Sergey Nivens - Fotolia
When evaluating backup products, you are making a long-term decision. Backup and archive platforms tend to be the stickiest products in enterprise IT, as you usually keep them longer than your data retention requirements. After you stopped storing new data, these archives must remain accessible for a long time.
Modern, converged data protection and secondary storage platforms have some unusual characteristics to consider as you look to gain some new backup capabilities. These include the scale-out nature of converged backup, as well as public cloud support and the simplicity of operations.
Converged data protection has a lot in common with hyper-converged compute; it uses a scale-out architecture with multiple physical servers forming a cluster. Ask yourself the following questions.
What does scale mean in your organization? Does it align with the converged backup vendor you are considering? You might require a vast backup volume in your large data center. Has the vendor deployed its offering for the number of nodes you need for capacity or performance?
You might need to implement data protection in many small locations. Can the converged backup vendor provide a single console and a single set of policies to manage all those locations? Can they scale down to the size of your small sites? Maybe you should use small physical nodes in a cluster, a single physical node at each location or a virtual appliance version of its converged secondary storage platform.
Another consideration is whether the converged backup product requires you to buy hardware or enables you to reuse some existing servers and load the software. If you must buy the vendor's servers, what does the hardware support look like? Can the hardware support integrate with your existing servers, support contracts and mechanisms?
Data platform support
All of the converged data protection vendors started with vSphere support, which is excellent if every application you use is in a vSphere virtual machine (VM). What about application awareness? Do you have database servers or email servers that you need to protect? Does the backup vendor have application-aware agents? If you have physical servers, then you will need agent-based backups.
What about public-cloud based resources? Can the backup vendor protect your SaaS platforms, such as Salesforce and Office 365? For data-aware backups, what is the granularity of protection and restore? You may not want to restore an entire database server to recover one small table, for instance.
Virtualization platform support
Different virtualization platforms have different capabilities. Mostly, converged data protection vendors support vSphere the best. But what about if you use Hyper-V? Or some form of Linux and kernel-based VM virtualization? Does the converged backup vendor have the same level of support for those hypervisors? In particular, the instant restore features that are a great part of vSphere may not be available with other hypervisors. Does the platform support file restore back into the VM file system with your hypervisor?
Newer backup products tend to have a cloud-friendly design, particularly to address off-site data copies and long-term archive using object storage in the public cloud. Look for granular control of this public cloud usage -- principally, control of retention, as public cloud storage is charged per gigabyte per month.
Restore to the cloud is another great option, using your converged backup platform as part of your disaster recovery (DR) plan or a DevOps workflow. Evaluate your needs for automation of these processes. Evaluate both for DR failover and any routine data migration, such as using the public cloud for reporting by restoring production database servers into the cloud.
If you plan to use DR to the public cloud, investigate what failback looks like. It is usually much harder to fail back from the public cloud to on premises than to fail over during a disaster.
In addition to using hyper-converged infrastructure, converged backup also follows the simplification model of HCI. The aim is to minimize the amount of human effort involved in data protection.
Look for simple, policy-based protection management, rather than complex schedule balancing. Expect that routine backup management should be about managing a small number of exceptions where the policy was breached, rather than checking logs for success.
Another useful element is an API to manage data protection programmatically rather than manually. Another aspect of simplification is the ability to automatically protect new VMs based on their identity rather than a help desk ticket to add them to a backup job.
Check for gaps
Converged backup is a relatively new category, and most of the vendors are relatively new to the data protection landscape. Features that you assume in more established products may not be available. The top example is that few converged backup platforms include backup or archive to tape. If you have a specific requirement for tape as a long-term archive, then your converged data protection options will be limited.