3.7 KiB
title, type, source-type, category, tags, date-added, video-source, audio-source, status
| title | type | source-type | category | tags | date-added | video-source | audio-source | status | |||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Public Cloud Learning Sessions-Storage Cost Optimization - 20240305 160037-Meeting Recording | cloud-learning | video | DevOps & SRE/05_FinOps |
|
2026-04-14 | nas:///volume2/work/Public Cloud Learning Sessions/Public Cloud Learning Sessions-Storage Cost Optimization - 20240305_160037-Meeting Recording.mp4 | summarized (Gemini 摘要) |
Public Cloud Learning Sessions-Storage Cost Optimization - 20240305 160037-Meeting Recording
Source: NAS /volume2/work/Public Cloud Learning Sessions/Public Cloud Learning Sessions-Storage Cost Optimization - 20240305_160037-Meeting Recording.mp4
Type: VIDEO | Category: 05_FinOps
Status: 🟡 Awaiting Whisper transcription → Summary
Storage Cost Optimization
This session covers storage cost optimization best practices across various AWS storage services: Amazon EBS, Amazon EFS, Amazon FSx, and Amazon S3. It includes an optimization example from ADM.
Key points include choosing the right storage for your workload, considering API costs and data transfer costs in addition to price per gigabyte, and understanding the different tiers available within each service.
Amazon EBS
EBS has SSD and HDD volumes. GP3 volumes are recommended as the default for general-purpose SSD due to being 20% more cost-effective than GP2. With GP3, you can scale IOPS and throughput independently of the volume size. For migration from GP2 to GP3, automation tools should be updated to create GP3 volumes by default. EBS snapshots have standard and archive tiers, with the archive tier offering 75% lower costs but higher restore times and a 90-day retention period. Automation via Data Lifecycle Management (DLM) or AWS Backup is recommended for managing snapshots, including setting retention policies and migrating to the archive tier.
Amazon EFS and FSx
FSx considerations include data deduplication, compression, and tiering. EFS offers standard, one-zone, and infrequent access tiers, with lifecycle policies to move files between tiers. The infrequent tier has a minimum billable object size of 128KB. EFS archive is a new tier, similar to Glacier, with a 90-day minimum duration and a 128KB minimum billable object size. FSx for NetApp ONTAP has SSD and HDD tiers (capacity pool), with automatic tiering between them.
Amazon S3
Choosing the right storage class is crucial for S3 cost optimization. S3 Standard is for frequently accessed objects, with no retrieval fees, minimum retention, or minimum billable object size. Glacier tiers (Instant Retrieval, Flexible Retrieval, Deep Archive) are for rarely accessed data, with varying retrieval times and costs. Intelligent Tiering automatically moves data between tiers based on access patterns, with no transition fees between tiers within Intelligent Tiering. With intelligent hearing we can automatically move data from warmer to colder color storage tiers and it will be based on the object less access data. Lifecycle policies can transition objects between tiers, expire non-current versions, and delete incomplete multi-part uploads. Data transfer charges should be considered, and PrivateLink can be leveraged to stay within the AWS network. Storage Lens, CloudWatch, S3 Inventory, and access logs can be used to monitor and optimize S3 usage.
ADM Optimization Example
ADM migrated NetApp file shares from on-premises to AWS. The initial migration to OpenZFS was inefficient. A second migration to a self-managed NetApp on EC2 instances incurred high data transfer costs. The final migration to AWS FSx for NetApp ONTAP resulted in a 60% cost reduction.