Increasingly more organizations are adopting cloud strategies that involve hosting IT infrastructure in multiple locations, including public clouds, private clouds, and on-premises environments. According to a recent study, this hybrid cloud approach has grown from 19% of organizations to 57% of organizations. In fact, Gartner predicts that hybrid cloud strategies will be the most common cloud deployments by 2020.
There are many questions organizations need to address as they develop their hybrid cloud strategies. Two of those key questions that this post focuses on include:
- Which workloads should we migrate to the public cloud versus keep in their current on-premises or private cloud environments?
- How do we migrate workloads to the public cloud in a way that doesn’t disrupt application connections or degrade performance?
Addressing these questions with accuracy and efficiency requires analyzing the right metrics. Here are 5 analytics you need to make confident and precise decisions to inform your hybrid cloud strategy:
Discovery and Dependency Mapping: What does your infrastructure look like? Having visibility across your entire on-premises and private cloud infrastructure is critical for a successful hybrid cloud strategy. You need to be aware of every application you have and how they are all connected. Conducting a thorough discovery and application dependency mapping process is the first step in planning what to move to the public cloud and how to move them in a way that doesn’t disrupt application connections.
Performance Analysis: What are your current application characteristics? To decide which applications to move to the public cloud versus keep on-premises or in the private cloud, you first need a detailed picture of the current performance profile of every application across your on-premises and private cloud environments. This includes CPU utilization, RAM usage, IOPs, and throughput. These metrics should be collected in five minute intervals over a period of at least two weeks and include peaks and valleys, not averages.
Hybrid Cloud Performance Benchmarking: Which applications will cost less and perform the same or better in the public cloud than they are currently? Hybrid cloud performance benchmarking enables companies to precisely project compute, storage, and network performance in the public cloud for each application based on current observed performance, available compute and storage options in the cloud, and predictive analytics.
With this prediction, it’s possible to compare an application’s current performance profile to its projected performance in the public cloud. You can identify suitable options for compute and storage in the cloud that enable you to meet application performance requirements at a lower cost than current on-premises or private cloud infrastructure. By the same token, these analytics can also enable organizations to determine whether certain applications should be kept on-premises because they may be either more costly or unable to meet performance requirements in the cloud. This process helps companies determine what to move to the public cloud versus what to keep on-premises and in the private cloud.
Usage analysis: Which applications will provide the biggest cost benefit by moving to the public cloud? Knowing how often compute and storage resources are idle, on, and unused will help you identify workloads that are best suited for migration to the public cloud. Those with bursty IOPS tend to be among the best cloud candidates because they enable you to take advantage of auto-scaling. When auto-scaling is enabled, you don’t have to worry about overspending by over-provisioning your resources.
Hybrid Cloud Connectivity: Are your migrated workloads operating in the cloud as they should be? The dependency mapping you completed earlier comes in handy here. Hybrid cloud connectivity tests validate that the application connections of your migrated applications are operating in the cloud as they were in their historical on-premises or private cloud environment. This analysis enables you to immediately identify and rectify any gaps in application connectivity in the cloud, and run follow-up tests to ensure the connections are working as they should.
A successful hybrid cloud strategy requires analysis of your current on-premises and private cloud infrastructure, as well as predictive analytics for the cloud. These insights are critical for knowing what to migrate to the public cloud, what to keep where it is currently, and to ensure public cloud migration in a hybrid environment goes smoothly.