Google bets on multi-cloud

The cloud is like Tater Tots: So good, you can’t have just one — and according to analyst firm 451 Research, nearly every enterprise today supports multiple clouds.

In fact, some 98% of enterprises are using, or plan to use, at least two cloud infrastructure providers, with 31% using four or more, according to a recent report by 451 Research, part of S&P Global Market Intelligence, commissioned by Oracle Cloud Infrastructure.  

Multi-cloud is an opportunity for Google, the perennial third place in the cloud market, according to a Q4 2022 report by Synergy Research Group. Amazon Web Services (AWS) is the longstanding market leader in the cloud, with 32% market share; Microsoft Azure is second place, with 23%, and Google Cloud lags a distant third, at 11%. 

Analysts told Silverlinings about a few directions Google could take to beat the competition, including beefing up its channel strategy and learning to speak the language of legacy IT.

Multi-cloud drives Google forward as well. 

“Multi-cloud comes up in almost every conversation I have with customers,” said Bobby Allen, Google group product manager and self-described “cloud therapist” for the cloud provider. 

In a recent multi-cloud win for Google, Uber struck deals with both Google and Oracle to shift critical workloads from its own datacenters to the two cloud platforms. The ridesharing company also works with AWS. The Google and Oracle deals are reportedly partly driven by Uber’s difficulty acquiring data center equipment. 

Cloud native? It's ‘cloud runtime’ now

Allen describes his job as talking with customers about their needs, with a particular focus on “cloud runtime” applications — applications that run on containers and use serverless architecture. Allen explains that this category of application was formerly known as “cloud native,” in less linguistically sensitive times.

Allen compares Google’s cloud runtime strategy to cooking with an air fryer. 

“I love Tater Tots in the air fryer. Then my mother heated up a croissant for me in the air fryer, and it was amazing,” Allen said. “I’d had croissants, and I’d used the air fryer, but I didn’t put the two of them together before.”

And that’s the similarity between Google’s cloud runtime strategy and air fryers, “We want to support your doing something amazing that you might not have thought about,” Allen said. 

We’re not sure the whole air fryer/cloud runtime comparison works, but we’re also hungry now for croissants and Tater Tots. So maybe we’re not thinking clearly. 

Cracking the enterprise market

“Multi-cloud is a really big deal for Google,” said 451 Research analyst Melanie Posey. “Cloud-native applications that run on Kubernetes in clusters is Google’s sweet spot. They’re looking to grab customers that want to do new things, not take what they had on premises and stick it in the cloud.”

Google is particularly focused on helping enterprises use data analytics to get business insights, as a means of Google’s cracking open the enterprise market. Multi-cloud is essential to that strategy, because that kind of data often does not reside in Google Cloud, Posey said. 

In addition to Uber, Google’s multi-cloud support has attracted customers including the Roche multinational healthcare company, Major League Baseball (MLB) and Moderna. 

What’s driving enterprises to the cloud?

The need for multi-cloud is driven by the long-term transition of cloud apps from dev-test and experimental apps to mission-critical applications. “People have very valuable assets in the cloud,” he says. They want freedom of choice between multiple cloud providers to protect those assets.

Additionally, M&A activity means merged companies have different assets in different clouds. 

Another multi-cloud driver: Regulatory concerns, with jurisdictions requiring customer data be stored locally. 

Additionally, users want to mix and match the strengths of different cloud platforms, Allen said.  

Surprisingly, portability — the ability to move assets from one cloud to another — is not paramount importance. Users want to be able to move workloads between clouds, but not instantaneously. They’re not moving assets on a daily basis to arbitrage costs.

For Google, hybrid cloud, with servers located on customer premises, is part of multi-cloud. 

Anthos, BigQuery and other key technologies

The open source Anthos platform, for managing containerized applications in the cloud, is key to Google’s multi-cloud strategy. Google developed Anthos to manage its own applications, and now makes the software available to run on AWS, Microsoft or private clouds running on enterprises’ existing servers. 

Other multi-cloud tools that are strategic for Google are BigQuery for big data analytics, competing with Snowflake and Teradata, Apogee for API management, and Google Kubernetes Engine (GKE). All of those tools run on Google, Amazon, Microsoft and on-premises. 

“We respect the fact that customers need autonomy and options,” Allen said. 

Enterprises can run into problems managing multi-cloud platforms. Typically, they connect to three clouds, in Allen’s experience — two public platforms and one private. 

One problem comes when enterprises let individual departments and groups spin up their own cloud applications without central planning. A related problem is failure to plan for the long term. 

Typically, enterprises connect to three clouds, in Allen’s experience — two public platforms and one private. 

Enterprises also can have unrealistic expectations. For example, some enterprises will split individual apps across multiple clouds, which does not work well. Individual apps should run in separate clouds, Allen added.