Zurück
Werbung
Next ’26 Partner Talks: How Rakuten Cloud and Google Distributed Cloud are redefining infrastructure
June 11, 2026
Google Cloud Next ’26 was an exciting show for Rakuten Cloud with a message and a product suite that resonated with customers. One of the highlights of our time at Google Cloud Next ’26 was the fireside chat with Mike Ensor, Technical Lead for Google Distributed Cloud Solutions, conducted by Anirban (Oni) Chakravartti, Rakuten Symphony’s SVP and Global Head of Sales (Enterprise).
What set this interview apart? Both Mike and Oni are on the front lines of edge deployments, involved in some of the largest deployments worldwide. With that experience level, this interview became an update from the front lines of edge computing.
Building the edge for AI
The historic selling point for edge was its ability to modernize and improve existing applications and use cases. But as customers build out their AI applications, they are looking to their edge compute infrastructure to process some of that data – changing the requirements for edge infrastructure. This is a “new-ish” trend, especially among restaurant and retail chains that are among the early adopters of GDC services.
“I think we can all recognize that the word AI is now in front of everything, so we probably don't even have to mention that it's new. But especially for the areas that we are working in with the edge, I think (AI) is a more appropriate use of edge technology,” Ensor said.
Data repatriation
Some customers who are considering building out their edge network are exceeding their “knowledge-transfer comfort level, because they have moved from on-prem to the cloud. And now we are telling them to come back to on-prem again,” Chakravartti said.
There are at least two general contributors to this push to repatriate data to the edge. The first is that an increasing number of workflows benefit from processing data close to where it is created or consumed for security or intellectual property protection reasons.
Then there is the concept of fast and slow data. As Ensor describes it, fast data is real-time processing in applications such as manufacturing, where assembly line machines need to process thousands of metrics per second to ensure efficient operation.
But that highly granular data doesn't mean nearly as much at a global or higher organizational level. Ensor discusses a theoretical slow-data example that takes the original 1,000 metrics per minute data flow and averages it to 1 per minute, which is then sent to the cloud for processing. The results of this slow data can be used in other factories where it becomes important for predictive analysis or other types of analytics. Fast and slow data are equally important to the customer: making real-time decisions while maintaining a long-term perspective.
Resilience and security are important
The discussion shifted to emerging approaches to securing edge computing and ensuring resilience.
One successful approach is creating smaller storage volumes specifically for active workloads. With Rakuten Cloud-Native Storage, that can be done dynamically. This is an important capability for agentic AI, where a new storage volume can spin up that is independent and it can have its own encryption. This extra layer of security is important because if one storage volume is compromised, the amount of information that can be read or used is minimal. This small data store is not particularly useful to the hacker, so the system is not a target.
Automation for day zero and beyond
With customers deploying hundreds, thousands or even tens of thousands of edge systems, automation of everything is critical. GDC already supports zero-day automation, allowing customers to plug in the hardware and have everything launched and deployed from a remote location.
This automation extends beyond day zero because problems can occur later when upgrades are needed for software, patches, or firmware. All of that can be managed from a central location using Google Site Reliability Engineering (SRE). The SRE teams get a notification if any aspect of the system is degraded – from someone tripping over a cable to a node going down. If the system can fix it, it will. But if a human needs to be involved, the SRE gets notified.
The synergy of Google Cloud and Rakuten Cloud
Beyond the technical specifications, the strength of the partnership between Rakuten Cloud and Google Cloud lies in their collaborative commitment. As one of Google Cloud’s select OEM partners, Rakuten integrates its proprietary software-defined storage with Google Distributed Cloud (GDC) to create a high-performance solution. More importantly, the two companies operate as a unified team, consistently going above and beyond to ensure customers are successful, even if it means working around the clock to meet complex customer needs.
What’s next for GDC?
The solution has seen great success in retail applications. “Our next major areas are healthcare, oil and gas, energy, and manufacturing,” said Ensor. “Those are the next major verticals we are going after, and each of them has some unique components.”
cloud.rakuten.com/blogs/...loud-are-redefining-infrastructure
June 11, 2026
Google Cloud Next ’26 was an exciting show for Rakuten Cloud with a message and a product suite that resonated with customers. One of the highlights of our time at Google Cloud Next ’26 was the fireside chat with Mike Ensor, Technical Lead for Google Distributed Cloud Solutions, conducted by Anirban (Oni) Chakravartti, Rakuten Symphony’s SVP and Global Head of Sales (Enterprise).
What set this interview apart? Both Mike and Oni are on the front lines of edge deployments, involved in some of the largest deployments worldwide. With that experience level, this interview became an update from the front lines of edge computing.
Building the edge for AI
The historic selling point for edge was its ability to modernize and improve existing applications and use cases. But as customers build out their AI applications, they are looking to their edge compute infrastructure to process some of that data – changing the requirements for edge infrastructure. This is a “new-ish” trend, especially among restaurant and retail chains that are among the early adopters of GDC services.
“I think we can all recognize that the word AI is now in front of everything, so we probably don't even have to mention that it's new. But especially for the areas that we are working in with the edge, I think (AI) is a more appropriate use of edge technology,” Ensor said.
Data repatriation
Some customers who are considering building out their edge network are exceeding their “knowledge-transfer comfort level, because they have moved from on-prem to the cloud. And now we are telling them to come back to on-prem again,” Chakravartti said.
There are at least two general contributors to this push to repatriate data to the edge. The first is that an increasing number of workflows benefit from processing data close to where it is created or consumed for security or intellectual property protection reasons.
Then there is the concept of fast and slow data. As Ensor describes it, fast data is real-time processing in applications such as manufacturing, where assembly line machines need to process thousands of metrics per second to ensure efficient operation.
But that highly granular data doesn't mean nearly as much at a global or higher organizational level. Ensor discusses a theoretical slow-data example that takes the original 1,000 metrics per minute data flow and averages it to 1 per minute, which is then sent to the cloud for processing. The results of this slow data can be used in other factories where it becomes important for predictive analysis or other types of analytics. Fast and slow data are equally important to the customer: making real-time decisions while maintaining a long-term perspective.
Resilience and security are important
The discussion shifted to emerging approaches to securing edge computing and ensuring resilience.
One successful approach is creating smaller storage volumes specifically for active workloads. With Rakuten Cloud-Native Storage, that can be done dynamically. This is an important capability for agentic AI, where a new storage volume can spin up that is independent and it can have its own encryption. This extra layer of security is important because if one storage volume is compromised, the amount of information that can be read or used is minimal. This small data store is not particularly useful to the hacker, so the system is not a target.
Automation for day zero and beyond
With customers deploying hundreds, thousands or even tens of thousands of edge systems, automation of everything is critical. GDC already supports zero-day automation, allowing customers to plug in the hardware and have everything launched and deployed from a remote location.
This automation extends beyond day zero because problems can occur later when upgrades are needed for software, patches, or firmware. All of that can be managed from a central location using Google Site Reliability Engineering (SRE). The SRE teams get a notification if any aspect of the system is degraded – from someone tripping over a cable to a node going down. If the system can fix it, it will. But if a human needs to be involved, the SRE gets notified.
The synergy of Google Cloud and Rakuten Cloud
Beyond the technical specifications, the strength of the partnership between Rakuten Cloud and Google Cloud lies in their collaborative commitment. As one of Google Cloud’s select OEM partners, Rakuten integrates its proprietary software-defined storage with Google Distributed Cloud (GDC) to create a high-performance solution. More importantly, the two companies operate as a unified team, consistently going above and beyond to ensure customers are successful, even if it means working around the clock to meet complex customer needs.
What’s next for GDC?
The solution has seen great success in retail applications. “Our next major areas are healthcare, oil and gas, energy, and manufacturing,” said Ensor. “Those are the next major verticals we are going after, and each of them has some unique components.”
cloud.rakuten.com/blogs/...loud-are-redefining-infrastructure
