Google’s Multi-Cloud Future

The search engine giant is staking its cloud success on a commitment to multi-cloud and data analytics, but many hurdles remain.

Karl Henrik Smith
Jun 21 · 8 min read

In April, Google Cloud announced a new open platform for managing applications called Anthos, a flexible and secure software-based solution. Anthos would provide enterprises with the capabilities of running apps on existing on-prem hardware or the public cloud. At its heart is open source platform Kubernetes, a key differentiator between Google Cloud Platform (GCP) and established players like Amazon Web Services (AWS) or Microsoft Azure. Kubernetes, which tech analyst Ben Thompson dubbed “the perfect antidote to AWS’ ten-year head-start in infrastructure-as-a-service,” is a fully portable open-source container cluster manager allowing GCP to effectively run microservices.

Anthos is a core component of GCP’s commitment to a multi-cloud ecosystem. When Anthos was launched at Google’s Next conference in April, the search engine giant cited research claiming that 88% of enterprises plan to use multiple cloud providers. And although several Google Cloud customers have existing software and infrastructure in place, they often leverage two or more cloud services at once to deploy their applications seamlessly.

In the case of Anthos, embracing open standards ensures that Google can deliver the benefits of a fully-managed service alongside the existing infrastructure capabilities of large enterprises. Its hybrid functionality is available both with the Google Kubernetes Engine (GKE) and in data centers with GKE on-prem, and will allow users to run multiple workloads; from the Google Cloud blog:

Anthos will also let you manage workloads running on third-party clouds like AWS and Azure, giving you the freedom to deploy, run and manage your applications on the cloud of your choice, without requiring administrators and developers to learn different environments and APIs.

In addition to furthering GCP’s multi-cloud strategy, the Anthos announcement also places additional emphasis on the need for partnerships. Thomas Kurian, who took over Google Cloud from Diane Greene earlier this year, also took the opportunity to unveil a new open-source partner program with leading data analytics and management companies including Confluent, DataStax, MongoDB, and Neo4j. These partners will be fully integrated into GCP to provide an improved user experience in several channels, from insights to billing and support.

This strategy to partner and share revenue with popular data analytics platforms could be beneficial to open source vendors, permitting them to build software without concerns over building their own cloud infrastructure; from ZDNet:

While the initial impression might seem like a manifestation of Google’s original ‘Do no evil’ strategy to paint a contrast to AWS, the reality is that this is more about GCP building customer loyalty as the preferred destination of the open source platform installed base who wants the best of both worlds: the full open source or open core stack of these platforms and the flexibility of having a choice around cloud providers. It’s another way to overcome the syndrome of enterprise customers about nobody getting fired for buying from the usual suspects.

These open-source integrations also draw a contrast of Google’s broader ecosystem with that of AWS, whose tack has been to repackage open-source projects under its own brand. Manvinder Singh, the head of infrastructure partnerships at Google Cloud, says that the best way to deliver open-source technologies as services in the cloud is “to work closely together with companies that have invested their resources in developing these open-source technologies.” Unlike AWS, Google’s strategy seems not to be launching an imitation service by gobbling up established data analytics and management companies. Rather, it argues that continued partnerships with such firms will be more conducive to creating sustainable open source communities.

Eggs <> Basket

One of the earliest drivers of multi-cloud was the inability for enterprise customers to easily transition between cloud services, known as vendor lock-in. In addition to the uncertainty around the reliability of a single cloud offering (and to prevent data loss and downtime), a multi-cloud approach would ensure that organizations are not being constrained by inefficient processes or cloud contracts.

Although a single cloud provider might offer the right mix of service level-agreements (SLAs), protocols, and uptime guarantee, this is not always enough. If users start requesting features that are exclusively available through a different application, enterprises may have to purchase that vendor’s public cloud solution. Multi-cloud setups typically combine at least two cloud providers:

Today, multi-cloud deployments are typically driven by geographical advantages of a specific cloud service or broader strategic or technical goals. Google Cloud lists a number of reasons why organizations would engage in a multi-cloud setup, including reducing IT spending, agility and flexibility in responding to market demands, or regulatory and legal compliance.

Multi-cloud approaches should be distinguished from the term hybrid cloud, which in the context of Google Cloud refers to a setup “in which common or interconnected workloads are deployed across multiple computing environments, one based in the public cloud, and at least one being private.” Typically, this also includes a private computing environment:

Google Cloud’s performance over the past month also points to why multi-cloud setups are becoming more common. In early June, an outage brought on by a routine configuration change and high levels of network congestion took down YouTube, Snapchat, Discord, and a number of other web services which all use GCP on the backend. The outage (and subsequent issues with service disruptions) underscores the fragile nature of modern computing architectures, and that placing one’s entire backend in the hands of a single company may not be the wisest solution.

The Looker Acquisition

Earlier this year, Google announced an investment of $13 billion on data centers and offices across the U.S. to bolster its cloud computing unit. Although it faces high competition in the cloud space from AWS and Azure, Google doubled its capital expenditure over the last fiscal year in its push for more cloud customers. The search engine giant reported on June 6th that it would be acquiring the data analytics company Looker for $2.6 billion. As both companies share over 350 joint customers (including BuzzFeed, WPP Essence, and Yahoo!), the integration of Looker will nicely complement Google Cloud’s existing analytics foundation.

On a fundamental level, Looker allows companies to understand their data stored in the cloud. The 2019 Wisdom of Crowds BI Market Study ranks the business analytics firm as a leader in categories including customer experience and vendor credibility. Looker’s CEO Frank Bien stressed the importance of committing to customer success, as the increasing complexity and volume of data makes integration at scale across the entire organization much more valuable. According to Bien, Looker had been trying to change the business intelligence (BI) and analytics market early on; from TechCrunch:

What we wanted to do was disrupt this pretty staid ecosystem of data visualization tools and data prep tools that companies were being forced to build solutions. We thought it was time to rationalize a new platform for data, a single place where we could really reconstitute a single view of information and make it available in the enterprise for business purposes.

In addition to its commitment to user experience, Looker says it will continue to work with all major cloud platforms in line with Google Cloud’s multi-cloud strategy. In a blog post about the acquisition, Kurian argues that Looker gives GCP the ability to more consistently define all metrics across different sources of data, such as supply chain in retailing or media analytics in entertainment. The integration of Looker will offer customers a far more complete analytics solution (see below).

Looker’s platform has a threefold focus: business intelligence, data applications, and embedded analytics. Data modeling will be available at any scale across hybrid and multi-cloud, alongside pre-configured data applications for several industries (IT, sales, HR) and improved visualization tools. By providing customers with an end-to-end analytics platform, the deal with Looker seems to be the inevitable next step forward in Google’s multi-cloud strategy.

Anti-trustworthy

Two weeks ago, the Wall Street Journal reported that the U.S. Justice Department was recently granted jurisdiction to engage in an antitrust probe of Alphabet (Google’s parent company). The deal with Looker, which is subject to government approval, is raising concerns among regulators over potential anti-competitive practices. But Kurian contends that the transaction poses no antitrust problems, since other data analytics firms remain in the market, and Google Cloud would not be obtaining any customer data in the process.

Antitrust regulators would have more grounds for a probe were it not for Google’s multi-cloud strategy, and its continued insistence that Looker will continue to work with AWS and Microsoft Azure. Google Cloud emphasizes that there are many analytics tools in the market, so “by just acquiring Looker, we’re not further concentrating the market in any sense.” And as multi-cloud usage becomes more prevalent for large enterprises, pushing cloud-neutral (or third-party) software firms like Looker will benefit GCP; from the Motley Fool:

Pushing a multi-cloud agenda would obviously benefit the laggard — in this case, third-place Google — by preaching diversification away from AWS and Azure and splitting workloads between all three. By purchasing Looker, Google indirectly benefits from the success of multi-cloud usage among its rivals, as long as they use Looker’s software. In other words, something is better than nothing, even if it means ceding some cloud workloads to competitors.

But a strategic pledge to multi-cloud might not be enough to fend off impending investigations into Google’s anti-competitive behavior. In its potential probe, the Justice Department will aim to uncover whether the search engine giant leveraged its position to unfairly take down its smaller competitors in markets like mobile operating systems and cloud computing. And although the Looker deal is unlikely to be deemed anti-competitive, Google should still be prepared for increased government scrutiny in the months ahead.

The Startup

Medium's largest active publication, followed by +479K people. Follow to join our community.

Karl Henrik Smith

Written by

Cloudflare | Copenhagen Business School ‘18

The Startup

Medium's largest active publication, followed by +479K people. Follow to join our community.