The statement is false . Huawei Cloud ECS supports multiple billing models, and cost optimization is not exclusive to pay-per-use billing. The principal modes include yearly/monthly, pay-per-use, and spot pricing , with each optimized for different workload characteristics.

Pay-per-use is useful for workloads with uncertain or fluctuating demand because resources are billed according to actual usage and can be released when no longer required. However, yearly/monthly billing can provide a lower effective price for stable, long-running workloads because Huawei Cloud applies subscription discounts; the longer the committed subscription period, the greater the potential discount. Therefore, for predictable production workloads running continuously, yearly/monthly billing can be substantially more economical than pay-per-use.

Spot pricing is another cost-saving mechanism. Spot instances use market-based pricing and can provide compute resources at a lower price where workload interruption characteristics permit their use. Consequently, several ECS billing models can reduce expenditure when aligned to an appropriate workload.

The incorrect portion of the statement is therefore the assertion that only pay-per-use can reduce costs .

Reference topics: Elastic Cloud Server; ECS Billing Modes; Yearly/Monthly; Pay-per-Use; Spot Pricing.

===============

1 → PaaS

2 → IaaS

3 → SaaS

The three mappings correspond directly to Huawei Cloud's definitions of the standard cloud service models. Infrastructure as a Service (IaaS) supplies foundational infrastructure resources such as compute, storage, and networking. Huawei Cloud services such as ECS, VPC, VPN, and OBS are representative infrastructure-level services. Therefore, the description referring to computers, storage devices, and networks maps to IaaS.

Platform as a Service (PaaS) provides managed platform capabilities used to develop, deploy, and operate applications. Huawei Cloud characterizes PaaS as including middleware, database platforms, development environments, microservice capabilities, and related application-platform resources. Consequently, the database/middleware/MySQL/MongoDB/Java-oriented description belongs to PaaS.

Software as a Service (SaaS) delivers complete applications to end users. Customers consume the application rather than maintaining the underlying operating system, middleware, runtime, and infrastructure. Enterprise applications, messaging systems, email services, collaboration applications, and similar ready-to-use software fit this model.

The correct progression is therefore PaaS → IaaS → SaaS for the three descriptions shown in the question.

Reference topics: Cloud Service Models; IaaS; PaaS; SaaS; Huawei Cloud Service Architecture.

===============

Option A is incorrect because Huawei Cloud Auto Scaling is designed to automatically adjust ECS instances and bandwidth resources , not Bare Metal Server instances. Huawei's AS documentation defines an AS group as a collection of ECS instances and states that scaling actions add or remove ECS instances according to alarm, scheduled, or periodic policies. BMS provides dedicated physical compute and is provisioned through the BMS service rather than being dynamically created and destroyed by an ECS Auto Scaling group.

The other relationships are valid. Cloud Eye supports BMS operating-system monitoring after the Cloud Eye Agent is installed; the BMS reports detailed CPU, memory, disk, and network metrics to Cloud Eye. EVS is integrated with BMS and additional supported EVS data disks can be attached when more storage is required. IMS is likewise an integral BMS service: images can be used to create consistently configured BMSs and to reinstall or change their operating systems, reducing repetitive deployment work. Huawei's BMS architecture explicitly lists VPC, IMS, EVS, Cloud Eye, and backup services as integrations supporting BMS operation.

Reference topics: Bare Metal Server Architecture; BMS and Other Cloud Services; Auto Scaling; Cloud Eye; IMS; EVS.

===============

Huawei Cloud establishes a clear parent-child relationship between a Huawei Cloud account and the IAM users created under that account. The account owns the cloud resources and is the billing entity , whereas IAM users consume or manage resources according to permissions assigned by the account administrator.

Therefore, option C is correct: IAM users do not independently pay for the resources they use. Charges generated through their authorized operations are billed to the parent account. Options B and D are also correct because the account administrator retains authorization control and can modify or revoke permissions assigned to IAM users.

Option A is incorrect because it incorrectly treats the IAM user as a payment entity. Authorization determines what an IAM user is permitted to access or operate , not who is responsible for payment. Huawei Cloud explicitly distinguishes resource ownership and billing from delegated resource access. An IAM user does not receive an independent bill merely because access to a resource has been authorized.

Accordingly, the correct relationship is: account owns and pays; IAM users operate resources according to delegated permissions; the account controls those permissions .

Reference topics: IAM Users; Huawei Cloud Accounts; IAM Authorization; Billing Responsibility.

===============

The statement is TRUE . Huawei Cloud makes an explicit architectural distinction between Elastic Cloud Servers and Bare Metal Servers regarding the underlying compute resources. An ECS is a virtualized compute instance. Multiple ECS instances can therefore consume resources provided by the same underlying physical host, with virtualization providing logical isolation between tenants and workloads.

A BMS, by contrast, provides exclusive access to physical compute resources. Huawei Cloud states directly that BMS users can exclusively access physical resources, whereas ECS users share those physical resources. This makes BMS particularly suitable for mission-critical systems, big-data clusters, enterprise middleware, high-performance workloads, and applications that require predictable performance or direct access to physical CPU and memory resources.

Huawei's BMS comparison also identifies Exclusive resources = Yes for BMS and traditional physical servers, but No for ECS. BMS additionally provides cloud characteristics such as automatic provisioning, VPC integration, image-based deployment, monitoring, and cloud storage while retaining physical-server performance.

Therefore, the differentiation presented in the question is technically accurate.

Reference topics: Bare Metal Server Fundamentals; BMS versus ECS; ECS Virtualization; Exclusive Physical Resources.

===============

The statement is FALSE because directly binding an EIP to an ECS is only one of several Huawei Cloud mechanisms for providing Internet access. An ECS with its own EIP can communicate directly with the public network, but ECSs can also access the Internet without having an EIP individually bound to each server .

The principal alternative is a public NAT Gateway using an SNAT rule . Huawei Cloud states that servers in a VPC subnet can share an EIP through a public NAT gateway. The NAT gateway translates their private source addresses into the configured public EIP, enabling outbound Internet connectivity while avoiding the need to expose or allocate an individual public IP address to every ECS.

Huawei's ECS documentation also explicitly describes “Internet connectivity for an ECS without an EIP bound” and recommends NAT Gateway as the standard solution. A proxy/SNAT ECS can also provide an Internet access path in suitable architectures.

Therefore, binding an EIP directly to the ECS is a valid method, but it is not the only method , making the statement false.

Reference topics: Elastic IP; Public NAT Gateway; SNAT; ECS Internet Connectivity; VPC Networking.

Option D is incorrect . The HCIA-Cloud Service V3.5 comparison positions SFS Capacity-Oriented as the large-capacity, high-bandwidth, cost-sensitive file-storage option. Huawei describes its principal advantages as large capacity, high bandwidth, and low cost , making it particularly appropriate for media processing, large-scale file sharing, HPC, and backup workloads.

The performance characteristics also support A and B. Huawei's documented SFS Capacity-Oriented implementation provides substantially greater maximum aggregate bandwidth than the previous-generation SFS Turbo Standard class. However, its access latency is approximately 3–20 ms , whereas SFS Turbo Standard provides lower latency, historically around 2–5 ms . SFS Turbo is therefore more appropriate where low-latency access and workloads involving large numbers of small files are important.

Capacity-Oriented storage can scale to petabyte-scale capacities and is specifically optimized economically for large-volume storage. Consequently, for equivalent expenditure, it can provide significantly greater capacity than performance-oriented SFS Turbo configurations.

Option D reverses this economic relationship. SFS Capacity-Oriented is not positioned as the more expensive alternative for equal capacity; low cost at large scale is one of its defining characteristics .

Reference topics: SFS File System Types; SFS Capacity-Oriented; SFS Turbo Standard; File Storage Performance and Cost.

===============

The statement is TRUE. Huawei Cloud OBS provides several resource-based authorization mechanisms, including bucket policies and object policies. Their principal distinction is the scope of resources to which they apply.

Huawei's official OBS Permission Configuration Guide explicitly states that an object policy applies to only one object in a bucket, whereas a bucket policy applies to multiple objects or all objects within the bucket. This enables administrators to select the authorization mechanism that best matches the required granularity.

An object policy is useful when permissions must be configured specifically for an individual object, such as granting another principal read or write permissions to one particular resource without broadening access to the rest of the bucket.

A bucket policy supports broader authorization scenarios. Its policy statements can define principals, actions, resources, effects, and conditions to grant or restrict access to selected objects, prefixes, or the entire bucket. Typical uses include allowing another Huawei Cloud account to access specified OBS resources or granting IAM identities access to defined bucket content.

The question therefore correctly distinguishes object-level policy scope from bucket-level multi-object policy scope.

Reference topics: OBS Permission Management; Object Policies; Bucket Policies; Resource-Based Authorization.

===============

The statement is false because its description of infrastructure ownership does not correctly characterize a hybrid cloud. A hybrid cloud combines different cloud environments—most commonly private cloud and public cloud —and enables workloads, applications, or data to be distributed between them according to security, capacity, latency, compliance, or business requirements.

Huawei Cloud describes hybrid cloud as a blend of private and public cloud environments. The private portion may be dedicated to a particular customer and may host confidential or sensitive workloads, while public cloud resources are delivered through infrastructure operated by a cloud service provider. Consequently, the combined infrastructure is not inherently owned and operated by a single organization .

This ownership distinction separates hybrid cloud from a purely private-cloud model. A private cloud can be dedicated to a single organization, whereas hybrid architecture deliberately combines environments that may have different infrastructure ownership and operating responsibility.

The statement correctly recognizes that hybrid cloud is flexible and that workloads can use different cloud environments, but because it makes an absolute and incorrect assertion about single-organization ownership, the entire statement must be evaluated as FALSE .

Reference topics: Cloud Deployment Models; Public Cloud; Private Cloud; Hybrid Cloud; CloudPond Service Overview.

===============

Statements B and C are false . Huawei Cloud distinguishes alarm-based, scheduled, and periodic Auto Scaling policies. An alarm policy triggers a scaling operation in response to monitored metrics such as CPU utilization. Scaling “every three hours” is time-interval behavior and therefore belongs to a periodic policy , not an alarm policy. This makes B incorrect.

Statement C is also incorrect. The expected number of instances is the desired number of ECS instances that should currently run within an AS group. Huawei specifies that this value must remain between the configured minimum and maximum instance counts. It can be modified manually or changed through scheduled, periodic, or alarm-based policies. It is not synonymous with the minimum instance count during peak hours.

Statement A is valid because an AS configuration can use a Huawei Cloud public image; Huawei's AS configuration parameters explicitly support public, private, and shared images.

Statement D is also valid. Huawei specifies that when an existing ECS is manually added to an AS group, the ECS must not already be a member of another AS group .

Consequently, the two false statements are B and C .

Reference topics: Auto Scaling Configurations; AS Policies; Expected Instances; AS Groups.

===============