20Which AWS service is mainly used for object storage?
AWS
Easy
A.Amazon EC2
B.Amazon RDS
C.Amazon S3
D.Amazon Lambda
Correct Answer: Amazon S3
Explanation:
Amazon Simple Storage Service (S3) is an object storage service used for files, backups, and other data.
Incorrect! Try again.
21An online store automatically adds computing instances during a festival sale and removes them when demand falls. Which cloud characteristic is being demonstrated?
Cloud Fundamentals
Medium
A.Measured service
B.Location independence
C.Rapid elasticity
D.Broad network access
Correct Answer: Rapid elasticity
Explanation:
Rapid elasticity allows resources to expand and contract automatically in response to changing demand.
Incorrect! Try again.
22A service runs on four instances, each providing 99% independent availability. A load balancer can serve users as long as at least one instance is available. Which principle primarily improves the service availability?
Cloud Fundamentals
Medium
A.Consolidation onto one server
B.Encryption of stored data
C.Redundancy across instances
D.Compression of network traffic
Correct Answer: Redundancy across instances
Explanation:
Using independent redundant instances reduces the chance that a single failure makes the entire service unavailable.
Incorrect! Try again.
23A development team wants to deploy application code without managing operating systems, runtime patches, or virtual machines. Which service model is the best fit?
Cloud Service
Medium
A.Colocation as a Service
B.Software as a Service
C.Infrastructure as a Service
D.Platform as a Service
Correct Answer: Platform as a Service
Explanation:
PaaS manages the infrastructure, operating system, and runtime while allowing the team to focus on application code.
Incorrect! Try again.
24A company uses an Infrastructure as a Service virtual machine. Under the shared responsibility model, which task is normally the customer's responsibility?
Cloud Service
Medium
A.Securing the provider's data center
B.Replacing failed physical disks
C.Patching the guest operating system
D.Maintaining the host hypervisor
Correct Answer: Patching the guest operating system
Explanation:
In IaaS, the provider manages physical infrastructure and virtualization, while the customer usually manages the guest operating system and applications.
Incorrect! Try again.
25A bank keeps regulated customer records in its private cloud but temporarily uses a public cloud for additional computing capacity during month-end processing. Which deployment model does this represent?
Deployment Models
Medium
A.Hybrid cloud
B.Private cloud
C.Public cloud
D.Community cloud
Correct Answer: Hybrid cloud
Explanation:
A hybrid cloud integrates private and public cloud resources so workloads or data can be distributed according to requirements.
Incorrect! Try again.
26Several government agencies with similar security and compliance requirements jointly operate a cloud platform available only to those agencies. Which deployment model is most appropriate?
Deployment Models
Medium
A.Public cloud
B.Community cloud
C.Hybrid cloud
D.Private cloud
Correct Answer: Community cloud
Explanation:
A community cloud serves multiple organizations that share common security, policy, or compliance requirements.
Incorrect! Try again.
27A workload uses an average of 12 vCPUs. Each host provides 8 vCPUs, and administrators target at most 75% average host utilization. Based only on this average requirement, what is the minimum number of hosts needed?
Guiding principle with respect to utilization
Medium
A.4 hosts
B.3 hosts
C.6 hosts
D.2 hosts
Correct Answer: 2 hosts
Explanation:
Each host should carry vCPUs of average load. Therefore, the requirement is hosts.
Incorrect! Try again.
28A background-processing system receives jobs at unpredictable rates. Which scaling signal would most directly align resource utilization with actual workload?
Guiding principle with respect to utilization
Medium
A.Disk manufacturer
B.Queue length
C.Application version
D.Server purchase date
Correct Answer: Queue length
Explanation:
Queue length reflects pending work, so scaling workers from this metric closely matches capacity to current demand.
Incorrect! Try again.
29An order service is being redesigned so that consumers do not depend on its programming language or internal database schema. Which SOA principle should guide the redesign?
SOA
Medium
A.Shared access to internal tables
B.Loose coupling through contracts
C.Dependence on local libraries
D.Deployment as one executable
Correct Answer: Loose coupling through contracts
Explanation:
A stable service contract hides implementation details and reduces dependencies between service consumers and providers.
Incorrect! Try again.
30A service provider must introduce a breaking change to an API while existing consumers still use the previous message format. What is the most appropriate approach?
SOA
Medium
A.Disable schema validation
B.Version the service contract
C.Expose the internal database
D.Delete the previous endpoint
Correct Answer: Version the service contract
Explanation:
Contract versioning allows older consumers to continue working while new consumers adopt the changed interface.
Incorrect! Try again.
31A research team needs 1,000 compute instances for six hours each month to process independent simulation tasks. Which cloud benefit is most relevant?
Applications of Cloud Computing
Medium
A.Manual server procurement
B.On-demand parallel capacity
C.Permanent hardware ownership
D.Fixed local resource limits
Correct Answer: On-demand parallel capacity
Explanation:
Cloud platforms can provision many instances temporarily, making them suitable for parallel and short-lived batch workloads.
Incorrect! Try again.
32A company stores encrypted backups in a cloud region far from its primary data center and periodically tests restoration there. Which application of cloud computing does this best illustrate?
Applications of Cloud Computing
Medium
A.Local network routing
B.Disaster recovery
C.Source code formatting
D.Desktop publishing
Correct Answer: Disaster recovery
Explanation:
Geographically separate backups and tested restoration procedures support recovery after failure or loss of the primary site.
Incorrect! Try again.
33An organization must let a private-cloud application securely access a managed database in a public cloud without sending traffic over the open internet. Which design is most suitable?
Design and Implementation of Public and Private Cloud Environments
Medium
A.A public endpoint without filtering
B.A shared administrator password
C.A local DNS cache only
D.A private VPN or dedicated link
Correct Answer: A private VPN or dedicated link
Explanation:
A VPN or dedicated connection provides controlled private connectivity between the two cloud environments.
Incorrect! Try again.
34In a multi-tenant private cloud, two departments must use the same physical infrastructure while keeping their network traffic isolated. Which implementation choice best meets this requirement?
Design and Implementation of Public and Private Cloud Environments
Medium
A.A common public subnet
B.Separate tenant virtual networks
C.Identical administrator accounts
D.One shared flat network
Correct Answer: Separate tenant virtual networks
Explanation:
Tenant-specific virtual networks provide logical isolation while allowing departments to share the underlying physical infrastructure.
Incorrect! Try again.
35An OpenStack administrator needs to launch and manage virtual machine instances. Which core service handles these compute operations?
OpenStack
Medium
A.Nova
B.Swift
C.Cinder
D.Neutron
Correct Answer: Nova
Explanation:
Nova is OpenStack's compute service and manages the lifecycle of virtual machine instances.
Incorrect! Try again.
36A virtual machine in OpenStack needs a persistent block volume that remains available after the instance is deleted. Which service should provide it?
OpenStack
Medium
A.Neutron
B.Cinder
C.Glance
D.Keystone
Correct Answer: Cinder
Explanation:
Cinder provides persistent block storage volumes that can exist independently of compute instances.
Incorrect! Try again.
37Users of an OpenStack cloud must authenticate once and receive authorization tokens that other OpenStack services can validate. Which service performs this role?
OpenStack
Medium
A.Horizon
B.Heat
C.Ceilometer
D.Keystone
Correct Answer: Keystone
Explanation:
Keystone supplies centralized identity, authentication, service discovery, and token-based authorization for OpenStack.
Incorrect! Try again.
38A web application currently runs on one EC2 instance in one Availability Zone. Which change most directly improves resilience to an Availability Zone failure?
AWS
Medium
A.Attach a larger volume in the same zone
B.Distribute instances across multiple Availability Zones
C.Assign another IP to the same instance
D.Increase the size of the existing instance
Correct Answer: Distribute instances across multiple Availability Zones
Explanation:
Deploying instances across Availability Zones prevents one zone failure from removing all application capacity.
Incorrect! Try again.
39An application needs highly durable storage for millions of images that are accessed as objects through an API. Which AWS service is the best fit?
AWS
Medium
A.Amazon EC2
B.Amazon EBS
C.Amazon S3
D.Amazon VPC
Correct Answer: Amazon S3
Explanation:
Amazon S3 is durable object storage designed for files such as images, backups, and other unstructured data.
Incorrect! Try again.
40An AWS application receives unpredictable request spikes. Instances should be added automatically when average CPU usage remains high and removed after demand decreases. Which combination best satisfies this requirement?
AWS
Medium
A.AWS IAM with CloudTrail
B.Route 53 with Amazon S3
C.Auto Scaling with CloudWatch
D.Amazon VPC with Direct Connect
Correct Answer: Auto Scaling with CloudWatch
Explanation:
CloudWatch monitors CPU metrics, while EC2 Auto Scaling adjusts the number of instances according to configured scaling policies.
Incorrect! Try again.
41A service experiences a predictable three-minute traffic spike at the start of every hour. New instances require six minutes to become ready, so a reactive CPU-based autoscaler repeatedly violates the latency SLO. Which change best preserves elasticity while addressing the violation?
Cloud Fundamentals
Hard
A.Shorten the monitoring interval while continuing to scale only after saturation
B.Schedule scale-out before each spike and retain reactive scale-in afterward
C.Increase the autoscaler's maximum instance limit without changing its trigger
D.Replace horizontal scaling with permanent vertical scaling of every instance
Correct Answer: Schedule scale-out before each spike and retain reactive scale-in afterward
Explanation:
Because provisioning takes longer than the spike, reactive scaling cannot add capacity in time. Predictive or scheduled scale-out provides capacity beforehand, while reactive scale-in preserves elasticity.
Incorrect! Try again.
42An architect places two instances, each advertised as 99.9% available, in the same availability zone and claims their parallel availability is approximately 99.9999%. What is the primary flaw in this calculation?
Cloud Fundamentals
Hard
A.Instance failures and zone failures are not statistically independent
B.Load balancing reduces the advertised availability of both instances
C.Parallel instances always provide exactly the availability of one instance
D.Availability percentages cannot be combined for virtualized infrastructure
Correct Answer: Instance failures and zone failures are not statistically independent
Explanation:
The calculation assumes independent failures. Instances in one zone share power, networking, and control-plane failure risks, so correlated zone failures invalidate that assumption.
Incorrect! Try again.
43A regulated workload requires a custom kernel module, a host-level forensic agent, and complete control over guest operating-system patch timing. Which service model most directly satisfies these requirements?
Cloud Service
Hard
A.IaaS with customer-managed virtual machines
B.PaaS with a managed language runtime
C.SaaS with tenant-specific application configuration
D.FaaS with a custom deployment package
Correct Answer: IaaS with customer-managed virtual machines
Explanation:
IaaS gives the customer control of the guest operating system and installed agents. PaaS, FaaS, and SaaS normally abstract or restrict kernel-level access.
Incorrect! Try again.
44An enterprise adopts a fully managed SaaS human-resources platform. Under the shared-responsibility model, which responsibility normally remains primarily with the enterprise?
Cloud Service
Hard
A.Replacing failed physical disks and maintaining storage firmware
B.Patching the provider's application binaries and operating systems
C.Securing the provider's data centers and backbone connections
D.Managing employee identities, access roles, and data classification
Correct Answer: Managing employee identities, access roles, and data classification
Explanation:
The SaaS provider manages the underlying application and infrastructure, but the customer remains responsible for its users, access configuration, and appropriate handling of organizational data.
Incorrect! Try again.
45Several independent healthcare institutions jointly fund an infrastructure available only to those institutions. It is governed according to their shared compliance and data-sovereignty requirements. Which deployment model is the best classification?
Deployment Models
Hard
A.A community cloud serving organizations with shared concerns
B.A public cloud with logically isolated subscriptions
C.A private cloud owned by one participating institution
D.A hybrid cloud joining unrelated public cloud providers
Correct Answer: A community cloud serving organizations with shared concerns
Explanation:
A community cloud is dedicated to organizations sharing requirements such as mission, security, policy, or compliance, regardless of whether one member or a third party operates it.
Incorrect! Try again.
46A company retains regulated records in its private cloud but sends tokenized, stateless workloads to a public cloud during demand peaks. An orchestrated workflow moves tasks between the two environments. Which model does this represent?
Deployment Models
Hard
A.A public cloud using dedicated hosts for sensitive workloads
B.A hybrid cloud combining distinct but connected cloud environments
C.A multi-cloud architecture using several public cloud vendors
D.A community cloud shared by departments of one company
Correct Answer: A hybrid cloud combining distinct but connected cloud environments
Explanation:
The architecture binds private and public clouds to support workload portability and cloud bursting. That combination is characteristic of a hybrid cloud.
Incorrect! Try again.
47A service receives 240 requests per second, and each request occupies one processing slot for an average of 250 ms. Using Little's Law, what is the minimum number of slots required to keep average slot utilization at or below 75%, assuming negligible queueing?
Guiding principle with respect to utilization
Hard
A.320 slots
B.80 slots
C.60 slots
D.75 slots
Correct Answer: 80 slots
Explanation:
Little's Law gives occupied slots. To keep utilization at 75%, capacity must be slots.
Incorrect! Try again.
48A private cloud consolidates virtual machines to improve host utilization. Four customer-facing VMs have strongly correlated peak demand, while four batch VMs peak during off-hours. Which placement policy best improves utilization without greatly increasing overload risk?
Guiding principle with respect to utilization
Hard
A.Place all batch VMs together and disable workload migration
B.Assign each virtual machine a separate dedicated physical host
C.Mix customer-facing and batch VMs across the available hosts
D.Place all customer-facing VMs together and reserve one idle host
Correct Answer: Mix customer-facing and batch VMs across the available hosts
Explanation:
Combining workloads with complementary demand patterns reduces aggregate peak variance. Packing correlated workloads together raises the probability that their peaks will exhaust the same host.
Incorrect! Try again.
49An order service invokes payment and inventory services. Payment succeeds, inventory fails, and neither service supports distributed locking. Which SOA design most appropriately restores business consistency?
SOA
Hard
A.Share one database schema between payment and inventory services
B.Retry inventory indefinitely while permanently retaining the payment
C.Require all services to participate in one global database transaction
D.Use a saga that invokes an idempotent payment-compensation operation
Correct Answer: Use a saga that invokes an idempotent payment-compensation operation
Explanation:
A saga coordinates local transactions and compensating actions when a later step fails. Idempotency makes compensation safe when messages or requests are delivered more than once.
Incorrect! Try again.
50A provider must add a field to a widely used service response without forcing simultaneous upgrades of all consumers. Which contract-evolution strategy best minimizes coupling?
SOA
Hard
A.Rename existing fields and publish only the revised contract
B.Add a mandatory field and reject requests from older consumers
C.Change existing field meanings while keeping their original names
D.Add an optional field while preserving existing field semantics
Correct Answer: Add an optional field while preserving existing field semantics
Explanation:
An additive, optional change allows tolerant older consumers to continue operating. Mandatory additions, renaming, or semantic changes can break existing integrations.
Incorrect! Try again.
51A research organization must analyze 2 PB of immutable data already stored in one cloud region. Compute is 20% cheaper in another region, but moving the data incurs high transfer cost and delays. Which design is generally most efficient?
Applications of Cloud Computing
Hard
A.Download the dataset to analyst workstations before processing
B.Copy the dataset continuously between regions for each worker
C.Run elastic compute near the stored data and export only results
D.Move the entire dataset to the cheaper region before every analysis
Correct Answer: Run elastic compute near the stored data and export only results
Explanation:
For data-intensive applications, data gravity often outweighs modest compute-price differences. Moving compute to the data minimizes transfer cost, latency, and duplication.
Incorrect! Try again.
52An event-driven image-processing application receives object-created events with at-least-once delivery. Occasionally, the same event launches multiple functions. Which design prevents duplicate output records without discarding legitimate images?
Applications of Cloud Computing
Hard
A.Disable retries for every function and event-source integration
B.Increase function memory until each invocation completes faster
C.Use the event identifier in a conditional idempotency record
D.Delete each input object immediately after starting processing
Correct Answer: Use the event identifier in a conditional idempotency record
Explanation:
At-least-once delivery permits duplicates. A conditional record keyed by a stable event or object identifier ensures that only the first invocation performs the logical operation.
Incorrect! Try again.
53A private cloud and public cloud both use the CIDR range 10.20.0.0/16. After a VPN is established, workloads cannot reliably route to remote addresses in that range. Which remediation is architecturally most robust?
Design and Implementation of Public and Private Cloud Environments
Hard
A.Disable private DNS resolution on both sides of the connection
B.Renumber one environment to a non-overlapping address range
C.Add identical routes for both networks to every routing table
D.Increase the VPN encryption strength without changing either network
Correct Answer: Renumber one environment to a non-overlapping address range
Explanation:
Overlapping CIDR ranges create route ambiguity. Renumbering provides unambiguous end-to-end routing; translation can be a temporary workaround but adds operational complexity.
Incorrect! Try again.
54A private cloud control plane uses a consensus-based state store. Which controller topology provides the strongest tolerance of one controller or fault-domain failure while avoiding split-brain decisions?
Design and Implementation of Public and Private Cloud Environments
Hard
A.Deploy one controller and restore it from frequent external snapshots
B.Deploy two active controllers together in the same equipment rack
C.Deploy four controllers split evenly across two isolated partitions
D.Deploy three controllers across fault domains using quorum consensus
Correct Answer: Deploy three controllers across fault domains using quorum consensus
Explanation:
Three members provide an odd-sized quorum and can tolerate one member failure. Separating them across fault domains also reduces correlated infrastructure failures.
Incorrect! Try again.
55An OpenStack request authenticates a user, locates an image, schedules a VM, attaches persistent block storage, and connects a tenant network. Which service mapping is correct?
Keystone provides identity, Glance manages images, Nova manages compute, Cinder provides block storage, and Neutron provides networking.
Incorrect! Try again.
56An OpenStack instance has a fixed IP on a private tenant subnet. It must accept inbound traffic from the internet without assigning a public address directly to its virtual interface. What is required?
OpenStack
Hard
A.A Glance proxy, image endpoint, and permissive project quota
B.A Keystone federation rule, service token, and public identity endpoint
C.A Cinder gateway, ephemeral address, and unrestricted volume policy
D.A Neutron external router, floating IP, and permitting security-group rule
Correct Answer: A Neutron external router, floating IP, and permitting security-group rule
Explanation:
A Neutron router connects the tenant subnet to an external network, while a floating IP provides address translation. Security-group rules must also permit the inbound traffic.
Incorrect! Try again.
57A database instance in OpenStack may be deleted and rebuilt, but its data must survive the instance lifecycle and later attach to a replacement instance. Which storage design best meets this requirement?
OpenStack
Hard
A.Store the data on the compute host's ephemeral root disk
B.Store the data inside a customized read-only Glance image
C.Store the data in Nova metadata attached to the instance
D.Store the data in a Cinder volume retained after deletion
Correct Answer: Store the data in a Cinder volume retained after deletion
Explanation:
Cinder volumes have a lifecycle independent of compute instances when deletion-on-termination is disabled. Ephemeral disks and instance metadata do not provide equivalent persistence.
Incorrect! Try again.
58A company serves private Amazon S3 content through CloudFront. Users must not bypass CloudFront by requesting objects directly from S3. Which design best enforces this requirement?
AWS
Hard
A.Make the bucket public and restrict access using object name entropy
B.Use CloudFront Origin Access Control and a restrictive bucket policy
C.Use an internet gateway and restrict the bucket to one AWS Region
D.Enable S3 static website hosting and allow only common HTTP methods
Correct Answer: Use CloudFront Origin Access Control and a restrictive bucket policy
Explanation:
Origin Access Control allows CloudFront to authenticate to S3. The bucket policy can then deny direct public access while permitting requests from the designated distribution.
Incorrect! Try again.
59A transactional application uses Amazon RDS and requires automatic failover after an Availability Zone failure while preserving a single database endpoint. It does not require additional read throughput. Which configuration is most appropriate?
AWS
Hard
A.A manual snapshot copied periodically to another Availability Zone
B.A Multi-AZ deployment with a synchronous standby instance
C.A cross-Region read replica with asynchronous replication
D.A larger single-AZ instance with provisioned storage performance
Correct Answer: A Multi-AZ deployment with a synchronous standby instance
Explanation:
RDS Multi-AZ is designed for high availability and automatic failover through the same database endpoint. Read replicas primarily provide read scaling and generally use asynchronous replication.
Incorrect! Try again.
60Workers consume debit events from an Amazon SQS Standard queue and update account balances in DynamoDB. A worker can finish the debit but fail before deleting the message, causing redelivery. Which design guarantees one logical debit per event?
AWS
Hard
A.Increase the queue visibility timeout beyond the average processing time
B.Configure a dead-letter queue after the first unsuccessful receive
C.Scale the worker fleet down to a single processing instance
D.Use a DynamoDB transaction with an event-ID conditional marker
Correct Answer: Use a DynamoDB transaction with an event-ID conditional marker
Explanation:
SQS Standard provides at-least-once delivery. Atomically writing an event-ID marker and updating the balance makes the debit idempotent, so redelivered events cannot apply it again.
Incorrect! Try again.
Did this save you a night before the exam?
LPU Notes is free, and it stays free. Ads cover part of the server bill.
The rest comes out of a student's own pocket: the domain, the storage,
and keeping the site up through the weeks everyone needs it at once.
The payment button didn't load. An ad blocker or a filtered network is the usual reason.
to try again.
Nothing here is ever locked, and nothing unlocks. Chip in only if it was worth it.
What it pays for →