Recently at Dell World, I was part of a conversation about what would be utopian disaster recovery and where we are today in the state of the industry. But where we are today is transforming, with a new name that encompasses many technologies. We are now using the term “data protection” (DP) to mean much more than just disaster recovery (DR), backup, business continuity (BC), replication, data loss prevention, and replication, but also the basic functions of confidentiality, such as encryption. The main goal of data protection is to provide a way to use your data as quickly as possible wherever it is needed and with minimal or no loss.
Some tools for disaster recovery provide just one of these features, and some provide a few, but the goals are similar. The benefit of these tools to the organization is to provide a way to recover in all forms. However, to just recover is quite a bit of work, depending on where you want to recover: to new hardware, new or different hypervisors, or even into the cloud. Disaster recovery not only waits in the wings to be implemented, but is often a part of migration from one data center to another, from a data center to a cloud, or even from one cloud to another. This is the today, but what is the utopian view of the future of DP?
Utopian Disaster Recovery
Utopian Disaster Recovery is just another name for fully operable data protection that is not only about the data, but also about the application, the user, and how the user interacts with the application and hence the data. Utopia is reached when disaster recovery is hardly ever done, because the other aspects of data protection have already taken place. In addition to providing business continuity, these aspects provide replication and backup of everything needed to restore application and user interactions at another location, whether that be a cloud, a hot-site, another part of your data center, or even just more hardware. Utopian Disaster Recovery also requires that the data protection tool automatically determine what applications are in use, what devices are in use, what data is in use, what the users want, and ultimately what jurisdiction the data and users are within, so that the appropriate applications and data can be presented and made available at any time.
Automation meets legal in many ways. With clouds in use and the ease in which jurisdictional borders can be crossed, legal requirements around data and use of that data need to be considered as a part of data protection and disaster recovery. In addition, considering the volume of data, applications, and users, we must start to automate disaster recovery, making it possible to immediately launch applications and provide access to data from anywhere. No longer should recovery take the right king of tape devices and tools; it should be made as simple as possible. We also need to trust our automation and widen the scope of what we protect to include all aspects of the application.
Transformation
Currently, we are in the midst of a transformation from the old way of thinking about DR—that it is just for emergencies—to a new, holistic view of data protection. Here is a list of products and where they are:
Dell AppAssure – One of the tools in Dell’s data protection portfolio that concentrates on specific applications (such as Microsoft Exchange, SQL, or SharePoint) to ensure their data is backed up. However, the list of applications are only those from the major vendors. If there is a customer application built on top of one of their defined applications, it is currently out of scope for AppAssure. AppAssure provides the ability to recovery test the defined applications as well as the ability to restore such data onto any hypervisor or any hardware, or into select clouds. Dell AppAssure is hypervisor agnostic. (DR, BC, application aware but missing application awareness outside predefined applications, multi-hypervisor)
Zerto – Zerto provides near real-time data protection by performing replication within a VMware vSphere virtual environment only. Such replication can occur between machines in the same data center, in different data centers, and into a growing number of VMware vSphere-based clouds. Replicated virtual machines are ready to be run on the replication target as needed. Definition of the application is left up to the administrator in the form of virtual machine groups. (DR, BC, replication, not application aware, single hypervisor)
Symantec Backup Exec – Backup Exec uses a variety of targets and sources for backup of data and even full virtual machines; however, it is not application aware. Backup Exec is mainly used for backup and recovery. For recovery purposes, Backup Exec can peer into virtual environments to aid in recovery. There is no real application awareness. Administrators have to define the applications using backup groups. Backup Exec can restore to any hypervisor. (DR, backup, not application aware, single hypervisor)
Hotlink DR – Hotlink provides an interesting set of cross-cloud management tools that are hypervisor agnostic. They have extended that concept to disaster recovery by backing up virtual machines from VMware vSphere environments to Amazon S3, transforming the virtual machines and making them available as Amazon Machine Instances (AMI) as necessary. All are managed from within VMware’s vSphere platform. Not only does Hotlink DR provide replication, but also migration and other services. Selection of the virtual machines to back up, however, is not yet application aware. However, it is integrated into the base management console, VMware vCenter, of VMware vSphere environments. (DR, BC, replication, not application aware, multi-hypervisor)
Veeam Backup and Replication – Veeam provides an ever-increasing set of tools that span multiple hypervisors, but it does not allow for the automatic restoration between them. Veeam was the first to introduce automated testing (SureBackup) and near-continuous replication between sites (including into clouds), and continues to provide means to ensure your backup is ready to use as needed, either as part of DR or as a part of BC. Not limited to just VMware vSphere, Veeam works within Hyper-V as well. They also now write to tapes, as well as recover data from storage snapshots. (DR, BC, automated testing, application aware for testing, no multi-hypervisor transformation)
EMC Avamar – Avamar is the basis for VMware’s own data protection tool. It has many interesting tie-ins to the VMware vSphere virtual environment , but it is not yet application aware. Avamar provides backup capabilities and replication between Avamar nodes. (DR, backup, not application aware, single hypervisor)
Neverfail – Neverfail provides continuous data protection as well as tools to architect a good disaster recovery solution using their IT Continuity Architect. The IT Continuity Architect tries to determine if service level agreements are being met and how many virtual machines and systems are actually being backed up. However, there is no automatic detection of an application. (DR, BC, CDP, not application aware, single hypervisor)
Closing Thoughts
There are many data protection tools available out there. The questions you need to ask are about the future, how easy are they to recover data from, how automated, and how do I know I have a safe, sane, and working backup that covers my entire application. This often starts with data classification, but ideally it should be automated. Our data protection tools should automatically know how our applications are defined by looking at them from a user perspective. I would like to see a marriage between a definition of an application and how the users use the application, with data protection, so that we can eventually achieve utopian disaster recovery.
We are far from there; tools are just starting to branch out and cover more than one hypervisor, provide their own definitions of an application, do automated testing, and provide automated recovery. Just getting the data is not enough. We need to be able to easily stand up a user-centric application at the push of a button that has been tested and can be restored anywhere without needing to worry about the underlying hardware or hypervisor or jurisdictional issues.
Where are you with achieving utopian disaster recovery? What tools do you use today, and are they application aware?