One of the key components of any disaster recovery plan is to verify that plan and make sure it works. The problem is that these tests are expensive, are not real world and are performed infrequently. In this article, we will discuss how data centers can leverage virtualization, disk backup and the cloud to significantly reduce the cost and complexity of DR testing so that it can be done more frequently.
DR Testing Dilemma
Testing disaster recovery is often a task that falls to the bottom of the IT priority list. In the past, many organizations would invest the time and resources to test their recovery plans but even then, these tests would at best take place once a quarter, but in most cases only occur once or twice a year. While this frequency of testing was arguably inadequate, it’s gotten worse given the fact that there are fewer dollars in today’s IT budget to allocate towards testing and IT people are already consumed with other, more pressing business tasks.
DR Testing On-Demand
Fortunately the flexibility of virtualized infrastructure now makes it relatively easy to clone or migrate business applications across private data center environments or out into the public cloud. In fact, through data protection software that directly integrates with the hypervisor, businesses can non-disruptively capture business data in near real-time and mirror it locally or off-site to create a separate DR test bed environment. This capability gives IT the wherewithal to conduct DR testing on-demand without committing huge sums of money to a DR provider or spending inordinate amounts of time planning and then conducting tests. In essence, DR testing can become a routine exercise that takes minutes or hours instead of days or weeks.
Backup Leveraged DR Testing
One way to implement a cost-effective DR testing plan is to utilize virtualization specific backup technology, like Veeam or EMC’s Avamar, to backup virtual machine (VM) data and replicate it off-site. These backup solutions integrate into the hypervisor to allow businesses to capture highly granular backups that place little to no overhead on the guest VMs that they’re protecting. This data can then be copied to a protection storage resource, like a disk backup appliance with deduplication, for efficient local backup image retention and bandwidth friendly off-site replication. Some of these backup solutions provide a feature called “recovery-in-place”. This means that the VM image that is stored on the backup appliance (or other backup disk resource) can be mounted by a VM to provide rapid application recovery capabilities.
Pro-Active DR Testing
An application or backup administrator can then periodically test the ability of their applications to recover from a failure by performing a recovery-in-place; whether the test is done locally or remotely. The benefit is that the business can demonstrate conclusively whether they are meeting application recovery service level agreements (SLAs) or not. If a problem is discovered, then corrective action can be taken before a real failure or disaster occurs; helping to reduce the risk of a prolonged outage. Ideally, these tests should be performed off-site, as well as onsite, so that businesses can determine all the other factors that could impact application service levels during a disaster – like network latency.
DR Testing Cloud Savings
If a business doesn’t have a secondary data center to test from, then this backup data can be replicated to a cloud service provider that offers backup or DR-as-a-Service (BaaS/DRaaS). That way an organization doesn’t have to incur the expense of building out a secondary site. Instead they can leverage the provider’s multi-tenant environment and only pay for the actual server and storage resources they consume.
Some of these providers, like VMware, allow the client to spin-up test VMs into their facilities so that they can emulate their production applications in a DR test scenario. Technologies like EMC RecoverPoint, for example, help enable non-disruptive DR testing by replicating data from the primary data center to a secondary data center or cloud provider facility. Since these environments are completely separate from the primary production data center, application owners can conduct risk-free DR testing as frequently as they wish. This can enable businesses to validate that their DR service levels are being met to satisfy both internal and external compliance audits.
VM Level DR Protection Policies
Through RecoverPoint, individual VMs can be selected for continuous data protection (CDP) and have their application data replicated into a cloud provider facility, like VMware’s, for DR protection. For example, VM#1 may host a critical business application while VM#2 represents a non-critical system. With RecoverPoint, the application administrator can designate minute by minute CDP for VM#1 and choose to ignore VM#2 or instead, give it a lower protection policy – once a day, for example. This helps to ensure that the most important business systems are given resource priority both during the protection process and during a recovery event.
As Phil George from EMC explained in a recent Storage Swiss chalk talk video, businesses need to move towards offering data protection-as-a-service (DPaaS) to simplify how their internal customers consume IT. One way of accomplishing this is by leveraging data protection technologies, like RecoverPoint and Avamar, that have been fully integrated into VMware’s vCloud Director solution. By creating virtual data centers (VDCs) with vCloud Director, businesses can abstract all the manual configuration work that is typically required to protect virtualized applications and fully automate the data protection provisioning process. Application owners can simply choose a service level policy (Platinum, Gold, Silver, Bronze, etc.) to obtain the data protection that is most appropriate for their application.
DR For The Edge
As importantly, VMs in far flung office locations can be protected without the need to deploy additional backup hardware at the remote site. This helps to reign in data on the edge and centrally protect it. And since RecoverPoint and Avamar offer a “software-defined” approach to data protection, DR testing can be performed within the enterprise or within a cloud provider’s facility.
Regular, consistent disaster recovery testing is now within reach of all organizations given the advanced capabilities of data protection software technologies and replication offerings. These solutions can be utilized within a private cloud or a hybrid cloud to provide businesses with the maximum flexibility for testing their DR readiness. And this can all be done without impacting business applications or requiring a huge financial outlay.
Sponsored by EMC