Cloud Storage Tiers Explained: Hot, Cool, Archive and Backup

Cloud Storage Types

Cloud storage is divided into four fundamental types, each optimized for a different access pattern and cost profile:

• Block storage — equivalent to a hard disk attached directly to a server. Used for OS volumes, databases, and applications requiring file-system-level access. Billed per GB provisioned.
• Object storage — scalable storage for unstructured data (files, images, backups, logs). Accessed via HTTP API (S3, Azure Blob, GCS). Billed per GB stored plus per-request charges.
• File storage — shared network file systems (NFS, SMB). Used for shared application data across multiple instances. More expensive than object storage.
• Archive storage — extremely low-cost storage for data accessed rarely or never. Long retrieval times (minutes to hours). Billed per GB stored with retrieval fees.

Block Storage: SSD vs HDD

All major cloud providers offer two primary block storage tiers:

SSD (Premium) block storage — backed by solid-state drives. Provides consistent, low-latency IOPS (Input/Output Operations Per Second) suitable for production databases, transactional applications, and OS boot volumes. AWS gp3 EBS volumes deliver up to 16,000 IOPS baseline; Azure Premium SSD v2 supports up to 80,000 IOPS. Cost: approximately $0.08–0.12/GB/month.

HDD (Standard) block storage — backed by spinning magnetic disks. Much lower IOPS performance (typically 40–500 IOPS) but significantly lower cost. Suitable for log archives, large batch processing workloads, or data warehousing inputs. Cost: approximately $0.04–0.05/GB/month — roughly 50% cheaper than SSD for the same capacity.

Object Storage Tiers

Object storage (S3, Azure Blob, GCS) is the most commonly used cloud storage type by volume. All three providers offer multiple access tiers within their object storage services:

Frequent Access (Hot) tier — lowest latency, highest availability (99.99%+), highest per-GB cost ($0.018–0.023/GB/month). Best for actively read data: website assets, application files, current backups, and data accessed multiple times per day.

Infrequent Access (Cool) tier — lower per-GB storage cost ($0.01–0.015/GB/month) but charges a per-GB retrieval fee ($0.01/GB). Data must not be accessed for at least 30 days without incurring early deletion fees. Best for backups, disaster recovery data, and assets accessed a few times per month.

Intelligent Tiering — AWS S3 Intelligent-Tiering automatically moves objects between frequent and infrequent access tiers based on actual access patterns, charging a small monthly monitoring fee ($0.0025/1,000 objects). This is recommended for data with unknown or changing access patterns.

Archive and Cold Storage

Archive tiers offer dramatic cost reductions for data that is rarely or never accessed but must be retained for compliance, legal, or historical reasons:

Backup and Snapshots

Block storage snapshots (point-in-time copies of EBS/managed disk volumes) are stored as incremental backups in object storage, billed at standard object storage rates. The first snapshot is a full copy; subsequent snapshots capture only changed blocks, making the ongoing cost much lower than the full volume size suggests.

AWS Backup, Azure Backup, and GCP Backup and DR provide managed backup policies across multiple services. Pricing is typically $0.05–0.10/GB/month for backup storage, plus data transfer costs for cross-region replication. For compliance-driven retention requirements (7-year financial records, HIPAA 6-year medical records), archive storage tiers are strongly recommended over standard backup storage to minimize long-term retention costs.

Storage Cost Optimization Strategies

• Enable lifecycle policies: Automatically transition objects to cheaper storage tiers after a defined period. A common pattern: Standard → Infrequent Access after 30 days → Glacier after 90 days → Glacier Deep Archive after 365 days. This can reduce long-term storage costs by 80–95%.
• Delete unused volumes: EBS volumes that are not attached to running instances still accrue charges. A regular audit of unattached volumes is essential hygiene for cost management.
• Compress before archiving: Data compression before archiving to Glacier/Azure Archive can reduce storage volumes by 40–70% for text-heavy data types (logs, documents, backups).
• Right-size SSD provisioning: EBS gp3 volumes allow you to provision IOPS and throughput independently of capacity. Many workloads over-provision volume size to get adequate IOPS — gp3 eliminates this by allowing 3,000 IOPS at any size.

Complete Storage Pricing Comparison 2026

Storage pricing varies significantly across tiers and providers. The key insight most architects miss is that the cheapest storage tier per GB is not necessarily the cheapest option — retrieval fees, minimum storage durations, and access frequency all affect the true cost. The table below shows the headline per-GB storage cost plus the critical secondary costs that determine total expense.

Automated Lifecycle Policies: The Highest-ROI Storage Optimization

Lifecycle policies automatically transition objects from more expensive to cheaper storage tiers based on age, and can delete objects that are no longer needed. For most organizations, implementing lifecycle policies is the single highest-ROI storage optimization — requiring a few hours of configuration work and delivering thousands of dollars per month in savings.

A typical pattern for application log data: store in Standard for 30 days (active analysis), transition to Standard-IA for days 31–90 (occasional reference), transition to Glacier Instant Retrieval for days 91–365 (compliance retention, rare access), and delete after 12 months. For a workload generating 1 TB of logs per month, this tiered approach costs approximately $23 in month 1, dropping to $4 as data ages into Glacier — versus $276/year if all data remains in Standard storage.

AWS S3 lifecycle rules are configured via the S3 console, CLI, or CloudFormation. Azure Blob lifecycle management policies work identically. GCP Object Lifecycle Management policies use condition-based rules. All three providers support transitioning by object age, last access time, and object size.

Request Costs: The Hidden Storage Expense

Storage pricing has two components most architects overlook: per-request costs and data retrieval costs. AWS S3 charges $0.005 per 1,000 PUT/COPY/POST/LIST requests and $0.0004 per 1,000 GET/SELECT requests. For an application that makes 100 million GET requests to S3 per month, request costs alone reach $40/month — comparable to the storage cost itself for smaller datasets.

For archive tiers, retrieval costs can dramatically exceed storage costs for data accessed more frequently than anticipated. S3 Glacier Flexible Retrieval charges $0.0025/GB for expedited retrieval (1–5 minutes) plus $0.03 per 1,000 requests. If a 10 TB archive is fully retrieved in an emergency, expedited retrieval costs $25 + request fees — negligible. But if large portions of a Glacier archive are being accessed monthly because someone moved data there prematurely, retrieval costs quickly exceed the savings from cheaper storage.