Sign in to practice all 132 questions with progress tracking and AI explanations.
1. A database workload requires consistent, high IOPS performance that does not burst. Which EBS volume type is most appropriate?
- A.io1/io2 Provisioned IOPS SSD volumes deliver consistent, user-specified IOPS up to 64,000 per volume, ideal for latency-sensitive databases.✓ Correct
- B.gp3 General Purpose SSD volumes provide a baseline of 3,000 IOPS and can be configured up to 16,000 IOPS, but performance may vary under heavy sustained load.
- C.st1 Throughput Optimized HDD volumes are designed for sequential big-data workloads and do not provide the high, consistent IOPS needed for databases.
- D.sc1 Cold HDD volumes offer the lowest cost but deliver very low IOPS, making them unsuitable for any performance-sensitive database workloads.
Explanation
io1/io2 Provisioned IOPS volumes are purpose-built for sustained, high-IOPS database workloads, guaranteeing the configured IOPS level. gp3 can reach 16,000 IOPS but is a general-purpose tier, while st1 and sc1 are HDD-based with far lower IOPS ceilings.
2. A media company needs cost-effective storage for large sequential read workloads such as video processing pipelines. Which EBS volume type should they choose?
- A.sc1 Cold HDD volumes are optimized for infrequently accessed data and deliver the lowest throughput, making them a poor fit for active video processing.
- B.io2 Provisioned IOPS SSD volumes maximize IOPS for random I/O but are significantly more expensive and unnecessary for sequential streaming workloads.
- C.st1 Throughput Optimized HDD volumes are designed for frequently accessed, sequential workloads like big data and media processing, delivering up to 500 MB/s throughput.✓ Correct
- D.gp2 General Purpose SSD volumes burst to 3,000 IOPS but their throughput is capped at 250 MB/s, which may bottleneck high-throughput sequential video workloads.
Explanation
st1 is specifically designed for large, sequential, high-throughput workloads like video processing, delivering up to 500 MB/s at a fraction of SSD costs. sc1 is for cold/infrequent access, and io2 is unnecessarily expensive for sequential I/O patterns.
3. An application writes 256 KB random I/O operations and requires 10,000 IOPS. What is the required EBS throughput in MB/s?
- A.Throughput = IOPS × I/O size = 10,000 × 256 KB = 2,500 MB/s, which exceeds EBS limits and would require multiple volumes in a RAID configuration.
- B.Throughput = IOPS × I/O size = 10,000 × 256 KB = 2,500 MB/s, which is achievable with a single io2 Block Express volume at maximum configuration.
- C.Throughput = IOPS × I/O size = 10,000 × 256 KB = 2,560,000 KB/s = 2,500 MB/s, but EBS normalizes all I/O to 16 KB chunks, so actual IOPS consumed is 160,000.
- D.Throughput = IOPS × I/O size = 10,000 × 0.25 MB = 2,500 MB/s; however, EBS throughput limits mean you need to provision multiple volumes striped with RAID 0.✓ Correct
Explanation
Throughput = IOPS × I/O block size; 10,000 × 256 KB = 2,500 MB/s. A single EBS volume cannot deliver this, so RAID 0 striping across multiple volumes is required to aggregate throughput. EBS IOPS are normalized to 16 KB units, further increasing logical IOPS consumed.
4. A gp2 EBS volume is 500 GB. What is its baseline IOPS performance, and when does burst apply?
- A.gp2 volumes deliver a flat 3,000 IOPS for all sizes regardless of provisioned capacity, so a 500 GB volume always performs at 3,000 IOPS without bursting.
- B.gp2 volumes provide 3 IOPS per GB of provisioned storage, giving 1,500 IOPS baseline for a 500 GB volume; volumes under 1,000 GB can burst to 3,000 IOPS using I/O credits.✓ Correct
- C.gp2 volumes provide 5 IOPS per GB baseline, so a 500 GB volume delivers 2,500 IOPS continuously and can burst to 10,000 IOPS if I/O credit bucket is full.
- D.gp2 volumes are credit-based and always burst; baseline IOPS do not apply unless the volume has been running for more than 24 hours and credits are exhausted.
Explanation
gp2 provides 3 IOPS per GB baseline (1,500 IOPS for 500 GB) and can burst to 3,000 IOPS using an I/O credit bucket. Volumes ≥1,000 GB always have 3,000+ IOPS baseline and never burst. gp3 is the newer alternative with a flat 3,000 IOPS baseline regardless of size.
5. An application requires maximum disk I/O throughput and can tolerate data loss if a single EBS volume fails. Which RAID configuration should be used?
- A.RAID 1 mirrors data across two EBS volumes, providing redundancy and protecting against single-volume failure, but does not increase throughput beyond one volume.
- B.RAID 5 stripes data with parity across three or more volumes for both performance and redundancy; AWS recommends this for critical database workloads on EBS.
- C.RAID 0 stripes data across multiple EBS volumes with no redundancy, aggregating throughput and IOPS; data is lost if any single volume fails.✓ Correct
- D.RAID 6 provides dual parity protection with moderate performance gains; it is recommended when both high throughput and fault tolerance are required on EBS.
Explanation
RAID 0 stripes I/O across volumes, multiplying throughput and IOPS, but any single volume failure loses all data. AWS recommends against RAID 5/6 on EBS because parity operations consume IOPS and the parity advantage is already provided by EBS's internal replication.
Practice all 132+ questions in this domain
Start free practice →