4 Reasons Multicloud Networking Is So Difficult

Within the definition of “cloud computing,” cloud service models can be boiled down into three main categories: Software as a Service, Platform as a Service, and Infrastructure as a Service.[1] While these aren’t new concepts, it’s important to remember that “the cloud” is really just someone else’s technology that’s offered “as a Service.”

This means that you’re beholden to the decisions that your chosen cloud provider makes. These decisions can impact how your workloads function, how well they perform and how easily they interoperate with other environments. This is especially true when choosing multiple clouds.

In fact, multicloud networking may be one of the most complex problems IT leaders will ever have to solve. The root of the issue is that infrastructure and services from different cloud providers inherently weren’t built to play nice with one another. In this blog post, we’ll explore four categories of challenges that could arise when trying to connect different cloud environments:

1. Technical challenges
2. Operational challenges
3. Security challenges
4. Cost challenges

We’ll also look at how your choice of networking technology could exacerbate or mitigate multicloud challenges.

1.    Technical challenges: Ensuring performance and reliability

Each cloud is like a world unto itself, with its own unique terminology, architectural design, management and monitoring tools, and even certification options for people to prove they can operate effectively within that specific ecosystem. When these different worlds collide, it often leads to technical challenges like poor performance and reliability.

In one survey from Enterprise Strategy Group (ESG), business leaders were asked to name why they use more than one public cloud provider. The two most common answers were performance flexibility and reliability.[2] Thus, networking complexity could prevent them from realizing some of the main benefits that attracted them to multicloud in the first place.

Consider the performance needs of multicloud applications. Different workloads within an application may have certain bandwidth and latency requirements, as well as different data storage and access needs. If those requirements aren’t met, the application won’t function as intended, and user experience will suffer. Running workloads in multiple clouds could greatly complicate things, since each cloud offers different capabilities, from different locations, using different management practices.

Trying to connect different cloud environments via the public internet may intensify built-in performance challenges. With the internet, multicloud traffic won’t always follow the most direct route, as shown in the diagram below. When traffic has to bounce between various internet exchange points, it will inevitably lead to slower data transfers between clouds, often with limited visibility into where the performance or reliability issues are occurring.

2.    Operational challenges: Implementing multicloud observability and change management

In multicloud networking, one set of problems may lead to another. For instance, operational challenges like a lack of comprehensive multicloud observability can lead to technical challenges such as poor performance. That’s because proprietary observability tools can only give you visibility into one piece of your multicloud environment. Many of these tools offer limited—and in some cases, severely limited—diagnostic, troubleshooting and performance-related data. You could be experiencing networking problems that limit performance, and you wouldn’t even know it.

Even if you do know about these issues, it may still be difficult to address them. You may not be able to quickly change your existing cloud connections or create new ones. Many businesses look to the public internet to enable multicloud change management. Because the internet is everywhere, they assume it’s easy to start in any one location and connect to any other location whenever the need arises. It’s true that internet connectivity can be useful for certain multicloud use cases. As an Equinix colleague once put it, the internet works—until it doesn’t. If you start using the internet just because it’s the “easier” option, you may not be prepared for just how difficult it could get.

For example, a third-party observability tool can help provide cloud-agnostic visibility, but these tools were never intended to function as an internet overlay. There are simply too many unknowns with the public internet—unknowns around performance, routing, reliability and more. Choosing private interconnection instead can help remove these unknowns and thus put the full benefits of these tools within reach.

3.    Security challenges: Protecting data in motion

Moving data between cloud environments inevitably creates some degree of risk. Enterprises need to think about the acceptable and unacceptable risks they face, and act accordingly. For instance, all businesses have crown jewel data assets that must be protected at all costs. Also, failing to meet compliance or data sovereignty requirements represents unacceptable risk for most businesses. They must do everything in their power to meet these requirements, and that means looking long and hard at how their cloud data is stored, transferred, accessed and processed.

A multicloud environment gives rise to security challenges because it creates a wider attack surface for businesses to protect. Also, each individual cloud will have its own shared responsibility model, making it difficult for businesses to know exactly where their security responsibilities end and the cloud providers’ responsibilities begin.

For many businesses, the natural inclination is to put their most vulnerable datasets in on-premises environments where they can maintain more control. There’s nothing inherently wrong with keeping sensitive data out of the cloud, but it does represent its own kind of risk: The risk that businesses won’t be able to use their data to its full potential because it’s isolated from enterprise-class cloud services.

Businesses must be able to move or copy their datasets into different clouds as needed without having to sacrifice control or visibility. They can achieve this by establishing their own data storage environment with low-latency private interconnection to different clouds. In this arrangement, they can move copies of their data into the clouds that they choose, while maintaining authoritative datasets in an environment that they control. This also means that companies can ultimately be deliberate about what data goes where. Instead of having all their data in one cloud, they can granularly move data to wherever it can be used most effectively.

4.    Cost challenges: Egress, licensing and more

The cost of data egress is widely recognized, but it’s certainly not the only multicloud cost driver that businesses have to deal with. In fact, all the challenges we’ve discussed so far could require serious financial resources to address. Organizations may find themselves paying licensing fees for multiple tools, such as VPNs and SD-WAN devices, to help overcome various technical, operational and security challenges.

Using the internet for multicloud connectivity could intensify many of the cost challenges that businesses already face in their multicloud environments. From an egress standpoint, they’d find themselves paying based on the amount of data moved, while a private interconnection solution could instead provide more predictable pricing.

Moving data over the public internet also leads to additional security costs. The only way to truly ensure data security over the internet is to crack each individual packet and perform a deep packet inspection. Not only does this decrease performance; it’s also very time-consuming for the security and networking teams and could require them to purchase costly tools.

Finally, poor performance can lead to higher costs in unexpected ways. For instance, companies that need to migrate very large datasets between clouds will likely have to pay for duplicate data storage in both clouds during the move. This means that completing the transfer quickly is of the utmost importance. Uber was able to achieve this when they used an Equinix interconnection solution to migrate 6.5 petabytes of cloud data on an aggressive timeline.

Find the right balance for your multicloud networking needs

There’s no single solution that can easily address all the multicloud networking challenges outlined here. It’s all about striking the right balance for the unique needs of your business and your different workloads. Both the public internet and private interconnection have their own capabilities and strengths, which means that it could make sense to choose a mix of both for different purposes.

To learn more about what’s involved with overcoming multicloud networking challenges, read the ESG analyst report Solving the Hybrid Multicloud Networking Puzzle.

[1] Peter Mell and Timothy Grance, The NIST Definition of Cloud Computing, National Institute of Standards and Technology (NIST), Special Publication 800-145, September 2011.

[2] Jim Frey, Solving the Hybrid Multicloud Networking Puzzle, Enterprise Strategy Group, now part of Omdia. Commissioned by Equinix and distributed under license from TechTarget, Inc. July 2025.