Billing Instructions for BYOC

Referencing the billing summary, this article offers an in-depth explanation of the various billing items associated with the AutoMQ Cloud product within the BYOC environment.

The terms 'cloud providers' and 'public cloud providers' used in this article denote major cloud service providers, including AWS, Google, Cloud, Azure, Alibaba Cloud, Tencent Cloud, Huawei Cloud, and others.

Software Service Fees (payable to AutoMQ)

In a BYOC setup utilizing AutoMQ Cloud, software service fees are calculated based on the actual size of the user's cluster. The scale metric for AutoMQ software service fees comprises the following:

• Message Processing Specification AKU (AutoMQ for Kafka Unit): Mandatory, AKU is employed to quantify the scale of computing resources allocated for message processing within the Kafka cluster. The AKU specification correlates directly with the cluster's scale pressure.

AKU Cost for Message Processing Specification

Metric Constraints

AKU denotes the computing resources allocated for message transmission within the cluster. AKU accounts for computing processing, storage IOPS, and network throughput resource consumption.

Selecting the appropriate AKU specification for each AutoMQ instance is crucial for achieving the desired message transport throughput. Based on benchmark performance tests, each AKU specification enhances performance as follows:

Capability per 1 AKU	Capability per 1 AKU	Description
Message Writing Throughput	20 MB/s	Core business metrics and stress of message read/write operations. Exceeding the specified capacity could slow down service responses, increase RT, or result in throttling failures.
Message Reading Throughput	20 MB/s	Core business metrics and stress of message read/write operations. Exceeding the specified capacity could slow down service responses, increase RT, or result in throttling failures.
Client Request Frequency	800 requests/sec	The frequency and stress of requests from applications accessing the server through Kafka Producer, Consumer SDK. Exceeding specified specifications may lead to throttling or increased RT. Request types include:
Produce FetchConsumer CommitOffset
Maximum Number of Partitions	1125	The number of partitions constrains the scale of metadata an instance can simultaneously carry. Exceeding the specified capacity may result in the inability to create new Topics or expand partitions.

It is advised in production environments to carefully monitor resource consumption according to these specifications and dynamically adjust resources as necessary to prevent overutilization, which could strain the cluster and compromise service stability.

Billing Rules (e.g., Hourly Billing)

Each AutoMQ instance (cluster) allows users to specify the required AKU specifications upon creation or when changing specifications.

• Billing scope: AKU counts for each AutoMQ Kafka instance.

• Billing method: Calculated hourly based on the peak consumption recorded during the billing cycle.

• Aggregation method: Aggregated and calculated at the per-instance level.

• Billing unit: AKU·hour, which means that the actual billing consumption is measured in hourly increments, with any fraction of an hour rounded up to the next full hour.

Billing Example (hourly Billing for an Instance)

Based on the rules outlined above, here are several typical billing scenarios:

Example one:

At 10:00, a user created an instance with a capacity of 12 AKUs. By 11:00, they had increased the capacity to 24 AKUs.

Between 10:00 and 12:00, the total AKU usage was:

12 1 hour + 24 1 hour = 36 AKU·hours.

This example demonstrates the real-time billing method prevalent in the cloud market, where software service charges are calculated on an hourly basis. Both monthly settlements and pre-purchased subscription licenses necessitate the evaluation of AKU consumption for each instance.

Cloud Resource Costs (fees Paid to Cloud Providers for the Resources Hosting the Instances)

In deploying AutoMQ instances, users not only pay software service fees but also bear costs for associated resources. For instance, using Alibaba Cloud:

Cloud Product	Specifications and Usage	Dependency Explanation
ECS (Virtual Server)	Specifications: Vary by region and may differ. Usage: Scales with user cluster size.	AutoMQ Kafka virtual server consumption varies with user cluster size. Even without creating an instance, at least one server is consumed for deploying the management interface.
Cloud Disk (Elastic Block Service)	Specification: PL1 Type Usage: Each Kafka data node consumes 40GB	AutoMQ Kafka utilizes cloud disk storage for temporary message data.
Object Storage	Specification: Standard Disaster Recovery Storage Usage: 2 buckets per environment, storage space scales with user environment resource consumption	The core storage dependency of the message cluster. S3 charges cloud providers based on actual space used and call volume.