buchachon - Fotolia
It's rare to find an IT department that relies on one technology or environment to support operations. As organizations switch to using more cloud-based applications, IT administrators must manage hybrid and multi-cloud infrastructures.
"It used to be that we controlled everything through beautiful bespoke systems of infrastructure components and technologies to create the perfect solutions. … Now we have this distributed monstrosity in some cases," said Angelina Troy, senior director at Gartner, during her presentation "Technical Insights: Managing IT in a Hybrid and Multi-Cloud World" at the company's Infrastructure, Operations & Cloud Strategies Conference.
A multi-cloud strategy helps organizations reduce vendor lock-in, support business continuity, mitigate risk, streamline transitions during mergers and acquisitions, and ease Microsoft Office 365 integrations. It's also something that generally occurs over time as organizations acquire more technology and software tools.
"The thought process behind [multi-cloud] happens organically," said Amr Ahmed, executive director of infrastructure and service resiliency, U.S. at EY. "You don't want all of your eggs in one basket."
With multi-cloud and hybrid IT, admins have four main levels to oversee: infrastructure, cloud/virtualization, applications and management. This covers network and storage hardware, workload bursting, virtual machine mobility, data integration, cost control, security governance and capacity management.
Overlay management tools help impose order, especially as organizations move toward a service-oriented or product-focused deployment model.
Moving to a service-oriented operational model enables organizations to optimize cost and industrialize its services, Troy said. This helps IT departments get out of siloed, pipeline thinking that primarily focuses on technology, instead of how an organization's infrastructure can help business operations.
Designing hybrid and multi-cloud architectures
When it comes to hybrid IT management, there is a spectrum of setups: One end is use-case-based provider silos, and the other is composite multi-cloud systems, Troy said.
Organizing supporting infrastructure or cloud technology by use is often the starting point for most organizations because it can be quicker to get up and running, it provides more cloud service choices and it doesn't require admins to combine applications.
On the other end of the spectrum, composite multi-cloud facilitates highly available infrastructure, provides cost transparency, lets users take advantage of high-level service application components and enables workload portability.
Ahmed added that organization leaders should decide why they want to run workloads in the cloud, and where organizations should run specific workloads.
For example, an IT department can run production workloads in AWS or back up data on VMware offerings. This planning should also include cost considerations because moving off premises to a hybrid or multi-cloud setup may also require a shift in financial practices.
"When it comes to infrastructure as a service, just because it is cost transparent doesn't mean that it is cost-effective," he said.
Structuring multi-cloud and hybrid IT management
Management tools are an essential part of IT operations; regardless of whether an organization already uses hybrid IT and multi-cloud architectures or is in the process of setting one up.
"It's rare to find [a management tool] that will cover everything," Troy said. "You don't have to integrate everything. Focus on what you need to get the job done."
The two main options to go with for multi-cloud management are central management or distributed management centers. Organizations may choose a model depending on department structure, current software tools, licensing terms and long-term goals.
With a centralized tool set, admins have a consistent view of infrastructure across all cloud instances. They can let operations focus on operations tasks, and there is a lower learning curve for line-of-business users.
Benefits of distributed management include not requiring a centralized support team or new management tools, quicker implementation and more software options. However, it does mean that there may be team silos and tool fragmentation.
There are plenty of options from VMware, AWS and HPE that place a layer over the production environment and help provide metrics that admins can use for daily, monthly and yearly management needs, said Ahmed. Using these offerings can reduce the back and forth admins must do to get essential information.
"You cannot rely one tool for management and another for production," he said.
Governance policies are another key component to hybrid IT management. Organizations can use programmatic controls, descriptive policy or fall somewhere in between depending on the use case or software application needs, Troy said.
With programmatic controls, organizations decrease security breach risk, potential cloud resource overspending and the chance of resource loss. These controls also enforce any compliance requirements, which makes them necessary for high-risk or mission-critical workloads.
For less critical scenarios, a descriptive policy gives users a high-level understanding of any procedures, requires less implementation planning and doesn't necessarily need code work or APIs to set up; these policies are also much easier for end users to grasp and are easier to change when needed.