Last reviewed: October 27, 2026 · Reading time: 10 minutes · Prices are list prices at the time of writing and change regularly; always confirm with the provider’s calculator.
Ten years ago, a second data center was the only serious way to recover from losing your primary one. Today Microsoft and Amazon both sell disaster recovery as a service, billed per protected server, so a 20-person firm can have the same failover capability as a bank. The two leading options are Azure Site Recovery (ASR) and AWS Elastic Disaster Recovery (AWS DRS). We deploy both, and the right answer depends far more on your existing environment than on the headline price.
- List prices are close: about $25 per protected instance per month for ASR and $0.028 per source server per hour (roughly $20.44 a month) for AWS DRS.
- The service fee is often the smaller part of the bill. Replica storage, snapshots, replication servers and failover compute are charged separately on both platforms.
- ASR includes the first 31 days free per protected instance.
- Choose based on where your identity, licensing and skills already live, then validate with a real test failover.
Headline pricing
On fees alone, AWS DRS works out around 18% cheaper per server per month. In practice that difference is often swamped by the separately billed items below, so compare total cost, not the line item.
What you pay for beyond the service fee
| Cost item | Azure Site Recovery | AWS Elastic Disaster Recovery |
|---|---|---|
| Service fee | Per protected instance per month; first 31 days free | $0.028 per replicating source server per hour; no upfront or minimum fee |
| Replica storage | Managed disks in the target region | EBS staging disks, one per source disk (AWS’s example uses gp3 at $0.08/GB-month or sc1 at $0.015/GB-month) |
| Recovery point snapshots | Billed by consumed capacity at managed disk snapshot rates | EBS snapshots ($0.05/GB-month in AWS’s example, 7-day default retention) |
| Replication infrastructure | Included in the service | Lightweight replication servers (t3.small in AWS’s example, about $0.0208/hour each) |
| Compute | Only during test failover or real failover | Only during drills or recovery |
| Data transfer | Egress charged only when traffic leaves an Azure region | Data transfer and conversion servers not included in the service fee |
AWS’s own worked example on its pricing page shows how the extras add up at scale: about $1,425 a month for 30 TB of staging storage and about $1,846 a month for snapshots, on top of the per-server fee. For a typical SMB with a few terabytes, those numbers are far smaller, but the principle holds: storage and retention choices drive the bill.
Capabilities compared
| Azure Site Recovery | AWS Elastic Disaster Recovery | |
|---|---|---|
| Best fit | Microsoft-centric firms already on Microsoft 365, Entra ID and Hyper-V or VMware | Firms already using AWS, or wanting to recover mixed physical, virtual and other-cloud servers into AWS |
| Sources protected | Hyper-V, VMware, physical servers, and Azure VMs between regions | Physical, virtual and cloud-based servers replicated into AWS |
| Replication | Continuous for VMware and physical; scheduled frequencies for Hyper-V | Continuous block-level replication with point-in-time recovery |
| Test without disruption | Test failover to an isolated network | Drill launches in AWS, included in the service fee |
| Orchestration | Recovery plans with ordered groups and scripts | Launch templates and post-launch actions |
| Failback | Supported to on-premises or the original region | Supported back to the source environment |
How we help clients choose
- Follow your identity and licensing. If your staff sign in with Microsoft 365 and Entra ID, ASR usually means fewer moving parts. If you already run workloads in AWS, DRS keeps everything in one console and one bill.
- Size the storage honestly. Retention of recovery points is the biggest swing factor in monthly cost on both platforms.
- Decide your RPO and RTO first. Both can meet aggressive targets; the question is what each system actually needs. See our guide to setting RTO and RPO.
- Don’t forget the data layer. Server replication is not a substitute for Microsoft 365 backup or immutable file backups, because ransomware replicates too.
- Run a real drill before you sign off. A test failover tells you more than any pricing sheet.
Running IBM i? Neither service is the natural home for those workloads; see our IBM i Cloud Disaster Recovery service instead.
Not sure which platform fits?
Ron Klink deploys both Azure Site Recovery and AWS Elastic Disaster Recovery and will give you a straight recommendation with a total-cost estimate.
Sources
- Microsoft, Azure Site Recovery pricing, accessed October 2026.
- Amazon Web Services, AWS Elastic Disaster Recovery pricing, accessed October 2026.
- Otava, What does disaster recovery as a service cost?, July 2026.
