As covered in the " Cloud Computing Trends " and " Huawei Cloud Solutions " modules of the HCIA curriculum, OpenStack and FusionCompute occupy different layers of the cloud stack. Statement A is correct: OpenStack is a global open-source project maintained by a community, whereas FusionCompute is Huawei’s proprietary commercial virtualization platform (hypervisor).4

Statement B captures a fundamental technical difference. OpenStack is anorchestratoror " Cloud OS " ; it lacks an internal hypervisor and must rely on external technologies (like KVM, Xen, or FusionCompute itself) to perform the actual virtualization. In contrast, FusionCompute is a standalonevirtualization suitethat includes the CNA (hypervisor) and VRM (management), allowing it to directly pool and manage physical hardware resources.

Statement C highlights the broader scope of OpenStack, which is designed to manage heterogeneous resources, including Bare Metal (Ironic) and Containers (Zun/Magnum). FusionCompute is a specialized tool primarily focused on Virtual Machines (VMs). Finally, Statement D refers to theMulti-tenancymodel. OpenStack was built for public and large private clouds, using " Projects " and " Tenants " to strictly isolate resources and billing between different organizations. FusionCompute is a lower-level resource manager that focuses on " Clusters " and " Resource Pools " rather than the complex tenant-based isolation logic found in high-level cloud platforms. In the Huawei ecosystem, FusionCompute often serves as the " Computing " component underneath a largerHuawei Cloud Stackor OpenStack-based solution, illustrating how these two technologies complement each other in a modern data center.

In OpenStack:

Swiftprovidesobject storage, designed for storing unstructured data such as images, videos, and backups.

Cinderprovidesblock storage, which is used by virtual machines as virtual disks.

Since Swift doesnotprovide block storage, the statement isFALSE.

In the official Huawei HCIA–Cloud Computing documentation, the core definition of virtualization centers on the abstraction of physical resources. The essence of virtualization technology is precisely the transformation of physical, hardware-based resources—such as CPUs, memory, network interface cards, and storage disks—into virtualized, logical entities. This process is facilitated by a software layer known as the Hypervisor or Virtual Machine Monitor (VMM). By creating this abstraction layer, virtualization effectively achieves the " decoupling " of the software environment (the Operating System and applications) from the underlying physical hardware.

This decoupling is a revolutionary shift from traditional computing architectures where an Operating System was tied to specific hardware drivers and physical configurations. In a virtualized environment, a Virtual Machine (VM) perceives a set of standardized, logical hardware components provided by the Hypervisor. Because the software is no longer bound to the physical machine, it gains significant mobility and flexibility. For example, in Huawei’s FusionCompute platform, this decoupling allows for features such as Live Migration, where a running VM can be moved from one physical host to another without any service interruption. This is possible because the VM ' s " logical " hardware remains consistent even as the " physical " host changes.

Furthermore, this transformation from physical to logical enables better resource utilization through overcommitment and sharing. Multiple logical servers can run on a single physical server, each isolated from the others. This fundamental shift is what provides the agility, scalability, and efficiency required for Cloud Computing. Without the decoupling provided by virtualization, the rapid provisioning and elastic resource scaling that define cloud services would be impossible to achieve. Therefore, virtualization serves as the indispensable technical foundation upon which all other cloud services and management layers are built within the Huawei ICT ecosystem.

In the official Huawei HCIA-Cloud Computing curriculum, specifically within the " Huawei Virtualization Platform Management " module, the lifecycle and resource adjustment of Virtual Machines (VMs) are covered in detail. While virtualization provides significant elasticity, certain physical-to-logical mappings have technical constraints. According to the training materials, while you can add disks or increase the capacity of an existing virtual disk (expansion), disk capacity cannot be reduced once it has been allocated and formatted by the Guest Operating System.

The reason for this limitation is rooted in data integrity. Reducing a disk ' s size at the virtualization layer (CNA/VIMS) would involve " chopping off " the end of the logical block range. Since the Guest OS manages the file system and may have placed metadata or data at the end of that range, shrinking the disk from the outside would almost certainly result in file system corruption and data loss. Therefore, Huawei FusionCompute does not provide a " shrink " function for virtual disks. If a user needs a smaller disk, they must create a new, smaller disk and migrate the data at the application level.

In contrast, other resources offer more flexibility. NIC quantities (A) can be reduced by deleting a virtual network adapter from the VM configuration. CPU resources (C) and Memory resources (D) can be adjusted in the VM settings; while increasing them is common, they can also be reduced (typically requiring a VM restart to allow the Guest OS to re-initialize the hardware map). This distinction is a key exam point: compute resources are elastic in both directions (with reboot), but storage capacity in FusionCompute is strictly " expand-only " to protect user data.

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

FusionAccess consists of multiple components, each with different roles and data persistence requirements.

WI (Web Interface)stores configuration and user access information and supports backup.

VLB (Virtual Load Balancer)maintains configuration data that can be backed up.

vAG (Virtual Access Gateway)also stores configuration-related data that supports backup and recovery.

TheLicensecomponent, however, isnot backed upas configuration or system data. Licensing information is typically re-imported or reactivated during recovery rather than restored from backup.

Therefore, the component that doesnothave backup data isLicense.

Huawei FusionCompute leverageshardware-assisted virtualization technologiessuch asIntel VT-x, EPT, and AMD-V, which significantly reduce the overhead associated with CPU and memory virtualization.

Memory virtualization traditionally involved complex software-based address translation, which introduced performance overhead. With hardware-assisted features such asExtended Page Tables (EPT), memory mapping is handled directly by the CPU, improving efficiency and performance.

This approach is explicitly highlighted in HCIA–Cloud Computing materials as a key reason why FusionCompute delivers near-native performance for virtual machines.

Therefore, the statement isTRUE.

As detailed in the Huawei HCIA-Cloud Computing technical documentation, FusionCompute offers a robust suite of resource management functions. Statement A is correct: FusionCompute datastores support multiple file system formats to accommodate different storage backends. VIMS (Virtual Image Management System) is Huawei’s high-performance cluster file system used for SAN storage; NFS is used for Network Attached Storage; and EXT4 is used for local disk virtualization on a single host. This variety allows administrators to build heterogeneous storage pools.

Statement B is also a verified function. FusionCompute supports dynamic resource adjustment. Users can modify memory capacity for both offline VMs (while the VM is stopped) and online VMs (while running). Online memory expansion requires the " Hot Add " feature and Guest OS support, while reduction typically requires the VM to be in a state where the kernel can release the address space. Statement C accurately describes Storage Live Migration. This feature allows a VM ' s virtual disk files to be moved from one datastore to another (or within the same storage device) while the VM remains running, which is essential for storage maintenance and load balancing.

Statement D is incorrect because of the word " only. " While traditional VM live migration (vMotion style) does require shared storage so that both hosts can see the same disk, modern versions of FusionCompute also support Shared-Nothing Live Migration. This technology migrates both the VM ' s memory state and its disk data simultaneously over the network, allowing a VM to move between hosts that do *not* share a common backend storage. Because D ignores this advanced capability, it is excluded from the list of correct functions in the context of FusionCompute ' s full feature set.

RAID 0

➡Multiple disks are combined to form a large-capacity storage device. This RAID level supports data striping to improve read and write speed, but does not provide data verification or redundancy backup.

RAID 1

➡Data on one disk is mirrored to another disk to provide complete redundancy backup.

RAID 5

➡Parity data is distributed on all disks in the array. Write operations can be performed on different disks at the same time.

RAID 10

➡Mirror backup is performed for each disk, and then the disks are combined to form a storage device, providing high read and write performance in addition to redundancy backup.

User data is stored in a network directory instead of a local host

➡Folder redirection

Redirects user folders (such as Desktop, Documents) to network storage

Prevents local data storage on virtual desktops

Enhances data security and centralized management

The client IP address segment is not configured. Users can access VMs from any IP address segment

➡Access control policy

Controls or relaxes access based on client IP segments

If not configured, access is allowed fromany IP address

Used to restrict or permit VM access based on network location

A parallel-to-USB cable can be used to connect the parallel port device to the client

➡Peripheral policy creation

Controls usage of peripherals such as USB, printers, scanners

Enables or disables device redirection

Supports special device scenarios via policy configuration