Companies used to prepare for disaster by running a duplicate server room off-site. Maintaining it cost tens of thousands a year. DRaaS (Disaster Recovery as a Service) keeps replicas of your important systems in the cloud. Those replicas can be running within minutes if the primary site fails.
What DRaaS does
DRaaS replicates your servers, data and applications into a cloud environment as they change. Hardware failure, ransomware, fire or flood can take the primary systems offline. The replicas then take over. Staff log in to the replicated environment and carry on, while the original systems are repaired or rebuilt.
Every recovery plan is measured against two numbers:
RPO (Recovery Point Objective): The volume of data you're prepared to lose. An RPO of one hour means replication runs at least hourly.
RTO (Recovery Time Objective): How soon systems must be usable again. An RTO of four hours means failover has to finish inside that window.
DRaaS vs backup
A backup stores copies of your data. DRaaS stores a working copy of the full environment.
When a server fails and backups are all you have, recovery looks like this:
- Obtain replacement hardware (days)
- Install the operating system and applications again (hours)
- Pull the data back from backup (hours)
- Check that everything works (hours)
Overall recovery time: days to weeks.
With DRaaS:
- Start the failover (minutes)
- Staff reach the cloud environment (minutes)
- Work continues
Overall recovery time: minutes to hours.
What gets replicated
The usual set is the systems the business actually runs on:
- File servers holding company files and shared data
- Application servers for CRM, ERP and other line-of-business software
- Domain controllers providing Active Directory and DNS
- Database servers such as SQL and PostgreSQL
- Email servers still hosted on site
Microsoft 365 and similar cloud-native services already run on Microsoft's own redundant infrastructure, so they don't need DRaaS.
What it costs
Pricing follows the volume of data you replicate and the compute reserved for failover:
- Small environment (2 to 3 servers, 500GB of data): £200 to £500 a month
- Medium environment (5 to 10 servers, 2TB of data): £500 to £1,500 a month
- Large environment (10 or more servers, 5TB or more of data): £1,500 to £5,000 a month
These figures are based on Azure Site Recovery or an equivalent platform. Running a physical DR site costs considerably more each month.
Common mistakes
Skipping failover tests. An untested recovery arrangement exists on paper. It has not been proven. Run failover tests at least twice a year.
Leaving some systems out. If the file server fails over and the domain controller does not, staff can't log in. Every system the others depend on has to be included.
Leaving connectivity out. After failover, people still need a route to the cloud environment. Remote access, VPN and internet connectivity all belong in the design.
Predictive monitoring
Predictive monitoring watches your systems for faults that are starting to develop. DRaaS is what you rely on if those faults still take systems down.
How Wanzo implements DRaaS
We replicate your important servers to Azure using Azure Site Recovery. We agree RPO and RTO targets with you, set up replication, prepare failover networking and run scheduled tests. If a disaster occurs, we carry out the failover and manage recovery.
What to do next
If you want DRaaS for the servers you run today, send us a message. We'll map what needs replicating, agree RPO and RTO figures, and set out a monthly cost.