Introduction
If AWS compute is about running your workloads, AWS storage is about keeping your data. Almost every application on AWS reads and writes data somewhere, which is exactly why storage is one of the most heavily tested areas in the Cloud Technology and Services domain of the CLF-C02 Cloud Practitioner exam — the largest domain at 34%. The good news: at the Cloud Practitioner level you are not expected to configure anything deeply. You are expected to know which storage service fits which situation, and to never confuse object, block, and file storage.
That single distinction — object vs. block vs. file — is the backbone of nearly every storage question on the exam. Get it right and you can reason your way to the correct service even on scenarios you’ve never seen. This guide walks the full CLF-C02 storage landscape the way the exam frames it: the three storage types, Amazon S3 and its storage classes, S3 Glacier for archival, Amazon EBS for block storage, Amazon EFS and FSx for file storage, AWS Storage Gateway for hybrid setups, and the Snow Family for offline data transfer.
This pairs naturally with the AWS Compute Services guide and the broader Core AWS Services overview. If you’re just starting your prep, the AWS Cloud Practitioner exam guide for 2026 sets the full scope.
The Foundation: Object vs. Block vs. File Storage
Before any service names, internalize the three fundamental storage models. This is the concept the exam tests over and over, dressed up in different scenarios.
| Type | What it stores | How you access it | AWS service |
|---|---|---|---|
| Object storage | Whole files as “objects” with metadata, in a flat namespace (buckets) | Over HTTP APIs (GET/PUT) | Amazon S3, S3 Glacier |
| Block storage | Raw disk blocks attached to one instance, like a hard drive | Mounted as a volume, formatted with a filesystem | Amazon EBS, instance store |
| File storage | Files in a shared hierarchical filesystem | Mounted over a network (NFS/SMB) by many clients | Amazon EFS, Amazon FSx |
The plain-English cues:
- Object = “I have files (images, backups, logs, video) and I want to store huge numbers of them cheaply and access them over the web.” → S3.
- Block = “I need a disk for a single server — a boot volume or a database’s data drive.” → EBS, attached to one EC2 instance.
- File = “I need a shared folder that many servers mount at once.” → EFS (Linux/NFS) or FSx (Windows/specialized).
If a question says “shared across multiple instances,” think file. If it says “attached to an instance” or “boot volume,” think block. If it says “store and retrieve any amount of data over the internet,” think object. Those keyword cues resolve most questions instantly.
Amazon S3: Object Storage
Amazon Simple Storage Service (S3) is AWS’s flagship object storage service and probably the single most-tested service on the whole exam. Data lives in buckets (globally unique names) as objects, each up to 5 TB, with essentially unlimited total capacity. You pay for what you store and for requests and data transfer — there is no capacity to provision in advance.
Key facts the CLF-C02 expects you to know:
- Durability: S3 is designed for eleven nines (99.999999999%) of durability. Objects are automatically stored redundantly across multiple facilities within a Region.
- Access: over HTTP(S) APIs, the console, CLI, or SDKs — not by mounting it as a disk.
- Use cases: backups, static website hosting, data lakes, media storage, log archives, software distribution, and storage for analytics.
- Security: private by default; you grant access with bucket policies, IAM, and Block Public Access. Encryption at rest is available (and on by default for new objects). Security specifics are covered in the Security & Compliance guide.
S3 Storage Classes
S3 isn’t one price — it’s a family of storage classes tuned for different access patterns. The exam loves cost-optimization questions where the right answer is “pick the cheaper storage class for infrequently accessed data.”
| Storage class | Best for | Trade-off |
|---|---|---|
| S3 Standard | Frequently accessed data | Highest storage cost, no retrieval fee |
| S3 Intelligent-Tiering | Unknown or changing access patterns | Small monitoring fee; auto-moves data between tiers |
| S3 Standard-IA (Infrequent Access) | Long-lived, infrequently accessed data | Cheaper storage, retrieval fee, still highly available |
| S3 One Zone-IA | Infrequent access, re-creatable data | Stored in a single AZ — cheaper but less resilient |
| S3 Glacier Instant Retrieval | Archives needing millisecond access | Low storage cost, higher retrieval cost |
| S3 Glacier Flexible Retrieval | Archives, minutes-to-hours retrieval | Very cheap storage, slower retrieval |
| S3 Glacier Deep Archive | Long-term retention (7–10+ yrs) | Cheapest storage, retrieval in hours |
Two exam heuristics:
- “Access pattern unknown / unpredictable” → S3 Intelligent-Tiering (it decides for you).
- “Cheapest for rarely-touched long-term archive” → S3 Glacier Deep Archive.
S3 Lifecycle Policies
You don’t move data between classes by hand. Lifecycle policies do it automatically on rules like “transition objects to Standard-IA after 30 days, to Glacier after 90, and delete after 365.” At Cloud Practitioner level you just need to know lifecycle policies exist and are the mechanism for automated, cost-saving tiering.
S3 Glacier: Archival Storage
Amazon S3 Glacier (the three Glacier classes above) is object storage optimized for archival — data you must keep but rarely access, like compliance records, medical images, or old media masters. The trade-off is always the same: the cheaper the storage, the slower and pricier the retrieval. Deep Archive is the cheapest storage AWS offers and is intended for data retrieved maybe once or twice a year, with retrieval measured in hours. When a scenario mentions “regulatory retention for years” and “lowest cost,” Glacier Deep Archive is almost always the intended answer.
Amazon EBS: Block Storage
Amazon Elastic Block Store (EBS) provides block-level storage volumes you attach to EC2 instances — think of it as a virtual hard drive for a server. An EBS volume behaves like a physical disk: you attach it, format it, and mount it.
CLF-C02 essentials:
- An EBS volume is generally attached to one EC2 instance at a time in the same Availability Zone (newer io2 volumes support limited multi-attach, but the exam treats EBS as single-instance).
- Volumes are persistent — data survives an instance stop/start (unlike instance store).
- You can take snapshots (point-in-time backups) stored durably in S3, and restore or copy them across AZs/Regions.
- Volume types trade performance for cost:
| EBS volume type | Backed by | Typical use |
|---|---|---|
| General Purpose SSD (gp3/gp2) | SSD | Boot volumes, most workloads — good balance |
| Provisioned IOPS SSD (io2/io1) | SSD | High-performance databases needing guaranteed IOPS |
| Throughput Optimized HDD (st1) | HDD | Big, sequential workloads (big data, log processing) |
| Cold HDD (sc1) | HDD | Infrequently accessed, lowest-cost HDD |
Contrast EBS with instance store: instance store is ephemeral storage physically attached to the host — very fast, but data is lost when the instance stops or terminates. Use it for scratch/cache data only. “Data must survive an instance stop” → EBS, not instance store, is a classic exam distinction.
Amazon EFS and Amazon FSx: File Storage
When many instances need to read and write the same files simultaneously, block storage won’t do — you need shared file storage.
- Amazon Elastic File System (EFS) is a fully managed, elastic NFS file system for Linux workloads. Multiple EC2 instances (across multiple AZs) can mount it at once. It grows and shrinks automatically — you never provision capacity. Use it for shared content, web-serving farms, and application data shared across a fleet.
- Amazon FSx is a family of managed file systems for specialized needs: FSx for Windows File Server (SMB, for Windows workloads with Active Directory integration) and FSx for Lustre (high-performance computing and machine learning). If a scenario says “Windows file share,” think FSx for Windows; “Linux shared file system,” think EFS.
| Service | Protocol | Platform | Shared across instances? |
|---|---|---|---|
| EBS | Block | Any (single instance) | No |
| EFS | NFS | Linux | Yes |
| FSx for Windows | SMB | Windows | Yes |
AWS Storage Gateway: Hybrid Storage
Not all data lives in the cloud yet. AWS Storage Gateway connects on-premises applications to AWS storage, letting them use cloud storage while keeping a local cache for low-latency access. It’s the exam’s answer to “a company wants to extend its on-prem storage to AWS” or “back up on-premises data to AWS without rewriting applications.” You don’t need the sub-types for CLF-C02 — just recognize Storage Gateway as the hybrid cloud storage bridge.
The AWS Snow Family: Offline Data Transfer
Sometimes you have so much data that sending it over the internet would take weeks or months. The AWS Snow Family solves this with physical devices AWS ships to you: you load your data onto the device and mail it back, and AWS imports it into S3.
- AWS Snowcone — the smallest, most portable device, for edge computing and small transfers.
- AWS Snowball — a rugged appliance for large-scale data transfer (tens of terabytes and up), also with compute options.
- AWS Snowmobile — a literal shipping container hauled by a truck, for exabyte-scale migrations. (Rarely the answer, but memorable.)
The exam cue is unmistakable: “transfer petabytes of data to AWS where network transfer would take too long” → Snowball / Snow Family. If the phrase “limited or no internet bandwidth” appears, think Snow Family too.
Putting It All Together: A Decision Cheat Sheet
The fastest way to answer a storage question is to match the scenario keywords to a service:
| Scenario keyword | Service |
|---|---|
| ”Store and retrieve any amount of data over the web,” “static website,” “data lake” | Amazon S3 |
| ”Access pattern unknown / changing” | S3 Intelligent-Tiering |
| ”Cheapest long-term archive,” “keep for 7 years” | S3 Glacier Deep Archive |
| ”Boot volume,” “disk for a single database server,” “persists after stop” | Amazon EBS |
| ”Fast temporary/scratch storage, ok to lose” | Instance store |
| ”Shared Linux file system for many instances” | Amazon EFS |
| ”Windows file share / SMB” | Amazon FSx for Windows |
| ”Extend on-premises storage to the cloud,” “hybrid backup” | AWS Storage Gateway |
| ”Migrate petabytes offline,” “no bandwidth” | AWS Snow Family |
Memorize this table and you have covered the large majority of CLF-C02 storage questions. The rest is just the object/block/file foundation applied to a new phrasing.
Frequently Asked Questions
What is the difference between object, block, and file storage in AWS?
Object storage (S3) stores whole files as objects in buckets, accessed over HTTP APIs — ideal for huge numbers of files. Block storage (EBS) provides raw disk volumes attached to a single EC2 instance, like a hard drive. File storage (EFS/FSx) offers a shared filesystem many instances can mount at once over NFS or SMB.
Which S3 storage class is cheapest?
S3 Glacier Deep Archive offers the lowest storage cost, intended for long-term archival data retrieved rarely (retrieval takes hours). For unknown access patterns, S3 Intelligent-Tiering optimizes cost automatically without a retrieval penalty on frequently accessed data.
What is the difference between EBS and instance store?
EBS is persistent block storage — data survives an instance stop, start, or termination and can be snapshotted. Instance store is ephemeral storage physically attached to the host; it is very fast but data is lost when the instance stops or terminates. Use instance store only for temporary/scratch data.
When should I use Amazon EFS instead of Amazon EBS?
Use EFS when multiple Linux EC2 instances need to read and write the same files at the same time — a shared file system. Use EBS when a single instance needs a disk (a boot volume or a database’s data volume). EBS attaches to one instance; EFS is shared across many.
What is the AWS Snow Family used for?
The Snow Family (Snowcone, Snowball, Snowmobile) is used for offline, physical data transfer into AWS when transferring over the internet would take too long or bandwidth is limited. AWS ships a device, you load your data, and mail it back for import into S3.
Is Amazon S3 highly durable?
Yes. Amazon S3 is designed for 99.999999999% (eleven nines) of durability, automatically storing objects redundantly across multiple facilities within a Region. This makes it suitable for backups, archives, and any data you can’t afford to lose.
Conclusion
AWS storage looks like a long list of services, but the CLF-C02 exam only asks you to do two things: recognize the three storage models — object, block, file — and match a scenario to the right service. Anchor everything on that foundation. S3 (with its storage classes and Glacier tiers) is object storage for anything you access over the web; EBS is block storage for a single instance; EFS and FSx are shared file storage; Storage Gateway bridges on-premises to the cloud; and the Snow Family moves data physically when the network can’t. Learn the keyword cues in the cheat sheet above and storage questions become some of the most predictable points on the exam.
To make these distinctions automatic under exam pressure, practice with realistic questions that force you to choose between similar services. The AWS Cloud Practitioner Certification-Ready Mock Exam Bundle mirrors the real CLF-C02 domain weightings, so storage questions appear alongside compute, networking, security, and billing exactly as they will on exam day — each with an explanation that reinforces the object/block/file reasoning. Combine it with the AWS Cloud Practitioner study plan and a round of practice questions to confirm the concepts have stuck. From here, revisit the Compute Services guide and the Shared Responsibility Model to round out the Cloud Technology and Services domain.