Conal Gallagher is the CIO & CISO at both Flexera and its division Revenera, where he manages IT and Information Security programs.
getty
Think about the last time you moved to new living quarters. Maybe you left college for a first apartment, moved to a new city or settled into a larger home to meet your family's needs. How well prepared were you? What could have been easier?
Moving to a new home is usually exciting, but the logistics get complicated. You need to hire movers — or rent a truck and hope that some pizza-loving friends will help. You carefully label boxes so that they land in the right spot in the new space — or awkwardly figure out where a hamper of random clothes and household items go. Then you set up new utility accounts, forward mail and identify where the neighborhood schools and stores are. You hope it all happens with as little disruption to your schedule as possible.
Just as a well-planned move is often the smoothest, the same holds true for cloud migrations. As you're packing up your data and workloads to transition business services to the cloud, consider three best practices for cloud migration planning to ensure an efficient migration:
1. Evaluate what you need in your new space.
Certain things will fit into your new space; some should be thrown out. There's no need to spend time, money or space on items — whether an old couch or outdated application architecture — that no longer fit your usage or storage needs.
A lift-and-shift move to the cloud may be appropriate for certain applications, but likely not ideal for the entire migration. Rearchitecting may be necessary. Here a few tips to avoid common challenges:
• Conduct a thorough cloud migration technical assessment. This helps identify if and where rearchitecting will be advantageous, the amount of effort required and the most effective order for migrating workloads to help meet business goals (e.g., cost savings, delivery of products and services).
• Identify performance requirements and the potential bottlenecks that may occur if an application is migrated to a new environment. On-premises environments and applications may require hardware changes to improve application performance. If premium storage systems with high input/output operations per second (IOPS) rates were implemented to improve database and/or filesystem performance, or if service level agreements (SLAs) were developed based on observed performance characteristics, these performance attributes must be considered in the move. Comparable IOPS rates may not be available — or perhaps they are, but not at an acceptable cost.
• Evaluate legacy applications that depend on outdated operating systems or tech stacks. For example, HP-UX and AIX are still encountered in the wild, but cloud-based options for these operating systems are hard to find. Those using such systems must, in the assessment phase, review applications not only for functional requirements and interdependencies but the underlying technologies, with the goal of ensuring a compatible target environment.
2. Know what goes where and why.
Like having boxes labeled correctly streamlines a physical move, knowing applications' destinations eases cloud migrations. Yet understanding application dependencies is commonly the most significant challenge. Is that application connected to log collection software or shared reporting services? Does it have another dependency that impacts multiple IT-delivered business services?
Most existing applications are integrated with, and dependent on, other applications within the IT environment. Failure to understand interdependencies connected to an application (and its infrastructure and component software) is a common pitfall. Migrating a single application may disrupt the functionality of the service(s) these interconnected applications deliver or introduce unacceptable latencies.
At enterprise scale, manual mapping efforts aren't practical. A configuration management database (CMDB), which can provide an application view, often isn't adequate, either; CMDBs are frequently outdated, incomplete and don’t identify business services interdependencies.
A full automated dependency mapping of the IT environment — prior to any migration — can help overcome this challenge. This identifies interdependencies (across on-premises, SaaS and cloud assets) and enables evaluation of which applications within the service can or can't be migrated, helps your organization make informed choices about what to migrate and facilitates ongoing cloud cost optimization.
3. Keep tabs on expenses.
Installing a smart thermostat in your new home helps keep the temperature comfortable, while eliminating waste from heating and cooling bills. Similarly, new tools and behaviors can help govern complex multi-cloud architectures and optimize costs.
According to the IDG 2020 Cloud Computing Study, 51% of respondents identify "cost management, reporting and optimization" as the top reason for selecting a multi-cloud management tool. But new tools must be supported with new practices. Organizations with legacy on-premises environments must discard old habits that aren't optimal for cloud. Her are a few practices worth reevaluating:
• Maintaining your own infrastructure. This incurs fixed costs (building leases, power, cooling, backup equipment, etc.) that don't vary based on the workloads running. As a result, IT professionals often aren't motivated to shut down resources when not in use. In the cloud, by contrast, you pay for resources that are operational, whether in use or not. The on-premises behavior of leaving resources running isn't cost-efficient in the cloud. Instead, automated cloud cost optimization policies (e.g., shutting down workloads after hours, rightsizing cloud instances, specifying expiration dates, using the lowest cost clouds and/or regions and maximizing provider discounts) are valuable approaches.
• "Buying the peak." IT professionals commonly request enough capacity to meet future needs — often much more than required immediately — as part of a "tech refresh cycle," in which new equipment is purchased every three to five years to replace outdated infrastructure. In the cloud environment, resources are theoretically infinite; they can be added (and removed) as needed. The paradigm shifts from buying the peak to"buying the steady-state and renting the peak."
With global end user spending on public cloud services forecast to reach $332.3 billion in 2021 (up more than 23% from last year) according to Gartner, optimizing cloud spend is and will continue to be an important part of cloud initiatives. Learning new behaviors can help achieve this goal.
Moving into a clean, well-organized, welcoming new home is a time of great opportunity. So is the move to the cloud.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Follow me on LinkedIn. Check out my website.