Earlier this month I argued that Microsoft should acquire Docker. After all, my reasoning went, Docker needs help in monetizing its hugely popular container runtime, while Microsoft needs a foil to Kubernetes, which is rapidly gaining the upper hand among developers as the orchestration tool of choice in managing Docker containers at scale. I thought it made sense.
Mark Balch, VP of Products at container appliance manufacturer Diamanti, however, did not.
Talking with Balch, he thinks Microsoft should avoid the orchestration wars altogether as it seeks to differentiate. Instead, he says, Microsoft should concentrate on making Kubernetes great for Windows rather than adding needless friction to the market by buying Docker. Microsoft, just like Google and Amazon, should focus on developing and extending platform services, he argues.
It's not (really) about orchestration
TechRepublic: Elaborate on how platform services win for the big three and sublimate orchestration....
Balch: I'd argue that 99% of public cloud value is the platform, assuming orchestration doesn't get in the way. Take Amazon Web Services, for example. AWS built the first scaled IaaS with essentially zero orchestration (from a user perspective). The customer would script the provisioning of individual VMs. It was a smashing success. Now Amazon leads price declines on IaaS and features value-added platform features like S3, EBS, databases, messaging, gaming, billing, and so on.
SEE Kubernetes leads container orchestration (ZDNet)
Nor is AWS alone. Google's cloud strategy is likewise far beyond IaaS because it's not where anyone can differentiate. Google positions analytics, AI/ML, databases, and much more as unique offerings that leverage their expertise in other parts of the business. Azure is the same, offering things like MSSQL, .NET, and enterprise-specific functions to complement their app portfolio.
In sum, cloud orchestration only matters if the services being orchestrated are compelling.
The Docker distraction
TechRepublic: Why not both for Microsoft? Buy Docker to better control the direction of the most popular enterprise container format AND invest in platform services?
Balch: Docker's container runtime is an exciting technology that they've lost control of and cannot monetize. Docker's orchestration stack would be an expensive boat anchor for any acquirer. Microsoft is embracing services on Linux (think MSSQL), so that it's not just a Windows shop anymore.
SEE Why Microsoft should acquire Docker to better compete against Kubernetes (TechRepublic)
Let Google keep funding the large investment in the core Kubernetes platform while Microsoft incrementally funds an engineering team to make Kubernetes properly support Windows Containers. Take the money saved and invest in Azure platform services.
On the flip side, Microsoft could buy Docker and reinvent all the Kubernetes features, but it won't make a dent in Amazon market share because winning is all about the platform services.
The Kubernetes train has jumped the Google tracks
TechRepublic: Some in the industry believe that Kubernetes is really a stalking horse for Google. Google open sourced Kubernetes to accelerate more adoption of GCP. The more workloads that run on Kubernetes, the easier to move them to GCP. Why should Microsoft not play the same game with Docker? Couldn't Microsoft use an unfair advantage at the orchestration layer too?
Balch: Kubernetes can run on any public cloud, so really Google has no advantage. Many people, including CoreOS and Heptio, support Kubernetes on AWS. Kubernetes has been a great marketing and community engagement vehicle for Google, but I don't buy into the concept of a lasting competitive advantage for GCP.
Public cloud providers need to spend as little as possible keeping the lights on and as much as profitability will allow developing high margin, sticky platform services. IaaS and orchestration is simply keep the "lights on" basic stuff - a gateway drug to get customers hooked on the good stuff.
It's important not to forget that venture investment and running a healthy cloud business are completely different worlds for a company like Microsoft, both financially and strategically. The person who runs Microsoft ventures is compensated on investment return. The person who runs Azure is compensated on market share/revenue, growth, profitability. Sometimes the two intersect and enable each other, but that's not necessary or frequent.
- Why Microsoft should acquire Docker to better compete against Kubernetes (TechRepublic)
- Linux job market booms even as the server market disappears (TechRepublic)
- Kubernetes leads container orchestration (ZDNet)
- Docker: The smart person's guide (TechRepublic)
- Microsoft to buy Kubernetes container-orchestration vendor Deis (ZDNet)
Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.