Este módulo transforma as questões em uma revisão orientada de fundamentos de storage em cenários práticos. Antes de responder, identifique o requisito principal do cenário, diferencie o modelo de acesso aos dados e elimine alternativas que resolvem outro problema técnico.
Questões neste módulo
10 questões ou exercícios, com correção e explicações mantidas no próprio simulado.
Como estudar
Responda sem consultar o gabarito. Depois, revise a justificativa e registre o conceito que levou à resposta correta.
Objetivo
Compreender o motivo técnico de cada escolha, em vez de apenas memorizar letras ou frases.
O que você revisará
SAN, NAS e DAS
Block, file e object storage
Flash, latência e IOPS
Thin provisioning e redução de dados
Escolha da arquitetura por workload
Estratégia para questões de cenário
Procure palavras que indiquem acesso em bloco, arquivo ou objeto; necessidade de baixa latência, grande throughput, compartilhamento, retenção, disponibilidade ou recuperação. Em seguida, escolha a tecnologia que atende diretamente ao requisito, sem adicionar complexidade desnecessária.
Material independente: conteúdo organizado por Jociano Baum para estudo e revisão. Não é material oficial de prova e não representa garantia de aprovação. As situações são educacionais e genéricas, sem referência a empresas ou ambientes reais.
Observações de autoria — Jociano Baum:
conteúdo organizado e complementado com base em materiais de estudo e revisão para Data Storage Associate (DSA).
Material destinado ao estudo independente e não apresentado como conteúdo oficial de certificação.
jociano.com.br
A virtualization team is replacing a legacy HDD-based storage array with an all-flash platform. The network bandwidth will remain the same, but the architect expects application response time to improve significantly. Which characteristic BEST explains the expected improvement?
A) Flash storage increases Ethernet bandwidth
B) Flash storage eliminates mechanical seek and rotational latency
C) Flash storage removes the need for multipathing
D) Flash storage converts file workloads into block workloads
✅ Correct: B) Flash storage eliminates mechanical seek and rotational latency
Why? HDDs depend on moving parts, seek time and rotational delay. Flash has no mechanical movement, so latency is much lower and IOPS are higher.
Why not A? Flash improves storage latency, not Ethernet bandwidth.
Why not C? Multipathing is still important for availability and path redundancy.
Why not D? Storage media does not automatically change the access protocol.
Exam Tip: If the scenario compares HDD vs flash and mentions response time, think latency reduction from no moving parts.
A company is migrating 1,000 virtual desktops to a new storage platform. Many desktops are based on the same golden image, and most users consume only a fraction of their assigned capacity. Which combination BEST reduces physical storage consumption?
A) RAID 6 and snapshots
B) Thin provisioning and inline data reduction
C) Erasure coding and synchronous replication
D) Full provisioning and compression only
✅ Correct: B) Thin provisioning and inline data reduction
Why? Thin provisioning delays physical allocation, while inline data reduction reduces duplicate and compressible data. VDI/golden image environments are strong deduplication candidates.
Why not A? RAID and snapshots help protection/recovery, not primary capacity reduction for duplicate VM data.
Why not C? Erasure coding protects data; synchronous replication protects availability/DR, not duplicate OS blocks.
Why not D? Full provisioning wastes capacity, and compression alone misses duplicate blocks.
Exam Tip: VDI + golden image + capacity efficiency = deduplication/data reduction plus thin provisioning.
A backup administrator is moving very large sequential backup files from production storage to a repository. The main concern is how much data can be transferred per second, not the number of small operations. Which metric is MOST relevant?
A) IOPS
B) Throughput
C) Queue depth
D) RPO
✅ Correct: B) Throughput
Why? Throughput measures the amount of data transferred over time, which matters for large sequential backup movement.
Why not A? IOPS matters more for many small operations, such as transactional databases.
Why not C? Queue depth affects outstanding I/O, but it is not the metric described.
Why not D? RPO is a recovery objective, not a transfer performance metric.
Exam Tip: Large sequential data movement = throughput. Small random transactions = IOPS and latency.
A database host is allowed to submit only a very small number of outstanding I/O operations to storage. During peak activity, the application cannot keep enough requests in flight to fully use the storage system. Which setting is MOST directly related to this behavior?
A) Queue depth
B) Snapshot retention
C) Object metadata
D) RAID stripe width
✅ Correct: A) Queue depth
Why? Queue depth controls how many I/O operations can be outstanding at once. Too low a value can limit performance under load.
Why not B? Snapshot retention controls recovery point duration, not host I/O concurrency.
Why not C? Object metadata describes objects, not block I/O queueing.
Why not D? RAID stripe width may affect layout, but the scenario specifically describes outstanding I/O count.
Exam Tip: “Outstanding I/O operations” is the clue for queue depth.
An application team wants to modernize storage access for latency-sensitive workloads. They are evaluating a technology designed for non-volatile memory and high parallelism, with an option to extend access across a network fabric. Which technology family BEST matches this description?
A) SMB and NFS
B) NVMe and NVMe over Fabrics
C) SATA and tape
D) Object storage and buckets
✅ Correct: B) NVMe and NVMe over Fabrics
Why? NVMe is designed for fast non-volatile storage with high parallelism. NVMe-oF extends NVMe access across supported fabrics.
Why not A? SMB and NFS are file protocols, not NVMe-based block access technologies.
Why not C? SATA and tape do not match the high-parallelism NVMe description.
Why not D? Object storage is API/object-based, not NVMe block access.
A storage design discussion compares two data protection methods. One is commonly associated with traditional disk groups and parity/mirroring, while the other splits data into fragments with parity-like information for capacity-efficient protection at scale. Which pairing is MOST accurate?
A) RAID and erasure coding
B) Compression and deduplication
C) Thin provisioning and snapshots
D) SAN and NAS
✅ Correct: A) RAID and erasure coding
Why? RAID uses mirroring/parity across disks. Erasure coding breaks data into fragments with protection information for resilient, capacity-efficient storage.
Why not B? Compression and deduplication are capacity efficiency techniques, not protection schemes in this context.
Why not C? Thin provisioning allocates capacity on demand; snapshots are point-in-time recovery copies.
Why not D? SAN and NAS are storage architectures/access models, not data protection schemes.
Exam Tip: Protection scheme comparison often points to RAID vs erasure coding.
Q17 • Difficulty: ★★★★ • Blueprint: Storage Fundamentals • Concepts: Block vs File vs Object
A customer asks why an object storage platform is not the best target for a latency-sensitive transactional database, even though it is highly scalable and durable. Which explanation is BEST?
A) Object storage is generally optimized for API-based unstructured data, not low-latency block I/O
B) Object storage cannot store large files
C) Object storage requires Fibre Channel zoning
D) Object storage is always volatile
✅ Correct: A) Object storage is generally optimized for API-based unstructured data, not low-latency block I/O
Why? Databases usually need low-latency block storage. Object storage is excellent for scalable unstructured data but is not the typical fit for transactional database volumes.
Why not B? Object storage can store large objects/files.
Why not C? Object storage does not require Fibre Channel zoning.
Why not D? Object storage is persistent, not volatile.
Exam Tip: Scalable and durable does not always mean best for databases; check access pattern and latency.
A storage engineer is explaining why two reduction technologies behave differently. One reduces the size of individual data patterns, while the other avoids storing identical blocks more than once. Which statement is MOST accurate?
A) Compression reduces duplicate blocks; deduplication reduces file permissions
B) Deduplication reduces repeated blocks; compression reduces the size of data representations
C) Compression and deduplication are both replication technologies
D) Deduplication is only used for encryption
✅ Correct: B) Deduplication reduces repeated blocks; compression reduces the size of data representations
Why? Deduplication identifies duplicate data and stores it once. Compression encodes data more efficiently to reduce its size.
Why not A? It reverses and confuses the concepts.
Why not C? They are data reduction technologies, not replication technologies.
Why not D? Deduplication is unrelated to encryption.
Exam Tip: Duplicate data = dedupe. Smaller representation = compression.
A customer asks what role storage controllers play in an enterprise array. Which answer is BEST for a DSA-level explanation?
A) They process storage I/O and provide storage services such as access, protection and management
B) They only provide DNS name resolution
C) They are used only for cloud object buckets
D) They replace host operating systems
✅ Correct: A) They process storage I/O and provide storage services such as access, protection and management
Why? Controllers manage I/O between hosts and storage media and provide array services.
Why not B? DNS is unrelated to the primary role of a storage controller.
Why not C? Controllers are not limited to object storage.
Why not D? Hosts still run their own operating systems.
An enterprise wants to maximize controller utilization and avoid having one controller idle during normal operations. Which architecture BEST aligns with that goal?
A) Active-passive controller architecture
B) Active-active controller architecture
C) Direct attached storage
D) Offline backup architecture
✅ Correct: B) Active-active controller architecture
Why? In active-active designs, multiple controllers can actively process I/O during normal operation, improving utilization and availability.
Why not A? Active-passive usually has one active controller and one standby.
Why not C? DAS describes attachment topology, not controller activity.
Why not D? Offline backup is unrelated to controller architecture.
Exam Tip: Both controllers serving I/O = active-active. One active, one standby = active-passive.
Ao concluir este bloco, revise os conceitos abaixo e retorne às questões em que houve dúvida. Uma resposta correta por eliminação só se transforma em aprendizado quando você consegue explicar o requisito e a tecnologia escolhida.
SAN, NAS e DASBlock, file e object storageFlash, latência e IOPSThin provisioning e redução de dadosEscolha da arquitetura por workload
Erro comum
Escolher a tecnologia mais avançada, mesmo quando o cenário pede simplicidade, custo menor ou apenas um tipo específico de acesso.
Próximo passo
Compare este módulo com outro domínio do DSA e observe como storage, rede, proteção, virtualização e cloud se relacionam.
Perguntas frequentes
Como revisar fundamentos de storage em cenários práticos?
Refaça as questões erradas, escreva em uma frase o requisito principal de cada cenário e compare as alternativas pelo modelo de acesso, desempenho, disponibilidade ou proteção exigidos.
É suficiente memorizar o gabarito?
Não. A utilidade do simulado está em entender por que uma alternativa atende ao cenário e por que as demais resolvem problemas diferentes.
Este conteúdo é oficial?
Não. É um material independente de estudo, criado para revisão conceitual e prática.