Leveraging Virtualization for Improved Availability and Disaster Recovery
Today, the vast majority of production workloads are run on virtualized infrastructure and most businesses have virtualized their local infrastructure using either VMware’s vSphere or Microsoft’s Hyper-V. VMs enable you to abstract your workload from the underlying server hardware. This degree of abstraction enables greater flexibility, efficiency, and availability than you can achieve running on bare metal. You can combine multiple workloads on a single physical server vastly reducing the total number of servers needed in your infrastructure. Fewer servers require less rack space, less power, less cooling and less networking equipment resulting in significant CapEx and OpEx savings.
Virtualization and Availability
Virtualization enables you to increase the availability of your workloads by reducing the planned downtime that accompanies system maintenance and patching. Virtualization also enables you to better meet your SLAs by dynamically moving your workloads to servers which have more available processing power.
Technologies like Live Migration or vMotion enable you to move VMs between hosts without incurring any end-user downtime. This improves availability by enabling you to seamlessly move all of the VMs on a physical server to another available host while you perform maintenance on the server. Likewise, Hyper-V’s dynamic load balancing and VMware’s Distributed Resource Scheduler (DRS) improve application availability and responsiveness by monitoring and balancing host resource usage according to administrative policies.
Virtualization and DR
Even in the age of virtualization, backup is still the core of all disaster recovery (DR) plans. While backups play an important role in preserving your data, backups alone do not make up a modern DR plan. Backups can certainly be used to recover your system in the case of a server or site failure. However, restoring from a backup will also usually entail data loss between the time the backup was taken and the time the restore process is enacted. Plus, system restores can take a considerable amount of time.
Today, with the use of virtualization for DR you can significantly reduce the time to recover a critical workload as well as the requirements for redundant hardware. Virtualization’s server consolidation capabilities vastly reduce the hardware required to implement a DR strategy. Technologies like Microsoft’s Hyper-V Replica or VMware’s vSphere Replication enable you to leverage the VM’s mobility by periodically replicating the changes made to a VM’s virtual hard disk to a VM in a remote DR location or to the cloud.
The replication process can happen as frequently as every few seconds ensuring that the remote VM’s state is very current. The remote DR VM replication target can then be powered on and used in the event of a failure in the primary system. Leveraging virtualization and a virtualized DR site enables you to vastly shrink your Recovery Time Objectives (RTOs) and your Recovery Point Objectives (RPOs). While virtualization replication does improve your RTOs and RPOs it is important to remember that it is not a substitute for backups. Backups can be used for archival purposes where you might restore an object from a backup that was taken weeks or even month’s ago. Replication is not intended for long term data archival. Both are required for a complete DR strategy.
While there’s no doubt that the use of the cloud and containers are rapidly growing, the reality is that server virtualization is and will remain the core technology for running the vast majority of enterprise applications in the foreseeable future. The advantages that it provides for availability and DR are essential for today’s business-critical applications.