Based on the company’s capital investments prior, it’s certainly not fair to date Google’s level of commitment to its cloud platform to Diane Greene’s hiring via acquisition in November of 2015. But that is just what many in the industry do. That event is viewed as an unofficial but unmistakable milestone in the evolution of what has become a growing and vital business for Google.
Whatever marker you pick, it’s worth taking stock of where the company is at present, and what better occasion than in the wake of its annual Next event? A showcase for Google’s enterprise efforts, Next is a show of relatively recent vintage. Up until a few years ago, Google’s cloud offerings were crammed uncomfortably into the company’s other, massive developer event I/O.
Over the past few years, however, Next has grown along with Google’s cloud business culminating in this year’s event where the attendance was such that the main stage moved from Moscone West to the still-under-construction South. If I/O is an event dominated by front end technologies like Android and Chrome, Next is about the back end infrastructure – GCP to G Suite and everything in between.
In 2018, when back end infrastructure is discussed, generally what people mean is cloud, and – with one important exception, which we’ll come back to – that is the case with Next. The event is the opportunity for Google to update the market on its progress in that and other areas.
Google may receive less attention than competitors such as AWS and Microsoft, but its aforementioned capital investments mark it as one of the few serious players in the market. As my colleague Rachel has written:
Competitiveness in the cloud market is largely reliant upon companies’ infrastructure investments. More specifically, given the size of the investment required it should be increasingly possible to infer actual competitive intent from the balance sheet in ways not possible in other technology markets such as software.
Having invested over thirty billion dollars over the last three years in its infrastructure from hardware to submarine cables, Google has bought itself a seat at the adult’s table.
The question at Next wasn’t, then, whether Google belongs in a conversation with the likes of AWS, Azure and, increasingly, Alibaba. The question is where Google is choosing to invest that capital, and how those investments are paying off.
To explore that question, here are five brief takeaways from Google Next.
- Google Goes on Premises:
With apologies to its front and center positioning of ethics in discussions of its various AI initiatives, the single most important news at Google Next was the announcement of GKE On-Prem. Kubernetes has become a veritable force of nature since Google released it as open source software in mid-2014, but it remains complicated enough to stand up and run that even high end developers are happy to have help managing it. This is the market opportunity that GKE was created in response to, along with AWS’s EKS, Pivotal’s PKS, Microsoft’s AKS, Red Hat’s OpenShift and startups like Heptio. To date, however, there have been clear lines of distinction between on and off premises solutions. Specifically, off premises, managed cloud solutions were limited to their own platforms.
With GKE On-Prem, Google is abandoning that distinction and embracing a hybrid, on or off premises world. This is a major announcement for Google, because it at once dramatically expands its potentially addressable market and complicates its relationship with partners such as Pivotal and Red Hat. The latter concerns may be overstated, however, both because coopetition is the industry norm, and because it’s likely that one or more cloud competitors will follow Google’s lead in this regard rendering the distinction moot.
Like all companies competing with Amazon Web Services, Google’s looking to expand where AWS isn’t – and on-prem is an interesting opportunity to do that. Importantly also, Google’s messaging here is not “Run like Google,” which can be offputting for companies who don’t see the need to do so, but the more compelling “we’ll bring the cloud to you.” For companies that want or need to maintain their own datacenters but want them to look more like the cloud, that sort of message will resonate. Now it’s up to Google to execute on their stated vision, which will be a challenge given the dramatically expanded number of environmental configurations they’ll need to contend with.
Apples & Oranges:
The good news for Google in its bid to contend with AWS is that it has a diverse array of assets to leverage. In G Suite, as one example, Google can – like Microsoft – offer enterprise customers a comprehensive suite of productivity and messaging tools to go along with its broad portfolio of infrastructure assets. The bad news for the company, however, is that attempting to message such radically different products with different buyers and consumption models at the same show is challenging.
From analysts to general attendees, trying to parse news about on premises Kubernetes offerings alongside AI integrations with calendaring is problematic. The solution to this often awkward arrangement is unclear; it’s probably not worth it for Google to spin up yet another conference, and offerings like G Suite and even Android or Chromebooks are relevant to the enterprise. But it’s a non-optimal conference structure for the time being.
For any new or emerging technology infrastructure platform, one of the most critical yardsticks by which to measure its success is its enterprise credibility. It’s difficult, after all, to sell to the enterprise if enterprises – and their partners of choice – don’t believe in the platform. This year’s Next clearly had this in mind, rolling out the likes of Accenture, Oracle and SAP on Wednesday. For its partner panel on Monday, in fact, three of the five participants were large systems integrators and one was a large enterprise vendor.
While Google made time for startups like Evernote or Spotify, then, this year’s Next was clearly an event aimed at showcasing Google’s hard core enterprise chops. Telegraphing this message to the market, however, is a necessary step but just the first of many to come. As businesses warm to the idea of Google as a legitimate business platform, they’ll expect to be courted in the manner which they have come to expect, which means that GCP – already the fastest growing part of Google as measured by headcount the last two years – will need to follow in the footsteps of other emerging enterprise players and poach enterprise sales talent from organizations with that experience.
One of the most notable aspects to the GCP portions of Next were the degree to which they focused on higher level abstractions. To be sure, there was discussion of underlying compute capabilities, network performance and the like, but it’s clear from Next that Google is focused up the stack. The base infrastructure is table stakes, but if this event was any indication, Google’s belief is that the winner won’t be determined at that level.
Instead, pieces like GKE, Istio and the newly announced Knative (pronounciation TBD) earned the lion’s share of the attention. Serverless in particular stood out as an area of focus, with Google (finally) making Cloud Functions generally available and announcing Knative along with IBM, Pivotal, Red Hat and SAP. Described as a set of serverless primitives (where serverless in this case essentially translates to FaaS) that sits on top of Kubernetes, it’s a response to AWS’ Lambda that seeks to differentiate itself both by drawing on Kubernetes’ surging momentum and offering multiple vendor support models. Bigger picture, Knative and even its GKE underpinnings point to a Google which is looking to frame the cloud debate above the base infrastructure, a frame which plays to the vendor’s strengths.
The open question at this point isn’t the level of abstraction targeted but for how much longer the serverless term itself can hold up under an increasingly crushing load of industry-wide marketing baggage. If you define serverless as encompassing base compute, containers and the function-as-a-service offerings typically covered by that term, the term serverless itself is indistinguishable from cloud and has lost all utility. Whatever it ends up being called, however, it’s clear that what we call serverless today is a model here to stay, and one that Google and its Knative partners are intently focused on.
Google and the New Kingmakers:
For a company that prides itself both on the quality of its engineers and the high end products and services said engineers produce, the company hasn’t internalized the idea of developers as kingmakers top to bottom to the degree that one might expect.
While Google has (and prominently featured) one of the single best developer advocates in the world in Kelsey Hightower on staff, competitors such as Microsoft are being notably more aggressive in their broader developer envangelism and recruitment efforts. Efforts that, based on Azure’s accelerating traction, appear to be paying off. Google’s devrel team is strong, but still small considering the market reality of AWS’ first to market status and Microsoft’s long tenure as a business supplier of choice.
It’s important for the business unit’s trajectory that GCP, for the reasons discussed above, be seen as catering to enterprise buyers in ways it has not in recent years. But it’s equally important to not forget that even within enterprise buyers, developers have enormous influence on purchasing decisions even if they’re not the ones signing checks. As one CIO at the event put it: “It wasn’t me that chose Google, it was the engineers.”
Like Microsoft, it is incumbent on Google to give its developer relations staff the resources to compete for – and win – developer time and attention. From outreach to product to advocacy, Google can’t rely on technical excellence to bring customers to it: it must go to where they are – and to do that it needs developer help.
Disclosure: Google is a RedMonk customer, as is Amazon, IBM, Microsoft, Pivotal, Oracle, Red Hat and SAP. Accenture and Heptio are not RedMonk customers.