- Rakuten Mobile told Fierce private cloud is its go-to for operations
- It does use a handful of public cloud services, but has concerns about cost and control if it were to move network functions over
- The operator doesn't necessarily think AI will pull it into the public cloud either
Japanese operator Rakuten Mobile has been hailed as one of a handful of truly cloud-native telcos. This is thanks in large part to the greenfield nature of its network, which began offering commercial service just five years ago. But there’s an interesting twist to its cloud strategy: the operator is doing almost everything in a private rather than public cloud environment.
“All of our network functions, as well as OSS/BSS systems, are hosted on our private cloud infrastructure,” Ryota Mibu, GM of Rakuten Mobile’s telecommunication cloud platform department, told Fierce. That includes its network core.
Rakuten’s approach is a bit different from a number of its peers, including Deutsche Telekom and Vodafone, which similarly keep their network functions on a private cloud but have opted to put workloads like BSS and IT functions in the public cloud.
Why has Rakuten gone this route? “Our concerns with the public cloud include higher costs, reduced control over platform versions and environments, and potential challenges in troubleshooting during incidents,” Ryota explained. “Given the scale and usage patterns of our infrastructure, we have found that a private cloud delivers the best ROI for our organization.”
That said, Rakuten Mobile does use a handful of public cloud services such as Office 365, Box and Domo for business operations.
The AI element
Like other operators, Rakuten Mobile is exploring how it can use AI in its operations.
Ryota said one of the key hangups right now is deciding “how much autonomy to allow in operational management and what level of AI accuracy is acceptable.”
Public cloud providers are some of the leading providers of AI models. Ryota said Rakuten Mobile is being cautious about introducing AI workloads on its private cloud due to economic concerns. But that doesn’t necessarily mean that Rakuten is planning to jump ship and deploy its AI workloads on the public cloud, either.
“We anticipate that, once use cases and the GPU market stabilize, there may be a shift from public to private cloud for AI workloads,” he said.
Ecosystem gripes
Public vs private decisions aside, Ryota said operating in the cloud generally offers perks like more flexible allocation of network resources and the ability to easily visualize and optimize resource usage. The latter is key because it helps control capex, he explained.
But Ryota echoed recent comments from Deutsche Telekom and Orange execs, who noted the telco ecosystem hasn’t yet caught up to the cloud paradigm.
“Many telco applications and ecosystems are still not fully cloud-native. While the evolution of telco may differ from that of the enterprise sector, both vendors and operators need to continue making improvements,” he concluded. “As telco business models and operations evolve, the ability to update systems and applications will be critical to survival.”
Catch the other installments in our telco cloud series here: