after the storage space increase. Storage space can only be increased, and cannot be decreased. The minimum increase increment is 100 GB.

Renewal

You can renew a resource package upon its expiration, or you can set auto-renewal rules for a resource package. For more information about renewing resource packages, see Renewal Management.