How to Approach Multicloud Networking

Cloud computing has turned the world upside down in terms of how infrastructure requirements are gathered and correlated, and how new deployments are delivered. Infrastructure changes that used to take many months can now be done on demand in minutes. And most organizations—94%, according to 2023 research from Enterprise Strategy Group[1]—are working with multiple clouds. The time has come to be intentional about architecting the networks that support workloads across increasingly diversifying environments. You have to effectively address multicloud networking to avoid risking poor application performance, inefficient use of resources or negative user experiences.

But multicloud networking can be quite complex, both technically and from a business point of view. Various stakeholders and teams need to be aligned: infrastructure, architecture and networking teams, as well as management and line-of-business leaders. There’s more than one way to think about how to approach multicloud networking. Depending on the needs of your business and how your infrastructure was designed, your approach could be driven by either specific workloads or specific infrastructure categories.

We’ve already covered some best practices for multicloud networking. In this blog post, let’s talk about how to solve for multicloud networking from both perspectives—workload-centric and infrastructure-centric—and share a few examples of how you might approach your multicloud network design.

Multicloud networking to address specific workloads and applications

Chances are, you’re not tackling multicloud networking for its own sake but to solve specific cloud-to-cloud networking challenges in your business. These workload requirements may drive how you solve for multicloud networking.

A workload-specific approach could include:

• Transaction processing
• Front-end or customer-facing applications
• Systems of record

Let’s look at a few possible scenarios where you’d approach multicloud networking to address specific workloads:

Example 1: When agility and security are paramount

Let’s say a global retailer has a very popular web-based customer-facing application with short development cycles and—because it’s popular and web-based—a large attack surface. It’s well known that consumers have very high expectations for their apps, so the retailer’s application needs to perform well to deliver an optimal customer experience.

In this scenario, the multicloud networking solution needs to be very agile and incorporate equally agile security functions that can keep up with the pace of new releases and the company’s global expansion. The retailer’s requirements for agility and adaptability would likely steer it toward a multicloud networking solution that uses virtual networking services that deliver the most flexibility, with on-demand deployment of networking resources adjacent to their diverse multicloud environment and other infrastructure.

Example 2: When throughput is your top priority

Let’s look at a different workload-focused example. Say a manufacturing company has a digital twin of its facilities that it’s using for predictive maintenance. To support the digital twin, the manufacturer has a non-customer-facing backend workload that combines a fast-growing data repository in one cloud and analytics in another.

While security is easier to handle in this scenario since the app isn’t customer-facing, throughput is an absolute priority. Therefore, the high bandwidth requirements are likely to steer the manufacturer toward physical networking resources, combined with physical cross connects, capable of consistently providing the highest possible performance and delivering stable, predictable latency.

Multicloud networking within your existing infrastructure

The above examples illustrate how specific applications or workloads might determine how you do multicloud networking. For other organizations, your approach might be based on your enterprise architecture. Specific infrastructure components like your existing compute, storage and network infrastructure can influence your direction.

In fact, it’s not uncommon for infrastructure to put some hard limitations on how an organization can address multicloud networking. After all, infrastructure is frequently built up prior to, or without regard for, all the workloads it might need to eventually support. In such cases, you may want to address multicloud networking from an infrastructure-centric point of view.

With this approach, your multicloud networking needs to take several questions into consideration:

• Where are your compute resources deployed? Are they in multiple clouds, on premises or in a colocation data center?
• Where do you currently store the data that’s relevant for your cloud-hosted workloads and applications?
• How do you expect any of the above to change in the foreseeable future?

The answers to these questions could lead to different decisions around how to do multicloud networking. For example, you may have some resources in the public cloud, while some remain in a private, on-premises data center. Most organizations have gravitated toward a hybrid multicloud architecture that combines the strengths of each environment to address their diverse requirements. It’s also becoming more common for companies to repatriate some data from the cloud into off-cloud environments because of high data egress fees resulting from storage in the cloud. Dedicated cloud adjacent storage makes it faster and easier to connect with public clouds while maintaining a high degree of control over your stored data.

Whatever multicloud networking solution you choose today, you’ll want it to still make sense in the future, as your infrastructure evolves. You should take an approach that will give you the most flexibility, allowing you to accommodate a hybrid multicloud architecture and adjust easily when things in your environment change.

Building the multicloud network you need

One thing we know is true: You need flexibility and choice around how you design your multicloud network. The complexity and demands of multicloud networking require problem solving and innovative approaches—and Equinix has been at work doing just that. We’re invested in delivering creative solutions to the requirements of multicloud connectivity.

We can’t make the complexity of multicloud networking disappear, but there are lots of creative ways to solve for it using the building blocks available on Platform Equinix®. We can help with a wide variety of cloud-to-cloud networking scenarios for all public and hybrid cloud environments, providing the right option for your current and future needs. With Equinix, you get the performance and control delivered by physical routing, the adaptability and cloud-like experience delivered by virtual routing, and the ability to extend either one with on-demand security functions and access to our vibrant ecosystem of service providers and partners.

Equinix has continued to evolve to deliver the tools and capabilities that allow you to implement robust cloud-to-cloud networking capabilities that can be highly tailored to your requirements:

• You can deploy virtual network functions (VNFs) like SD-WAN devices, routers, firewalls and load balancers from leading vendors on demand using Equinix Network Edge.
• Or use our new virtual routing capability—Fabric Cloud Router—within Equinix Fabric® for essential cloud-to-cloud routing.
• Equinix Fabric can be used as a virtual metro-based switching platform if you want to build multicloud diversity with two physical locations in a metro.
• With Equinix Fabric, you can also build a virtual connection to transparently carry MACsec traffic from port to port, to deliver a highly secure front and back end.
• And Equinix Precision Time™ can help you unify timing on your systems of record and transaction processing.

This more flexible approach offered by Equinix is quite different from what’s available in the clouds and won’t lock you in to a single cloud vendor. With Equinix, you get a new level of agility since you not only have options for what kind of networking you choose; you can also deploy infrastructure solutions faster and make changes dynamically.

Learn more about multicloud networking by downloading the Enterprise Strategy Group report Streamline Multicloud Networking.

In case you missed it, check out the other blog posts in our mini-series on multicloud networking:

What Is Multicloud Networking? Why Is It So Important?

8 Best Practices for Multicloud Networking

[1] Paul Nashawaty, Infographic: The Mainstreaming of Cloud-native Apps and Methodologies, Enterprise Strategy Group, August 11, 2023.