Part one of a two-part series examining these important cloud computing architectures.
As a technical offering manager with IBM Cloud, I find that it helps to establish common-ground with customers by understanding where they are in their cloud journey. In this post, let’s review the most common cloud-computing architectures:
- Hybrid cloud
- Distributed cloud
Hybrid cloud: Keep some applications on premises, experiment in the cloud
In my view, hybrid cloud is your choice if you want to move to a cloud native, cloud-based architecture without abandoning your existing applications. Essentially, maintaining legacy apps on-premises while taking advantage of the cloud to experiment on new applications. A simple use-case of hybrid cloud is exposing high-value assets that run on-premises to the cloud by using application programming interfaces (APIs).
The core benefits of using hybrid cloud include the following features:
- Ability to scale on demand
- Use of the latest cloud technologies, such as artificial intelligence (AI) and machine learning
- Ability to utilize cloud to modernize user experiences while continuing to run legacy assets on-premises
Multicloud: Enjoy more availability and flexibility
Multicloud allows you to spread your applications to several cloud environments and use more than one vendor. Here’s why multicloud can help you.
Say you’re an executive at a delivery company needing to scale for an increase in orders. You want to achieve this objective while also carrying out the following tasks:
- Use hybrid cloud to run on-premises
- Handle their additional load of data by using the cloud
- Avoid vendor lock-in for developing applications
- Be agile and keep up with the changing ecosystem
With multicloud, you can access multiple clouds and multiple data centers worldwide to accomplish all these activities. You gain availability and can do more workloads than hybrid cloud.
Consider multicloud to be a subset of hybrid cloud. At the same time, you can potentially use a hybrid cloud as one of its multiple cloud environments.
Multicloud also combats shadow IT, where certain employees prefer to work in a different cloud than what the enterprise uses. Shadow IT leads to a company adopting multiple clouds without overall guidance coming from the top level of leadership. Multicloud gives you better visibility and governance over shadow IT than hybrid cloud.
Distributed cloud: Use single control pane for cloud management and consistency
Gartner recently published a paper that said distributed cloud fixes what hybrid cloud breaks. I would argue that distributed cloud also fixes what multicloud breaks.
Let's take Kubernetes as an example. All major cloud providers support their own managed Kubernetes service, but the underlying technology is all essentially the same open source project. But, if you want to create a Kubernetes cluster with more than one vendor, you can encounter differences in the following areas:
- Data centers
- Ops teams with varying skill sets
If you use the latest versioning on IBM Cloud — which some of your other cloud providers don’t support — it may results in inconsistencies that your ops team will struggle to handle.
Some multicloud management solutions claim to have a single control pane to address this challenge. The truth is that these platforms allow you to control the operations of your Kubernetes clusters, but not fully manage the operations specific to the cloud provider where your cluster runs. For tasks such as changing access roles or security constraints, you’ll have to navigate to the individual ops dashboards of the various cloud providers.
In contrast, with distributed cloud, you can continue to use multicloud environments and access resources no matter where they’re located — all within a single control plane on one cloud.
You can use distributed cloud for multiple tasks, including the following activities:
- Creating and deploying Kubernetes clusters
- Monitoring the environment
- Making secure updates in that cloud
The following major advantages come through using distributed cloud:
- Consistency: One core location to manage and compute resources anywhere
- DevOps: The ability to deal with all of your clusters equally when deploying applications
- Governance: Compliance with regulatory concerns such as the General Data Protection Regulation (GDPR) by deploying a policy consistently across environments
Learn more about distributed cloud by reading "Understanding Distributed Cloud Architecture: The Basics."
In the second part of this blog, I’ll discuss how distributed cloud works with edge computing.