Multicloud adoption is slowly becoming a norm in today’s digital world and organizations are increasingly using hybrid cloud from multiple vendors to deploy their applications. According to various cloud surveys, less than 40 percent of the respondents uses a single cloud provider, while significant percentage of enterprises, use or in the process of utilizing two cloud providers. Multi cloud adoption is helping organizations avoid vendor lock-in, meet geographic performance requirements, meet regulatory requirements, and exploit best-of-breed solutions available in the market. The competitive price war is also bringing in some cost advantages along with risk mitigation options for the digital enterprise inclined toward multi-cloud adoption.
“Multicloud adoption helps organization to achieve resilience, avoid vendor lock-in and take advantage of best-of-breed service in the marketplace”
These rationales supporting multicloud, though valid, bring along a number of challenges from the operation perspective and hinders adoption. At the top, is the concern regarding monitoring and effective deployment across multi-vendor platform. But shift toward cloud native development, containerized hosting, and deployment automation are enabling organizations to deliver a portable solution that is DevOps-driven and free from vendor lock-in, thus removing one of key impediments of multi cloud adoption.
Let us first define multi and hybrid cloud as both are used interchangeably at times. In hybrid cloud, an organization uses a combination of public, private cloud, and on-premises services to achieve its business requirements with some degree of workload portability, integration, and orchestration between them. In multicloud, an organization uses multiple cloud services from more than one provider say Amazon, Microsoft, or RedHat to deliver a specific application or service and has more strategic emphasis on vendor lock-in, cost, exploiting best-of-breed solutions, and adherence to local regulations.
Multicloud Drivers and Challengers
All cloud providers (inclusive of AWS, Azure, GCP) even with multiple geographically dispersed, redundant data centres suffers outages from time to time. Enterprises understands the business risk and are betting on multicloud to achieve resilience.
One of the most widely cited drivers of multi-cloud adoption by many organizations is the desire to avoid getting locked into one cloud provider's infrastructure and their pricing model.
Though all cloud providers offer the essential compute, storage, and networking services, the desire to utilize best-of-the-breed services is driving multicloud adoption. For instance, while AWS service offering is growing day-by-day, covering the entire gamut of IT operations, GCP offers number of specialized services related to data analytics and machine learning.
Multicloud adoption bring along number of challenges from operation perspective, including monitoring capability, managing networks, and seamless deployment capability. It also brings additional challenges related to cost, right skill resources, security, and compliance as a result of bringing multiple vendor platforms into the existing IT landscape.
Most organizations have in the past invested in tools for monitoring their on-premise data centres, but these legacy tools now turn out to be misfit and fall short of providing the unified view. An enterprise faces challenges in workloads monitoring and in understanding the cost implications when multiple vendors are involved.
Most organizations also find it difficult to integrate and connect the multicloud environment to a single entity for network management as these isolated environments are subject to security breaches, cyber-attacks, and vulnerabilities.
Multicloud introduces additional complexity when it comes to application delivery and reflects the importance of standardization on tools and processes. An enterprise faces difficulties when it comes to moving an application between clouds as it sometimes involve significant effort for application reconfiguration and sometime necessitates different sets of tools and practices.
Beacon of Hope for Rapid Adoption
The recent developments in various vendor offering in the monitoring tool area and software-defined networking (SDN) are driving the multicloud adoption. The container platforms are becoming the beacon of hope when it comes to addressing the portability challenges.
As different cloud solutions run in different software environments, containers and container orchestrations became the ideal enablers of multicloud deployments because they isolate software from the underlying environment and empowers developers to build applications that can be deployed and scaled essentially anywhere.
Containers based on immutable image with runtime dependencies resolved within the image, facilitates portability while Kubernetes facilitates orchestration and scheduling and helps in running containers at scale. With the growing support for containers and Kubernetes by all major cloud providers including Google, Amazon, Red Hat and Azure, developers will be able to deploy and run portable solutions, regardless of chosen environment for deployment.
With the availability of ‘off-the-self’ container platforms such as Red Hat OpenShift container platform, which is based on enterprise Kubernetes solution and works with all major public cloud platforms including AWS, Azure, GCP, the adoption of multicloud will witness rapid-growth scenarios as they facilitate truly portable environments and ease the workload monitoring challenges for the IT teams.
On the deployment automation front, provisioning pre-defined environments on demand was a challenging ask, but with the concept of managing environment in the same way as applications by describing the entire infrastructure-as-code (IaC) unlocked the door to complete automation. With the emergence of multi-cloud continuous delivery platform, organizations are now equipped with the necessary capabilities to carry out faster and reliable rollouts across different cloud platforms.
The container platforms promise to provide an environment with execution, orchestration, integration, security, management, and deployment automation capabilities designed to provision and control container clusters across multiple cloud infrastructure platforms. However, enterprises still need to be cautious of the fact that offerings in the market are relatively young and likely to see innovations from different vendors. Hence, it is always recommended for organizations to take an incremental fail-fast approach based on ‘pilot-adopt-accelerate-sustain’ models.