For many enterprises, the cloud migration journey is becoming more complex, operationally speaking. As organizations move more applications, workloads and infrastructure to the cloud, they’re likely to be using services from multiple public cloud providers; 55% of organizations in IDG’s 2020 Cloud Computing Study currently use multiple public clouds.
Multi-cloud environments provide a multitude of benefits, including access to best-of-breed platforms and services and greater flexibility. But there are obstacles to this approach, in the form of increased complexity and the potential for increased costs.
As multi-cloud environments grow, IT teams will need to address these challenges to ensure their investments are meeting broader business objectives. We asked our community of B2B technology influencers for their thoughts on the best way to manage and optimize workloads across different public cloud providers in a multi-cloud environment. We grouped their responses into three main takeaways.
1. Get your priorities in order
IT leaders who have embarked on a cloud modernization journey understand the importance of prioritizing which applications and workloads to move from on-premises environments to the cloud. It’s equally important to make sure you’re choosing the right cloud service provider (CSP) for each of the applications or workloads you’re migrating.
“As part of a cloud migration strategy, establish a cloud decision framework to provide a methodical approach, based on workload characteristics and cloud platform capabilities, to guide workload placement in a multi-cloud environment,” said Sook Chua, Ernst & Young LLP Principal, Technology Consulting.
Jack Gold (@jckgld), President and Principal Analyst at J. Gold Associates, LLC, said the multi-cloud consideration set should include which services are available on each cloud instance (e.g., frameworks, app stacks from the various cloud providers, third-party availability as a service) and specialized hardware acceleration needs for specific workloads (e.g., artificial intelligence, machine learning, natural language processing, image processing).
Because different use cases can vary by cloud provider, it’s wise to take the extra time and steps to map workloads to the different CSPs, said Will Kelly (@willkelly), technical marketing manager for a container security startup. For example, Kelly said, for cloud storage, look for a CSP that offers the best pricing, while analytics workloads may be best served by a CSP with the best artificial intelligence stack.
Steve Prentice (@StevenPrentice), a technology integration specialist, offers some additional criteria for assigning applications to cloud platforms, including security, cost, practicality, mission critical and urgency, uniqueness and location. As IT teams work through these priorities, they may find that some workloads and applications, and the data they contain, may be better off spread across multiple CSPs. “Distribute your data across multiple providers to reduce risk and maximize redundancy, security, and resiliency,” he said.
2. Make it manageable and measurable
Because cloud services can be deployed easily across different parts of the business, it’s important for IT teams to take a holistic approach to managing a multi-cloud environment to keep costs in check. Monitoring and measuring performance across cloud services both are critical to ensuring optimal results.
“You can’t manage something you haven’t measured,” said Frank Cutitta (@fcutitta), CEO and Founder of HealthTech Decisions Lab. “So the first step is to do an audit of the components of the workloads [physical, technological, and economic]. The second step is identifying similarities of these workloads despite the cloud environment they are operating. The third step is to determine whether there is a way of streamlining or consolidating the cloud environments by redistributing similar security needs and processes across the most efficient cloud environments.”
To further optimize workloads between different public cloud providers, “it is very important to have a breakdown of the usage and specific metrics for applications or services that are going to be hosted on these resources,” said Martha Cisneros (@marthacisneros), an IT engineer and technology consultant. Also critical, she said, is an understanding of the provider’s data protection policies.
3. Orchestrate and optimize
Cloud orchestration — the process of integrating automated tasks and processes into consolidated workflows — is an important step in optimizing workloads across private and public clouds. “Using cloud orchestration to wrangle it all into a manageable [single] workflow is the key to optimization,” said Gene De Libero (@GeneDeLibero), Chief Strategy Officer at GeekHive.
Integrating operational data with cloud cost data will uncover opportunities to improve performance or reduce costs, said Isaac Sacolick (@nyike), President of StarCIO and author of Driving Digital: The Leader’s Guide to Business Transformation Through Technology. “The best way to optimize workloads is by establishing observability and monitoring standards, leveraging AIOps to correlate operational data, and automating service management workflows using single-pane platforms that work across public clouds,” he said.
Effective cloud optimization also requires steps to define KPIs, establish governance and drive accountability across IT teams, said Chua, from Ernst & Young LLP. “Cloud optimization is not a one-time activity, but rather a continuous process, relying on observability and automation to continuously discover, identify and optimize,” she added.
A final piece of advice: choose your providers wisely for initial deployments, because moving applications from one platform to another can be a big challenge. “In practice you can’t lift and shift applications and infrastructures from one cloud to another unless you’ve designed them to do so and accepted the trade-offs you need to use to support that portability,” said Simon Bisson (@sbisson), a technology journalist. “The future may well be different … with a combination of Kubernetes and containers for large-scale applications.”
As digital acceleration transforms industries and customer and employee expectations, CIOs have a vital role in deploying advanced technology and data to integrate functions and enable new business models seamlessly. Ernst & Young LLP Consulting teams with technology leaders to drive their agenda by delivering tech at speed and innovation at scale, while putting humans at the center.
For more information about how Ernst & Young LLP can help you unlock long-term value for your stakeholders and thought-provoking content for technology professionals visit ey.com/CIO.