In my role at Equinix, I often hear from global systems integrators (GSIs) who are trying to figure out how to help enterprises capitalize on cloud services while also limiting risk. Increasingly, this means they’re looking to apply the power of cloud to help their customers succeed with AI.
So much of what makes public cloud services great is ironically what could make them problematic from a sovereignty and privacy perspective. Their global reach can be very helpful in certain situations, like when enterprises need to replicate data across regions to ensure geo-redundancy. In other situations, enterprises need to be very careful about not moving data away from its country of origin.
GSIs recognize that helping their customers meet complex data sovereignty and privacy requirements will be a serious challenge, particularly when it comes to distributed AI workloads. To overcome this challenge, they’ll have to move away from the traditional cloud strategies that have served them well in the past—strategies built around public cloud services first, often provided by global hyperscalers. What they need instead is a sovereign cloud strategy that allows them to tap into cloud services on demand while also incorporating strong data governance principles.
A sovereign cloud strategy is all about ensuring flexibility and control to manage different workloads and datasets in different ways. Enterprises won’t be able to get this level of control if they rely exclusively on public cloud services. Many organizations are strongly considering a sovereign cloud strategy that incorporates private infrastructure alongside public cloud services. This hybrid approach will allow them to host different datasets in different places, positioning them to comply with data sovereignty requirements worldwide.
How is AI driving the sovereign cloud conversation?
Many of the factors that are leading businesses to closely examine their cloud sovereignty requirements are the same ones inspiring them to pursue a private approach to AI. After all, AI is inherently about data. To succeed with AI, enterprises need the right data, in the right amounts, processed in all the right locations. This means that they’ll need to move data quickly between data centers in many different places.
Having so much data in motion all at once can be risky. For instance, moving the wrong data into the wrong places can lead to non-compliance penalties or reputational damage. Data privacy is another concern: An organization’s AI data is an extremely valuable asset, so they need to be careful that they don’t unintentionally share it by moving it into the wrong location or using it in the wrong way.
Balancing the need for data mobility with the need for sovereignty and privacy controls won’t be easy. The fact that many enterprises see cloud services as an essential part of their AI strategy will only complicate things further. The challenge is to figure out how to get all the benefits of cloud infrastructure—the flexibility, the scalability, the access to specialized services—while also avoiding the potential drawbacks around data sovereignty and privacy.
In particular, many enterprises are considering GPU as a Service offerings that promise cloud-like access to advanced AI hardware. Of course, these GPUs still physically exist somewhere, even if customers can’t see them. Therefore, enterprises need to ensure they only work with GPUaaS providers located in the right places for their sovereignty strategy.
The good news is that private AI doesn’t have to mean “AI in isolation.” Enterprises can still make cloud services a part of their AI strategies, as long as they’re maintaining close control over what data moves into the cloud and why. To do this, they first need to deploy an AI-ready data architecture on private infrastructure. Then, they need to use secure, private interconnection to reach their cloud partners. This helps ensure that they’re only moving data into the public cloud when they truly need to do so, and that they’re able to get those datasets out of the cloud when their requirements change.
Access a global platform for sovereign cloud
To ensure cloud sovereignty, enterprises need to choose the right cloud providers in the right places, and they need the flexibility to change their minds. They can find that choice and flexibility on Platform Equinix®. Equinix is the global market leader in low-latency cloud on-ramps, so it’s easy for our customers to connect on demand with the cloud providers of their choice.
Our global footprint includes Equinix IBX® data centers in 74 metros across the world. This allows our customers to stand up private infrastructure in any location where they might need to store or process data. They can manage that private infrastructure to ensure they’re meeting their exact data sovereignty and privacy requirements. For instance, certain sensitive datasets can stay exclusively on private infrastructure, while other datasets can move to and from the cloud as needed.
For data that enterprises do need to move to the cloud—or to other partners in their AI ecosystems—they need to avoid the public internet, which has privacy shortcomings and doesn’t always follow the most direct route from Point A to Point B. In contrast, a private interconnection solution like the ones offered at Equinix can help enterprises feel confident that their data in motion is always protected, and that it won’t unintentionally pass through the wrong borders while in transit.
Whether it’s physical Equinix Cross Connects or Equinix Fabric® for virtual connections, our customers can choose the right mix of interconnection solutions to meet the performance and privacy needs of their AI strategy. They’ll also have no shortage of partners to connect with, as the Equinix ecosystem includes thousands of clouds, networks and enterprises.
Equinix partnerships enable sovereign cloud solutions
Public sector agencies often find it particularly challenging to access cloud services. They typically want to remain in their own on-premises data centers, but such data centers tend to be isolated from the cloud. This could make it difficult for them to scale infrastructure and tap into innovative new capabilities.
Thanks to Equinix private interconnection solutions and our vast partner ecosystem, public sector agencies can reach the cloud without having to leave their on-premises data centers. They can connect on demand with network service providers, knowing that any NSP they choose will already be at Equinix. Those NSPs can help them extend their network into an Equinix IBX colocation facility. We can then help them connect to clouds and other service providers while maintaining strict controls over which datasets are allowed to leave the country.
The diagram below shows an example of what this might look like for a public sector agency in Canada. The agency would be able to create a sovereign cloud within their own country, while also supporting global operations. For instance, they could establish a PoP in Japan that could one day lead to sovereign cloud infrastructure in that country as well.
Hyperscalers also recognize the need for cloud sovereignty, and have begun to offer solutions intended to meet that demand. With our global reach and unmatched interconnection portfolio, Equinix offers an ideal environment in which to deploy these sovereign cloud solutions.
Our ecosystem of partners, including GSIs, can help further simplify cloud sovereignty for our joint enterprise customers. For instance, we’ve worked with Kyndryl to help make it quicker and easier for enterprises to deploy Google Distributed Cloud in the locations of their choice. This solution extends Google Cloud services to the digital edge, allowing enterprises to meet their data sovereignty requirements while also enabling low-latency use cases such as AI inference.
Read the solution brief to learn more about Google Distributed Cloud with Equinix and Kyndryl.
Subscribe to the Equinix Blog