Disaster Recovery as a Service (DRaaS) looks great on paper. DRaaS vendors tell organizations that they can eliminate their disaster recovery site and recover on-demand, in the cloud. The potential for cost savings abounds. Organizations have the potential to reduce the cost of the disaster recovery site, the equipment (storage and servers) on standby at that site, and the operational overhead of maintaining that site.
Many DRaaS solutions claim push-button disaster recovery, but they don’t explain that a push-button recovery may not mean a quick recovery. The problem is that there are time compromises organizations need to live with if they expect to use the typical DRaaS solution.
Many DRaaS vendors use Amazon Web Services (AWS) as the place where they execute their customer’s disaster recovery. The problem is the overwhelming majority of an organization’s applications, and data exists within a VMware environment. AWS does not, natively, run the vSphere hypervisor, so any VMware VM needs conversion to run correctly within the Amazon cloud. Amazon does provide conversion tools to convert a VMware VM into an Amazon Machine Image (AMI). The challenge is the conversion process takes time, in many cases hours.
Another challenge is that most DRaaS vendors store the protected copies of data on Amazon’s Simple Storage Service (S3) to save money. S3, though, is not typically suitable as storage for applications or their data. As a result, before execution, IT, or the DRaaS solution, need to copy the VM from S3 storage to the higher-performing Elastic Block Storage (EBS). Depending on the size of the VM, the transfer time can also take hours to complete.
These are just two examples of challenges IT faces when trying to recover with DRaaS solutions that use Amazon AWS as part of their solution. These challenges are not unique to Amazon. DRaaS in almost any public cloud provider will present similar obstacles.
Other challenges to successful cloud DR include the reality that during the disaster, the IT team is operating in an environment foreign to them, rather than the VMware vCenter, which they know. Once the failure passes, they have to face the prospect of failing back, which means reversing all the changes made to the VMware VMs as well as the time involved to return to normal operations.
Datrium – A No Compromise VMware DRaaS
Datrium, makers of a disaggregated hyperconverged solution, DVX, that brings scale to previously limited hyperconverged environments, also delivers a DRaaS solution that eliminates most of the DRaaS recovery time challenges. In the past, using their DRaaS solution required implementing and migrating to Datrium’s hyperconverged solution. Now customers can leverage Datrium’s unique DRaaS capabilities in any VMware environment with its new DRaaS Connect.
The Datrium DRaaS solution leverages VMware Cloud on AWS. As a result, a disaster plan no longer needs to wait for conversion before executing. Also, during the disaster, the customer is running in an environment that is VMware and requires no additional training. In the event of a disaster, the Datrium orchestration software starts a software defined data center (SDDC) in VMware Cloud on AWS (VMC). To overcome the S3 to EBS time delay, Datrium uses the same technology it uses in its on-premises disaggregate hyperconverged solution, DVX. ESX hosts in the SDDC mount backups on S3 directly via a live NFS datastore exposed by Datrium software, and start executing workloads using the Datrium’s DR runbook orchestration software. The NFS datastore contains all the backups needed to execute the runbook. This is achieved by using NVMe flash & EC2 instances layered on top of S3 backups, for caching and IO processing. This technique is identical to the split-provisioned architecture on-premises.
Since the DR environment is running in a VMware infrastructure, it requires no conversion fail-back. Once the on-premises data center is ready to host applications, the solution replicates data back to it. There is no downtime or interruption of services, while Datrium is recovering to the new data center.
The new release of DRaaS Connect enables any organization to take advantage of Datrium’s DRaaS capabilities without standardizing on the Datrium’s DVX solution first. In some ways, though, DRaaS connect customers are using a flavor of DVX when they recover in a VMware cloud, and data is fed into it from S3. DRaaS Connect represents an enormous opportunity for Datrium to not only solve customer’s most critical problems but also to be a proven, trusted partner when the time comes to refresh VMware storage infrastructures.
In theory, organizations can replicate data from their on-premises data center to a VMware Cloud. The problem is that storing a secondary copy of data in EBS storage is cost-prohibitive for all but the most critical of VMs. The Datrium approach enables organizations to include all of their VMs in their DR plan thanks to cost-effective S3 storage and the company’s ability to make VMs instantly available with high performance.
Datrium’s maturing of the solution and especially their ability to make the solution universally available, make DRaaS Connect a solution that VMware environments looking to improve their DR strategies, should strongly consider.