2023 Jul 6

Cloud Unnative

What is “cloud-unnative”? It sounds like a tongue-in-cheek term promoting architectures that do not fit into the cloud, but it is not. Actually, I have spent much of my career helping companies get their architectures into forms that distinctly are cloud-native, much of it long before the term “cloud-native” existed. Many cases make perfect sense not to go into the cloud. The design principles, however, can be enormously helpful towards operating more efficiently and reliably, within or outside the cloud.

2021 Nov 12

Not Such A Paradox

Is the cost of cloud a “trillion dollar paradox”? Legendary venture capital firm Andreesen Horowitz thinks so, based on their blog post from late May. The above article has been making the rounds for several months, causing a lot of teeth-gnashing and lining up of parties on either side. Is it evident that cloud puts “pressure … on margins (that) can start to outweigh the benefits, as a company scales and growth slows”?

2020 May 12

Bullish on Automation

Every pundit under the sun has their theories about how the “world will change” post-Corona. Personally, I have never liked airy pronouncements about how “NOW it is different.” I prefer more measured responses, like Ben Thompson of Stratchery, who has said in his articles that the responses to Corona would accelerate existing trends. I do, however, think that one trend will reverse direction, and that has a major impact on markets: supply chains.

2020 Apr 2

Inception, Kubernetes Cluster Managers

Inception When the cloud-native world really got under way, especially the open-source part of it, much of what we used to do (and, likely, most companies still do) in custom and proprietary ways became standardized: defining workloads defining storage defining dependencies defining policies defining placement defining replicas load balancing rollout strategies packaging status many other things The tool to do this, of course, began with docker, which addressed, primarily the packaging and workloads part, but grew to Kubernetes as a basis for the rest of it.

2019 Dec 6

KubeCon San Diego 2019 Observations

Thoughts on KubeCon North America 2019 In late November, I spent several days at KubeCon/CloudNativeCon, for the fourth or fifth time. It certainly has grown over the years; San Diego’s conference was oversubscribed at 12,000 attendees. In the somewhat snarky words of a friend from the Linux Foundation, “this feels a lot like the OpenStack conference at its peak.” I am hopeful that it doesn’t go down the same path.

2019 Sep 3

Kubernetes Enabling Moving Up the Stack

Yesterday, I had the privilege of one of my many discussions on technology direction with Josh Mahowald. Our conversation turned towards why it is that so many interesting and enabling pieces of software have been built on top of Kubernetes. As Josh put it, there is an explosion of higher-level services and tooling. This will be the first in a series of posts examining what Kubernetes really is (ok, an API), where it fits in the history of technology, and why it has enabled this kind of “explosion”.

2019 May 7

Change Process vs Change Execution

Why do I need both GitHub/GitLab/Bitbucket and JIRA/PivotalTracker/etc.? Over the last several years, while working with companies that regularly deliver software, I have seen three basic patterns in use: Git-All-In: These companies, normally founded by engineers, run all-in on git platforms like GitHub/GitLab/BitBucket. The Issues tracker and Pull Requests/Merge Requests (I will stick with “PR”, but no offense to GitLab) are the primary tool they use to track activity. As they grow some, they use the various “project”, “kanban” and other organizational features added to these platforms, but the primary method for interacting with daily work and knowing their state is the Git-based platform.

2019 Mar 7

How to Run a Great Conference

How do you run a great conference? I spent the first three days of this week, Monday March 4th through Wednesday March 6th, attending and speaking at QCon London. In my case, I spoke about LinuxKit, a toolkit for composing lightweight, minimal and optimized runnable operating system images. Those who know my focus on technology operations might wonder why I gave a talk on so deeply a technical subject as composing operating system images.

2018 Nov 26

Serverless vs Containers is Silly

Serverless? Containers? Who will win??? In the week of aws reInvent, when 45,000 or so people are descending on Las Vegas, and two weeks before the big cloud-native conference in Seattle, the question of “which is the future, serverless or containers?” seems to be the “debate du jour”. For example, take last week’s debate: I am going to posit a different position, one which, in the true spirit of compromise, leaves everyone equally unhappy:

2018 Jun 19

DockerCon Observations

A few weeks back, I wrote an article on my observations on KubeCon/CloudNativeCon. A number of people asked that I follow up with similar observations about DockerCon. Last week, I had the unexpected pleasure of attending DockerCon in San Francisco. I say unexpected not because I did not expect to attend, but because I did not expect it to be so pleasurable. Indeed, I fully intended to cancel my ticket, until I was asked by several colleagues to attend with them.