Virtualization and Cloud Computing practice questions

Answer: A) Multi-tenancy

• AMulti-tenancy describes a cloud provider serving multiple separate customers (tenants) from shared underlying physical infrastructure, while using logical isolation (virtualization, access controls) to keep each tenant's resources and data separate from the others.
• BDedicated single-tenant infrastructure is essentially the opposite of multi-tenancy, which specifically involves multiple customers sharing the same underlying infrastructure.
• CA Type 2 hypervisor is locally run virtualization software on one machine, not the description of a provider serving many separate customers from shared infrastructure.
• DOn-premises computing is confined to a single organization's own facility, the opposite of a shared multi-customer cloud infrastructure.

Answer: A) Measured/metered service (pay-as-you-go)

• AMetered utilization means usage is tracked and billed based on actual consumption, a defining characteristic of cloud computing's pay-as-you-go pricing model.
• BHigh availability describes uptime/resilience guarantees, not how billing is calculated.
• CRapid elasticity describes automatic scaling of resources, a separate characteristic from how usage is billed.
• DResource pooling describes a provider serving multiple customers from shared physical infrastructure, not the billing method itself.

Answer: A) Cloud backup

Scheduled offsite copies that can be restored to an earlier point in time are cloud backup. File synchronization copies changes, including deletions and corruption, to every device, so it is not a backup.

• ACloud backup keeps copies offsite on a schedule, often as several point-in-time versions. The firm can restore data after hardware loss, a site disaster, or corruption.
• BSync keeps files consistent across devices, so a change made in one place appears everywhere. Corrupted or deleted files would be copied to every location, so sync alone does not provide restorable point-in-time versions.
• CResource pooling describes how a cloud provider shares its infrastructure among many tenants. It is a cloud characteristic, not a way to protect data.
• DRapid elasticity is the ability to scale resources up or down quickly as demand changes. It does nothing to preserve earlier copies of files.

Answer: A) CPU overcommitment, with VMs contending for time on the physical cores

Assigning more total vCPUs than the host has physical cores is CPU overcommitment. It works when VMs are mostly idle, but causes contention and slowdowns when many VMs need CPU at the same time.

• ACorrect. A vCPU is a share of time on the physical cores, not a dedicated core. Assigning more vCPUs than there are cores works while VMs are mostly idle. When several VMs need CPU at the same time, they have to wait for physical core time and performance drops.
• BIncorrect. Ballooning reclaims memory from guests. The scenario is about vCPUs outnumbering physical cores and slowdowns under simultaneous CPU load, not RAM.
• CIncorrect. Without hardware virtualization support, a modern hypervisor generally wouldn't run VMs at all. It wouldn't cause slowdowns only during busy periods.
• DIncorrect. A virtual switch problem would affect network traffic. It wouldn't explain slowdowns tied to the ratio of vCPUs to physical cores under heavy CPU load.

Answer: A) Egress charges billed per gigabyte of data leaving the provider's network

Transferring large amounts of data out of a cloud provider triggers egress charges, which are usually billed per gigabyte. Inbound transfer is typically free, which is why large outbound transfers often produce unexpectedly high bills.

• ACloud providers commonly charge for data transferred out of their network, often per gigabyte beyond a small free allowance. Moving terabytes to an on-premises site creates a large amount of billable outbound traffic.
• BIn this scenario the data left the cloud; it did not enter it. Major providers also typically do not charge for inbound transfer from the internet.
• CThe scenario says storage usage stayed the same, and copying data out does not add stored capacity. Storage charges do not explain the sudden increase.
• DThe scenario says compute usage stayed the same. Instance-hour charges depend on how long VMs run, not on how much data was transferred out.

Answer: A) Private cloud

• AA private cloud is dedicated entirely to one organization, whether hosted on their own premises or by a provider on dedicated hardware, giving maximum control and customization at the cost of not sharing infrastructure costs with anyone else.
• BA public cloud shares infrastructure among many customers on the provider's own multi-tenant hardware, the opposite of the dedicated, non-shared setup described here.
• CA community cloud is shared among several organizations with common interests, not dedicated to just one organization alone.
• DThe scenario explicitly wants dedicated, non-shared infrastructure — a hybrid cloud combines private and public elements, not a description of a single dedicated private environment.

Answer: A) A hypervisor shares out the host's CPU, memory, and storage, giving each guest system its own allocated portion

Several independent VMs can run on one physical server because a hypervisor gives each VM its own share of the host's CPU, memory and storage. Clustering, a dedicated server per department, and SaaS each describe a different model.

• ACorrect. In virtualization, a hypervisor sits between the physical hardware and the guest operating systems. It allocates part of the host's CPU, memory and storage to each VM, so several independent VMs can run on one physical machine.
• BIncorrect. The scenario has only one physical server. Giving each department its own physical server is the opposite of consolidating workloads through virtualization.
• CIncorrect. Clustering combines several physical machines into one logical system. This scenario does the reverse: it splits one physical machine into several logical servers.
• DIncorrect. SaaS delivers finished applications from a provider over the internet. In this scenario the company runs full VMs with their own operating systems on its own server.

Answer: A) Hybrid cloud

Keeping critical workloads on premises while bursting extra demand into a public cloud is a hybrid cloud. Private, public and community describe single deployment models, not this combination.

• AA hybrid cloud combines on-premises or private infrastructure with public cloud resources in one environment. Sending extra demand to the public cloud while keeping critical workloads in-house is called cloud bursting, a classic hybrid scenario.
• BA community cloud is infrastructure shared by several organizations with common requirements, such as regulatory needs. This scenario involves one organization combining its own data center with a public provider.
• CA private cloud is dedicated to one organization, often on premises. This company also uses a third-party public cloud, so private cloud alone doesn't describe the setup.
• DA public cloud is run by a third-party provider and shared among customers. Here the sensitive data stays in the company's own data center, so public cloud alone doesn't describe the setup.

Answer: A) Set a virtual machine reserve for the baseline and a virtual machine limit of 50%

In Hyper-V, the virtual machine reserve guarantees a minimum CPU share under contention, and the virtual machine limit is a ceiling that is always enforced. Relative weight only adjusts a VM's share when VMs compete and guarantees nothing.

• ACorrect. The virtual machine reserve ensures the VM gets at least its reserved CPU when there is contention. The virtual machine limit is always enforced, so the VM stays at 50% even when nothing else is running.
• BIncorrect. This reverses the settings. A limit sets a ceiling, not a guaranteed floor. A reserve sets a minimum and does not stop the VM from using more when spare CPU is available.
• CIncorrect. Relative weight only gives the VM a larger share during contention and provides no absolute guarantee. Adding virtual processors increases potential usage rather than capping it at 50%.
• DIncorrect. Relative weight guarantees no minimum, and a reserve is a floor, not a ceiling. When the host is idle, Hyper-V lets a VM use more than its reserve, so the 50% requirement would not be enforced.

Answer: A) Thick provisioning

• AThick provisioning reserves the VM's entire allocated virtual disk space on physical storage upfront, guaranteeing that capacity is available, unlike thin provisioning, which allocates physical space only as data is actually written.
• BThin provisioning allocates physical storage only as data is actually written, growing over time — the opposite of reserving the full size immediately.
• CA snapshot captures a VM's point-in-time state for later reversion; it's unrelated to how a virtual disk's storage allocation is provisioned.
• DA VLAN is a network segmentation concept, unrelated to virtual disk storage provisioning.