SAA-C03 - Design Cost-Optimized Architectures - Section 4.2

Optimise the cost of block, file and backup storage by right-sizing EBS volumes and managing snapshots and backups.

Right-size EBS volumes by matching volume type and provisioned capacity to actual workload IOPS and throughput, and use incremental EBS snapshots that charge only for changed blocks. Reduce EFS costs with the EFS Infrequent Access storage class for files not accessed recently, and centralise backups across services with AWS Backup to enforce retention policies and control cost.

EBS volume right-sizingEBS snapshotsAWS BackupEFS Infrequent Access

Practice question for this objective

Free sampleDesign Cost-Optimized Architecturesmedium

A company runs a fleet of EC2 instances each attached to a 1 TiB General Purpose gp2 EBS volume. CloudWatch shows the volumes consistently use under 200 GiB of capacity and never exceed 1,000 IOPS, yet the monthly EBS bill is high because gp2 performance scales with provisioned size. A solutions architect must reduce storage cost while preserving the same usable data and current performance. Which change should the architect make?

  • AMigrate each volume to Provisioned IOPS io2 at the same 1 TiB size and provision 1,000 IOPS, keeping the headroom available for occasional spikes in instance request volume.
  • BTake a snapshot of each volume and restore it to a larger 2 TiB gp2 volume so the higher provisioned size delivers more baseline IOPS for the same workload at no extra charge.
  • CMove the volume data to a Throughput Optimized st1 volume of 256 GiB, because st1 is the cheapest EBS type per GiB and easily covers the modest IOPS the workload currently records each month.
  • DMigrate the volumes to gp3 and right-size them to 256 GiB, since gp3 includes a 3,000 IOPS baseline independent of size and bills capacity separately from performance. Correct
Use gp3 and right-size EBS volumes to cut storage cost, because gp3 decouples a baseline 3,000 IOPS from provisioned capacity. gp2 ties IOPS to provisioned size, so a 1 TiB volume is often paid for just to reach an IOPS target the data does not need. gp3 provides 3,000 IOPS and 125 MiB/s as a free baseline regardless of size and bills capacity, IOPS, and throughput separately. Right-sizing a lightly used volume to 256 GiB on gp3 therefore keeps the same usable data and performance while removing the cost of capacity that exists only to buy IOPS.

Why A is wrong: Tempting because io2 lets you set IOPS directly, but keeping the 1 TiB size and adding provisioned IOPS pricing raises cost rather than reducing it, so the optimisation goal is not met.

Why B is wrong: Tempting because larger gp2 raises baseline IOPS, but doubling the size doubles the storage charge and the workload does not need more IOPS, so this increases cost instead of cutting it.

Why C is wrong: Tempting because st1 has a low per-GiB price, but st1 is an HDD type built for large sequential throughput and performs poorly on the small random IO this workload runs, so it is the wrong fit.

Why D is correct: gp3 decouples IOPS from size with a free 3,000 IOPS baseline, so right-sizing to 256 GiB keeps the data and performance while sharply cutting the per-GiB storage charge.

See more SAA-C03 practice questions, answers explained.

More in this domain

Back to all Design Cost-Optimized Architectures objectives, or the SAA-C03 cert hub.

Examworthy is not affiliated with or endorsed by Amazon Web Services. Original, blueprint-aligned practice material only.