One size does not fit all when it comes to meeting VMware-based virtual machine (VM) data protection requirements. Costs must be kept in check while shrinking recovery time and meeting recovery point objectives. IT professionals are pressured to deliver near-zero data loss and near-zero application downtime. They also need to comply with data privacy regulations such as the European Union’s (EU’s) General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA). Finally, the IT professional must develop strategies that protect against increasingly tenacious ransomware attacks.
The problem is that protecting mission-critical VMs with legacy approaches is complex and expensive. It is also likely to be sub-par when it comes to meeting more demanding protection service level agreements (SLAs). Many enterprises are turning to the cloud—hoping to cut infrastructure costs, accelerate recovery times, and simplify management. However, many cloud data protection solutions simply “bolt on” cloud support rather than being truly cloud native, and as a result fall short of the “cloud promise.” These solutions cannot provide key functionalities, including on-demand capacity scalability, utilizing in-cloud compute, instant data availability, minimal data loss, workload mobility, and improved data visibility.
The current state of VMware backup and recovery
Most disaster recovery and business continuity solutions for VMware environments on the market today are designed for legacy on-premises environments. Cloud capabilities were not included from the beginning of development. The primary problem is that these solutions do not leverage the cloud intelligently.
The first generation of cloud-enabled backup and recovery solutions effectively leveraged the cloud as a replacement for tape archives—a long-term repository for secondary data copies that in an ideal world never need to be retrieved again. But now, a number of backup solutions have taken the next step in terms of running backup software in a virtual machine in the cloud, using the cloud as a tier for old backups and even providing some in-cloud disaster recovery features. However, these “cloud-washed” solutions still lack a number of key capabilities.
Current solutions lack cloud tiering
The first problem is that some of these “cloud-washed” solutions often require all of the organization’s data to be backed up both on-premises and in the cloud. The cloud implementation becomes a mirror copy of the on-premises environment, rather than being used to shrink the on-premises infrastructure footprint. Capacity requirements effectively double because data is being stored twice, both on-premises and in the cloud.
Cloud-washed solutions lack automation
The operations problem is exacerbated by the fact that most cloud-washed solutions do not automate the highly laborious, manual processes required for recovery—which not only adds costs but also slows time-to-recovery. For example, recovering from a disaster, unlike a single server outage, requires multiple virtual machines to be recovered and in a specific order. The problem is most backup solutions don’t automate runbook execution, which means that IT must engage in the costly and time-consuming manual testing of processes.
Limited integration with cloud compute
Cloud compute cycles are more cost-effective and scalable than on-premises server implementations. With cloud compute the business pays only for the cycles that it uses, but without proper integration to leverage these cycles on-demand, “cloud-washed” solutions are costing you more than they should. These backup solutions were designed for servers that are ‘always-on’, so they consume cloud compute resources even when idle.
Cloud-washed solutions also miss another opportunity, which is to leverage scale-out compute in the operation of the data protection solution itself. Data protection requires multiple schedules to be managed and executed, and indexes to be maintained. Additional value-add services like search or legal hold require even more CPU resources. A cloud-washed solution is typically scale-up in nature and can’t leverage processing power beyond a single server or node.
Requirements for cloud-native VMware backup and disaster recovery
Truly cutting costs and improving business agility and continuity for VMware-based hybrid clouds requires using the cloud in a more strategic fashion.
Firstly, a cloud-native solution should provide the option to backup data from the on-premises production storage implementation directly to the public cloud. This can help to mitigate—and in some instances potentially even to eliminate entirely—the need for on-premises infrastructure. Not only can reduction in on-premises infrastructure cut both capex and opex costs for the organization, it can also help to avoid hardware vendor lock-in. Meanwhile, cloud resources are inherently more flexible and scalable than their on-premises counterparts. IT professionals can more quickly, easily and cost-effectively expand their disaster recovery solutions as the needs of the organization change and grow.
The benefit of cloud-native VMware backup and disaster recovery
To summarize, a backup and disaster recovery implementation that uses the cloud strategically and natively radically cuts costs because it eliminates entirely the need for a secondary data center. Additional infrastructure does not need to be purchased, deployed, and managed. The amount of data center floor space, which is at a high premium today, that IT needs to procure is minimized. IT staff also can focus their time spent managing, testing, provisioning, and troubleshooting hardware on the production data center. They also have more time to devote to more revenue-facing initiatives for the business.
To enable customers to implement the cloud in this way, Druva invested in cloud disaster recovery, which was purpose-built to run in the cloud. Data can be backed up directly to, and it can be booted up directly in, the AWS-based Druva implementation. Data is tiered automatically across various classes of cloud storage services. To further cut the cost structure, we include applicationa ware compression and de-duplication. Additionally, Druva adds a number of capabilities that support security and business agility. The platform itself is agentless and integrates directly with vCenter, for “set it and forget it” automated virtual machine discovery, provisioning and load balancing. Declaring a disaster and failing over to a recovery state is as easy as pushing a button.