With the rapid rise in popularity of virtualization and cloud technologies in modern business infrastructure, cloud-based disaster recovery is becoming increasingly important. As with any other disaster recovery (DR) approach, the main objective in cloud-based DR is restoring your data and getting your services back up and running in the shortest possible period of time. Does cloud-based DR make sense for your business? Today’s blog post discusses the benefits of disaster recovery in the cloud and disaster recovery as a service. The requirements for cloud-based DR and the benefits afforded by this strategy are also explored.
What Is Disaster Recovery?
Disaster recovery is a set of measures aimed at recovering data and service provision after failure. There are a number of reasons your virtual infrastructure could be taken out, including natural disasters, technological failure, human error, etc. Virtual machines (VMs) must be recovered on physical hosts with all the data and applications they use to provide different services. The services must continue running properly when the restore process is complete. It is imperative that a comprehensive business continuity plan and a disaster recovery plan are not just created before disaster strikes, but also thoroughly tested. RPO and RTO values must be defined. Classic disaster recovery presumes that if a physical host (or any other hardware component, such as attached storage or a network switch) fails, then you must use other compatible hardware, either on-premises or at a remote site, to recover your virtual environment. You would also need to install special software for backup and recovery. When using the cloud for VM disaster recovery, however, you don’t need to have any replacement physical hardware. The cloud provider’s computing environment can be used. Moreover, you don’t need to install software for backup and recovery if you are using disaster recovery as a service in the cloud.
What Is “Disaster Recovery as a Service”?
Disaster recovery as a service (DRaaS) is a complex third-party service provided by a managed service provider (MSP) in the cloud for cloud-based disaster recovery. The MSP usually also provides VM backup as a service (BaaS), replication as a service (RaaS), recovery, and failover to make disaster recovery possible. The VMs in your production infrastructure can be replicated to the secondary virtual infrastructure in the cloud with these services, in preparation for cloud-based disaster recovery. DRaaS can provide VM failover in a case of disaster, then failback for moving the workloads back to your primary site. DRaaS can be used for public clouds as well as private clouds.
Public cloud DRaaS is implemented when organizations use public cloud infrastructure that meets the requirements for disaster recovery. The administrator of the organization using DRaaS can usually manage and configure the provided services with the web interface of a self-service portal.
Private cloud DRaaS can be used by organizations that have private cloud infrastructure in their own datacenters. Managed services can be used for disaster recovery in the private cloud. As an alternative, appropriate software (which usually supports Multi-Tenant architecture) can be installed in the organization’s own physical infrastructure.
Hot Site vs. Cold Site
A hot site is a site that has hardware and software ready for immediate use if a disaster occurs. A hot site usually has VM replicas that are updated regularly with fresh recovery points after each replication job. Network configuration, hypervisor configuration (e.g. ESXi hosts, vCenter), and storage system configuration is ready to go. If the primary site holding your VMs fails, you can fail over to VM replicas at a hot site in the shortest time. After recovering the VM replicas (which should be the most important business-critical VMs in your virtual infrastructure), the other VMs can be recovered from their backups. A cloud environment can be used as a hot site for your cloud-based disaster recovery plan.
A cold site has the minimum configured software and hardware. Sometimes a cold site is an empty space that is reserved for the deployment of infrastructure if disaster strikes. Recovery to a cold site takes much more time, because almost all components must be configured from scratch. The cold site approach to disaster recovery can be used when the RTO values are high.
Thus, you can determine which type of site is better for DR in your case by weighing your RTOs and the costs associated with the upkeep of such sites. A hot site is more expensive than a cold site, but a hot site allows VMs and workloads to be recovered in the shortest possible timeframes.
Prerequisites for Cloud-Based Disaster Recovery
- Backup and replication to cloud. If you want to use failover to replica, you must first create VM replicas. To ensure that you can recover your VMs in the cloud, you must back up your VMs to a safe place (a cloud is suitable for this).
- High network bandwidth. Virtual machines can hold large amounts of data that must be transmitted over the network from the primary site (production site) to the secondary site (DR site) when preparing for disaster recovery. If these sites are located far away from each other, then a high-speed Internet connection is required. Don’t forget about security — use encryption for traffic between your primary and secondary sites.
- Select an appropriate managed service provider whom you can trust with your cloud-based disaster recovery plan. You must order the appropriate virtual infrastructure for disaster recovery in the cloud. Configure each component of your remote site in the cloud and make sure they are prepared for a potential disaster. You can rely on DRaaS or install and configure backup software by yourself.
Benefits of Disaster Recovery Based in Cloud
- Cloud-based DR minimizes dependencies on physical infrastructure. One of the general benefits of using virtual machines is their high compatibility for migration between hardware servers. Hardware used in the cloud is abstracted for customers; VMs can be migrated and run in the cloud just as they can run on physical servers. You don’t need to upgrade or support your physical infrastructure used for disaster recovery at the remote site if you use a cloud-based disaster recovery plan.
- High scalability is another result of abstracted hardware. Servers and infrastructure can be upgraded transparently for VMs. When using cloud services and MSPs, a “pay as you go” approach is used (also known as the pay-per-use model). If a customer needs more storage, memory, or processing power — for a disaster recovery site, for example — then the customer simply requests the MSP for such upgrade and pays more for the additional provided services. If the customer no longer needs some of the resources they have been assigned, and they remain idle, then the customer can just as easily ask the MSP to reduce the amount of disk space, memory, or processors associated with their subscription, lowering the costs for their services.
- Cloud-based DR also affords the possibility to run recovered VMs in different geographical regions, because you can connect to the cloud from anywhere in the world. Having physical infrastructure in different countries is not always possible; at best, it is complicated. By using clouds, you can buy infrastructure in any region you like and connect to that cloud-based infrastructure from anywhere via the Internet.
Despite the major benefits of disaster recovery based in cloud listed above, there are also some drawbacks associated with cloud-based disaster recovery:
- Cloud-based DR is not a good solution where you have limited network bandwidth. Internet connection is crucial. If your company’s staff use a 1-Gbit network connection during their working processes (e.g. for copying multimedia content from workstations to virtual servers and back during editing) they would notice significantly lower network speeds when failover to cloud occurs. Keep this factor in mind when planning your disaster recovery processes.
- Data is stored on third-party servers. If your company operates with sensitive data, you should consider whether you can completely trust the third-party managed service provider with this private data. If you can trust the MSP, then you can use cloud for disaster recovery. If not, then you should build your own remote disaster recovery site with physical infrastructure.
Cloud-Based Disaster Recovery Support in NAKIVO Backup & Replication
NAKIVO Backup & Replication includes the following features that can be used for disaster recovery of VMware VMs, Hyper-V VMs and Amazon AWS instances in the cloud.
- Multi-tenancy support. The edition of NAKIVO Backup & Replication for managed service providers uses a Multi-Tenant architecture that allows the logical isolation of customers. The product can be deployed in the cloud for the purposes of providing BaaS, RaaS, and DRaaS. The main benefits of this approach are security for customers and ease of management for MSPs.
- VM replication. Having a VM replica is imperative for the “failover to replica” strategy to be used. NAKIVO Backup & Replication can create VM replicas that are the identical copies of the source VMs. Each recovery point represents a regular VM snapshot. Space-saving as well as traffic reduction technologies are used, including Changed Block Tracking (CBT) for VMware VMs and Resilient Change Tracking (RCT) for Hyper-V VMs.
- Automated VM failover with Network Mapping and Re-IP. This new feature provides failover to replica. VM networks and IP addresses at your DR site might differ from those of your production site. Network Mapping helps map the target network to VM replicas automatically when failover occurs. Re-IP changes the IP settings inside the VM for automatic connection of the VM to the IP network of the DR site upon failover.
- Backup copy to cloud. The more copies of your data you have, the higher the probability of successful recovery. NAKIVO Backup & Replication can store copies of your backed-up VMs in the cloud. You can store these VM backups in the cloud that is used as a disaster recovery site or in another cloud.