TL:DR
- Public internet creates unpredictable routing, latency issues & surprisingly high egress fees that limit multicloud applications & cloud-to-cloud migrations.
- Private interconnection via virtual connections enables direct cloud-to-cloud data paths with predictable performance & enhanced security controls.
- Organizations improved scalability and reduced cloud costs by dynamically allocating resources through intelligent workload distribution, ensuring optimal performance during peak demands while maintaining cost-effectiveness.
Today’s IT leaders don’t need to be told the benefits of hybrid multicloud. They’ve learned from experience that no single cloud can be as good as all the clouds. They also know that to take advantage of services from different cloud providers, they need to avoid vendor lock-in and maintain control over their cloud data.
While the value of hybrid multicloud is clear, businesses often find it difficult to get the secure, high-performance connectivity they need to capture that value. They need networking solutions that can move data across multiple clouds and on-premises environments whenever the need arises, including:
- Moving small datasets regularly to enable hybrid multicloud applications
- Moving one very large dataset as part of a one-time cloud-to-cloud migration
Either way, the right networking solution can make or break a hybrid multicloud strategy.
When enterprises choose the public internet to move cloud data, they’ll run into a number of challenges, as we’ll explore below. A direct, private interconnection solution can help overcome these challenges. For most enterprises, it simply doesn’t make sense to pursue hybrid multicloud without interconnection.
The internet isn’t always the best choice
Some businesses move cloud traffic over the public internet using VPN tunneling because it seems like the easiest and most cost-effective networking option. However, there are several issues that make the public internet less than ideal for moving cloud traffic, including:
- Unpredictable routing and latency
- Throughput issues
- Potential security concerns over the public internet
- Surprisingly high egress fees, particularly in the case of large-scale data migrations or other use cases that involve moving large datasets consistently
As cloud traffic moves over the internet, it doesn’t follow the most direct route from Point A to Point B, making it impossible to ensure optimized performance. For this reason, the internet is a poor choice for any multicloud application that needs to provide a reliable user experience. High latency can leave users stuck waiting for an application to respond, inevitably causing frustration. The diagram below illustrates how cloud-to-cloud routing over the internet leads to an indirect path that limits performance.
The internet also isn’t a good choice for any cloud workloads with strict data privacy and sovereignty requirements. It’s an inherently public transmission medium, so it’s impossible to keep data fully protected against leaks or breaches. In addition, the internet doesn’t offer the control and observability needed to ensure that traffic stays within specific borders. Thus, any business operating in a heavily regulated industry or in a region with specific compliance requirements should think long and hard before using the internet for cloud traffic.
Interconnection enables a better way
Many organizations have learned that just because they start out using the internet for cloud connectivity doesn’t mean they have to stay there. When they realize that the internet is no longer good enough, they’ll likely turn to a private interconnection solution instead.
Interconnection means a dedicated IT traffic exchange between two points—such as two different cloud environments. Network on-ramp nodes from different cloud providers can often be found inside the same colocation data centers, thus allowing data to be moved via the most direct route possible. Also, since that data doesn’t have to share the connection with any other traffic along the way, interconnection offers better performance than the public internet by definition. It also provides more control over the route that data takes to reach its destination, which can help with meeting data sovereignty requirements.
To meet the needs of different networking use cases, interconnection is available via physical or virtual connectivity. While physical connections are ideal for exchanging consistently high levels of traffic with particular partners in specific locations, virtual connections are better for situations where networking needs are more unpredictable. Multicloud networking is one area where the flexibility and scalability of a virtual interconnection solution could prove invaluable.
After they’ve deployed virtual interconnection solutions, many enterprise leaders realize that multicloud networking isn’t just something they have to do. It can become a true driver of strategic advantage for the business:
- They can move data across clouds and on-premises environments quickly, thus enabling a consistent user experience for hybrid multicloud apps. For instance, they may want to run a customer database in one cloud, analytics tools in a different cloud, and sensitive processing on-premises. Those three pieces can function as a holistic application without high latency standing in the way.
- When their cloud strategy changes, they can execute a one-time migration between clouds. Even if they’ve been using the first cloud environment for years and have built up a massive backlog of data, they can get the necessary high-bandwidth connectivity to move everything quickly. Thus, they can avoid having to operate in multiple clouds simultaneously just because the migration isn’t finished yet.
Interconnection in action: Multicloud networking customer stories
For years, enterprises have used Equinix Fabric®, our Network as a Service offering, to quickly stand up private virtual connections. However, many of our customers wanted a solution that was optimized specifically for cloud use cases like hybrid multicloud applications and cloud-to-cloud migrations. In 2024, we introduced Equinix Fabric Cloud Router, a virtual routing solution available as a built-in component of Equinix Fabric, to meet this demand. The diagram below shows how Fabric Cloud Router connects cloud on-ramps from different providers that are physically colocated inside the same Equinix data center, thus shortening the path that cloud-to-cloud traffic travels.
Already, many Equinix customers are using Fabric Cloud Router to unlock the full value of multicloud and cloud-to-cloud networking. Let’s look at a couple of examples.
Enabling multicloud at Pexip
Pexip, a Norwegian software company specializing in videoconferencing and collaboration solutions, hosts elements of their platform in both Microsoft Azure and Google Cloud. The company has used Equinix Fabric in the past, but saw an opportunity to further improve their customer experience by using a dedicated virtual router as the primary method of connecting the two clouds.
The company was familiar with the unique benefits Equinix offers, including our global reach and market-leading portfolio of cloud on-ramps, with more than twice as many on-ramps as our nearest competitor. They also appreciated the fact that Fabric Cloud Router is available for a fixed monthly cost, unlike cloud native solutions that charge egress fees based on the volume of data moved.
During an extensive testing phase, the company experimented with using infrastructure automation tools like Terraform, and eventually built an automated provisioning template. When they transitioned from testing into production, this automation made it easy to deploy Fabric Cloud Router in multiple locations in less than two days.
As a direct result of deploying Fabric Cloud Router, Pexip improved application performance and customer experience. They achieved this without having to purchase or host their own physical networking equipment.
Completing a massive cloud-to-cloud migration at Uber
Uber, the world’s largest ridesharing company, decided to move from one public cloud provider to another. The challenge was that they had accumulated 6.5 petabytes of data in the first cloud, and this massive dataset needed to come with them when they moved.
As they evaluated their networking options, Uber wanted to avoid the costs and complexity that would undoubtedly arise from moving data over the public internet. They also needed to complete the migration quickly, to avoid paying for storage on two separate clouds at the same time. Finally, they needed a solution that wouldn’t impact the performance of routine production traffic running on their existing network backbone.
With Fabric Cloud Router, Uber was able to achieve all these goals. Since the solution is available as a service, they were able to spin it up in less than a day. This allowed them to meet their aggressive migration timeline, and then scale bandwidth back down once the migration was complete.
See how interconnection can optimize multicloud
Fabric Cloud Router is designed to help you avoid the challenges that would otherwise impede your hybrid multicloud strategy. Without the on-demand multicloud connectivity that Fabric Cloud Router provides, a cloud is just a cloud, isolated and unable to live up to its full potential.
The solution is available everywhere Equinix Fabric is—in 60+ strategic metros worldwide. This includes 39 metros that offer access to two or more clouds and 12 metros that offer access to five or more clouds. In addition, we recently announced Fabric Cloud Router Basic Package, designed to support smaller workloads while delivering the same powerful features.
Take the first step toward realizing your multicloud potential: Read our multicloud networking planning guide.


