Creating a hybrid cloud is much harder than organizations expect. Cloud infrastructures run different hypervisors than on-premises, which means it’s necessary to transform the workloads before they can run in the cloud, and then transform them back if the organization decides to migrate them back on-premises. Another problem is that a typical workload isn’t made up of just one server. In most cases, there are several virtual machines that need to be in place prior to a migration or disaster recovery. Additionally, these virtual machines need to start in the exact right order and need to have customizations done either to the VMs themselves or to the environment hosting these workloads. The result is that hybrid cloud recovery and migration is too complex for many organizations to support at scale.
How IT Does Hybrid Cloud Now
The first step in creating a hybrid cloud strategy is to decide why the organization is going to use the cloud. The most common answers, Cloud as DR, Cloud for Development/Testing (Dev/Test) and the cloud for bursting during peak loads. From a process perspective, each of these needs a written plan that executes whenever making data movements to the cloud.
The next step is to create a plan to move data to the cloud and keep that data updated as it makes sense but also not let those updates fall so far behind that the use case is operating on old data. At the same time, data movement to the cloud can’t impact on-premises performance. Additionally, IT should look to avoid building a completely new process and infrastructure to support cloud migration.
Ideally, the organization should look to leverage an existing data movement application like backup, replication or archive instead of implementing a new solution focused solely on that task.
The Importance of Automation
An important element in making hybrid cloud work for the organization is to leverage automation so that workloads can spin up quickly, in the correct order, with the right and most recent copy of data. Hybrid cloud is supposed to enable the organization to test new code quickly, burst to the cloud when peak loads occur and initiate a disaster recovery when the primary data center fails. Without automation, implementing a hybrid cloud becomes too time consuming to execute and too prone to error. Automation enables a push-button simplicity, which automatically transforms workloads to the cloud’s native hypervisor and starts virtual machines in the right order.
Introducing Cohesity Runbook
Vendors are aware of the challenges facing customers looking to create a hybrid cloud strategy and several are adding automation tools to their platforms to accommodate customer needs. One of the most unique is Cohesity’s Runbook. Runbook is available on the Cohesity Marketplace (Cohesity’s version of an App Store) and runs on the Cohesity DataPlatform. We’ve covered Cohesity for the past four years, with initial coverage starting in 2015 in our briefing note, “Stopping Secondary Storage Sprawl.” Cohesity also appeared in our Lightboard Video, “Managing Secondary Data Globally.”
Cohesity Runbook is a natural extension of the Cohesity GUI providing a seamless experience for administrators. It presents administrators with a canvas that feels similar to Microsoft Vizio, enabling them to drag, drop and draw their cloud data movement workflow. Runbook also validates the workflow to make sure it will execute correctly, and executes the workflow when necessary. Runbook streamlines complex tasks and administrators can create workflows for individual applications or the entire environment. Within the workflow, they can manage dependencies, timings and insert their own scripts for more detailed customization. It can also orchestrate the conversion between VM formats. It validates all of these steps before execution, improving confidence that test and actual executions will work as planned.
The primary focus of Runbook is to aid Cohesity customers with disaster recovery and hybrid cloud workload mobility, but future plans are to automate dev/test and to support containerized applications. Cohesity has a variety of ways it can leverage the cloud as we discussed in our briefing note “No Room in the Data Center – Secondary Storage Needs Cloud Integration.”
Runbook currently supports AWS with plans to support other major cloud providers.
In our opinion, the merits of using the cloud for the long term storing of data is debatable but leveraging cloud for more temporal use cases like disaster recovery, dev/test and cloud bursting is not. Cohesity has several methods of leveraging the cloud which benefit from Runbook automation. Runbook solves this problem and does so in an elegant fashion. The solution is an obvious choice for existing Cohesity customers and may be the final motivation that organizations interested in Cohesity need to decide to switch.