Direct answer: at official list prices on 8 November 2026 (USD, US East, excluding tax and discounts), cloud DR has two cost models: AWS Elastic Disaster Recovery (DRS) charges $0.028 per source server per hour, plus staging EBS, snapshots and replication EC2; Azure Site Recovery (ASR) charges $25 per protected instance per month (first 31 days free), plus replica storage, cross-region egress and test/failover compute; GCP Backup and DR charges per GiB for management and long-term storage, not per server. For 100 servers, 30 TB, 7-day retention and 3.3% daily change: the AWS DRS official example totals about $6,389/month (including EBS/snapshots/replication instances), Azure ASR about $4,036/month ($25 × 100 + 30 TB standard HDD replica), and GCP Backup and DR about $1,690/month (management + long-term storage at 30 TB). But their RPO/RTO differ: DRS/ASR are continuous replication; GCP is more backup/DR-oriented—do not compare monthly fees alone.
This guide is for cloud platform, SRE, disaster recovery, IT operations and procurement teams. It compares AWS DRS, Azure Site Recovery and GCP Backup and DR: billing units, 100-server/30 TB TCO, 10/50/100-server scale, hidden costs, RPO/RTO selection and a 90-day checklist. Prices come from the AWS DRS pricing page, Azure Site Recovery pricing page and Azure Retail Prices API, and the Google Cloud Backup and DR pricing page (all accessed 8 November 2026). Official price facts, editorial calculations and exclusions are kept separate. This is not procurement advice.
(1) Separate the models first: DRS/ASR are per-server continuous replication; GCP Backup and DR is per-GiB backup/DR—monthly fees are not directly comparable. (2) The monthly fee is only the base: staging storage, snapshot retention, cross-region egress, drill/failover compute, support and licenses all add. (3) Set RPO/RTO first: minutes-level RPO needs continuous replication; hours-level/backup-oriented can use cheaper per-GiB options.
Billing models: per server vs per GiB
AWS DRS charges $0.028 per hour for each actively replicating source server; the fee does not depend on disk count/capacity/drill count/target region, but staging EBS, EBS snapshots and replication EC2 instances are extra. Azure ASR charges $25 per protected instance per month (replication to Azure), with the first 31 days free; replica storage, cross-region egress and test/failover compute are extra. GCP Backup and DR charges a management fee per GiB of protected source capacity (about $0.01/GiB-month) plus long-term storage (about $0.045/GiB-month); recovery compute/network are extra. The units differ, so normalize to the same workload before comparing.
| Item | AWS DRS | Azure ASR | GCP Backup and DR |
|---|---|---|---|
| Billing unit | $0.028 per source server-hour | $25 per protected instance-month (first 31 days free) | Per-GiB management + long-term storage |
| Model | Continuous replication | Continuous replication | Backup/DR (RPO depends on policy) |
| Extra resources | Staging EBS, snapshots, replication EC2 | Replica storage, egress, drill/failover compute | Storage, recovery compute/network |
| Fit | On-prem/other cloud to AWS, AWS cross-region | On-prem/other cloud to Azure, Azure cross-region | GCP resource protection, backup/DR |
Sources: AWS DRS pricing page, Azure Site Recovery pricing page and Azure Retail Prices API (eastus, 8 November 2026), GCP Backup and DR pricing page (us-central1, 8 November 2026). Editorial summary; availability/regions/support follow the official pages.
100-server/30 TB TCO: official example and editorial estimate
The AWS official example for 100 servers, 200 disks, 30 TB, 7-day retention and 3.3% daily change: DRS $2,044 + EBS staging $1,425 + snapshots $1,846.50 + replication EC2 $1,053.54 = $6,389.03/month (excluding drill/recovery instances). Azure ASR editorial estimate: 100 × $25 = $2,500 + 30 TB standard HDD replica about $1,536 = $4,036/month (excluding egress/drill compute). GCP editorial estimate: 30 TB management about $307 + long-term storage about $1,382 = $1,690/month (per-GiB estimate; different RPO model).
| Option (100 servers/30 TB) | Components | Monthly cost (USD) |
|---|---|---|
| AWS DRS | $2,044 replication + $1,425 EBS + $1,846.50 snapshots + $1,053.54 EC2 | 6,389.03 |
| Azure ASR | $2,500 instances + $1,536 replica storage | 4,036.00 |
| GCP Backup and DR | $307 management + $1,382 long-term storage | 1,689.60 |
AWS is the official Pricing example #1 (US East, 100 servers/30 TB/7 days/3.3% daily change, excluding drill compute); Azure/GCP are editorial estimates (Azure $25/instance-month + standard HDD $0.05/GiB-month; GCP management 0.000013699/GiB-h × 730 ≈ 0.01, long-term storage 0.000061644/GiB-h × 730 ≈ 0.045). Excludes egress, drill/recovery compute, tax and discounts; GCP's RPO/RTO model differs from DRS/ASR.
Scale: 10 / 50 / 100 servers
Assuming 300 GB per server and total capacity scaling linearly: AWS DRS about $63.89/server-month, Azure ASR about $40.36/server-month, GCP about $16.90/server-month. 10 servers about $639/$404/$169; 50 about $3,195/$2,018/$845; 100 about $6,389/$4,036/$1,690. The per-GiB option looks cheaper at scale, but RPO/RTO, recovery orchestration, application consistency, cross-region networking and support also scale.
| Scale (300 GB/server) | AWS DRS | Azure ASR | GCP Backup and DR |
|---|---|---|---|
| 10 servers / 3 TB | about 639 | about 404 | about 169 |
| 50 servers / 15 TB | about 3,195 | about 2,018 | about 845 |
| 100 servers / 30 TB | about 6,389 | about 4,036 | about 1,690 |
Editorial calculation: linear scaling from the 100-server example; AWS uses the official example ratio, Azure/GCP use per-instance/per-GiB rates. Excludes minimums, regional differences, egress, drill/recovery compute, tax and discounts; the bill is authoritative.
Hidden costs: six items beyond the monthly fee
The DR bill often misses: staging/replica storage (full source disk, not just changes), snapshot/recovery-point retention (longer retention costs more), cross-region egress/replication traffic, drill/failover compute (recovery instances billed hourly), support/licenses (Business/Enterprise Support, OS/database licenses) and operations/tools. In the AWS example, EBS + snapshots + replication EC2 are about $4,325, more than the $2,044 DRS fee; Azure replica storage and egress also grow with capacity/regions.
| Cost item | Trigger | Control |
|---|---|---|
| Staging/replica storage | Full source disk resident | Remove unused disks, tier storage, replicate only needed systems |
| Snapshots/recovery points | Retention days/frequency | Set retention by RPO, clean up regularly |
| Cross-region egress | Cross-region replication | Compression, nearest region, budget alerts |
| Drill/failover compute | Launch recovery instances | Limit duration/size; release after drills |
| Support/licenses | Production DR, OS/DB | Check contract/license terms |
| Operations/tools | Orchestration, testing, on-call | Automation, drill schedule |
Editorial summary; specific billing/licensing follows AWS/Azure/GCP pages and contracts.
How to choose: RPO/RTO, platform and cost
Set RPO/RTO first: minutes-level RPO and fast failover need DRS/ASR; hours-level RPO and backup/restore may use GCP Backup and DR or object storage/snapshots. Then platform: on-prem VMware/physical to AWS → DRS; to Azure → ASR; resources already on GCP → Backup and DR; cross-cloud DR also needs egress/network costs. Finally operations: continuous replication needs staging/replication instances/drill orchestration; per-GiB backup needs recovery orchestration/application consistency—neither is "buy and forget".
| Scenario | Prefer | Check first |
|---|---|---|
| Minutes-level RPO/fast failover | AWS DRS / Azure ASR | RTO, replication lag, drills |
| Hours-level RPO/backup-oriented | GCP Backup and DR / snapshots + object storage | Restore time, retention, app consistency |
| On-prem VMware/physical | Target-cloud DR service | Agent/replicator, network, licenses |
| In-cloud cross-region | DRS/ASR/native replication | Egress, regional limits |
| Cost-sensitive/many cold systems | Backup/cold DR | Whether RTO is acceptable |
| Compliance/critical business | Continuous replication + regular drills | Evidence, audit, support |
Editorial selection framework, not a vendor recommendation; RPO/RTO/compliance follow business and contracts.
90-day readiness checklist
Use the first 90 days to close the "tiering, RPO/RTO, option, TCO, drill, monitor" loop. Weeks 1-2 inventory systems/dependencies/data; weeks 2-4 set RPO/RTO/tiers; weeks 4-6 choose continuous replication vs backup and calculate TCO; weeks 6-8 pilot replication/recovery drill; weeks 8-10 roll out/retention; weeks 10-12 monitor/cost review/audit evidence. Do not buy a DR license and skip drills.
| Phase | Action | Output |
|---|---|---|
| Weeks 1-2 | Inventory systems/dependencies/data/existing backups | DR list, dependency map |
| Weeks 2-4 | Set RPO/RTO/tiers by business | DR tiers, SLA |
| Weeks 4-6 | Choose continuous vs backup; calculate TCO/hidden costs | Option, budget |
| Weeks 6-8 | Pilot replication/recovery drill; measure RTO | Drill report, issue list |
| Weeks 8-10 | Roll out; set retention/egress/compute policies | Runbook, retention policy |
| Weeks 10-12 | Monitor/cost review/audit evidence | Dashboard, review |
Editorial readiness advice, not an official template; RPO/RTO/compliance/budget follow enterprise standards and vendor contracts.
FAQ
How is AWS DRS billed?
$0.028 per actively replicating source server per hour; staging EBS, EBS snapshots and replication EC2 are extra. The official 100-server/30 TB/7-day example totals about $6,389/month, excluding drill/recovery instances.
How is Azure Site Recovery billed?
$25 per protected instance per month (replication to Azure), first 31 days free; replica storage, cross-region egress and test/failover compute are extra. 100 servers are about $2,500/month plus about $1,536 for 30 TB standard HDD replica.
How is GCP Backup and DR billed?
Per-GiB management fee (about $0.01/GiB-month) + long-term storage (about $0.045/GiB-month); recovery compute/network extra; not per server. 30 TB is about $1,690/month (editorial estimate).
Why is the monthly fee so different?
Billing units and RPO/RTO differ: DRS/ASR are continuous replication per server; GCP is backup/DR per GiB. Continuous replication's staging/snapshot/replication-instance costs are higher; backup is cheaper but RPO/RTO may be longer.
What else should be budgeted?
Staging/replica storage, snapshot/recovery-point retention, cross-region egress, drill/failover compute, support/licenses and operations. In the AWS example, EBS + snapshots + replication EC2 exceed the DRS fee itself.
Continuous replication or backup?
Minutes-level RPO/fast failover → DRS/ASR; hours-level RPO/cost-sensitive → GCP Backup and DR or snapshots + object storage. Set RPO/RTO first, then platform/operations/compliance.
What does 100 servers cost?
At 100 servers/30 TB: AWS DRS about $6,389/month, Azure ASR about $4,036/month, GCP Backup and DR about $1,690/month; excludes egress, drill compute, tax and discounts; the bill is authoritative.
Can we buy DR without drills?
Not recommended. DR value is recoverability; without drills you cannot validate RTO/application consistency/dependencies. Budget drill compute/labor, test regularly and keep evidence.
Research method and limitations
This article separates "official price facts" from "editorial calculations." AWS DRS's $0.028/server-hour, EBS/snapshot/replication EC2 example and $6,389.03 total come from the AWS DRS pricing page. Azure ASR's $25/instance-month, first-31-days-free and separate storage/egress/compute come from the Azure Site Recovery pricing page and Azure Retail Prices API (eastus, 8 November 2026). GCP Backup and DR's per-GiB management/long-term storage rates come from the Google Cloud Backup and DR pricing page (us-central1, 8 November 2026). The Azure/GCP 100-server/30 TB figures, scale table, hidden costs and 90-day checklist are editorial.
Eight limitations: (1) only public list prices, excluding tax, discounts, support/licenses; (2) only US East/US Central; (3) assumes 100 servers/30 TB/300 GB per server/7-day retention/3.3% daily change; (4) GCP's RPO/RTO model differs from DRS/ASR and is not directly equivalent; (5) egress/drill/recovery compute/operations are excluded; (6) not all source types (VMware/physical/in-cloud) are covered; (7) no real replication/recovery benchmark is included; (8) prices/products change, so recheck the official pages and bill before purchase.
Sources and basis
- AWS, Elastic Disaster Recovery Pricing: aws.amazon.com/disaster-recovery/pricing. $0.028/server-hour, EBS/snapshot/replication EC2 example, $6,389.03 total. Accessed 8 November 2026.
- Microsoft Azure, Site Recovery Pricing: azure.microsoft.com/en-us/pricing/details/site-recovery. $25/instance-month, first 31 days free, separate storage/egress/compute. Accessed 8 November 2026.
- Azure Retail Prices API (productName contains Site Recovery, eastus): prices.azure.com/api/retail/prices. VM Replicated to Azure $25/month. Accessed 8 November 2026.
- Google Cloud, Backup and DR Pricing: cloud.google.com/backup-disaster-recovery/pricing. Per-GiB management and long-term storage. Accessed 8 November 2026.
Data cut-off 8 November 2026; prices/availability follow the AWS/Azure/Google official pages and bills; this is not procurement, commercial or legal advice.