When I met with Ben Golub and Solomon Hykes at the Doug Fir Lounge in Portland, it was 2013 and OSCON was buzzing. Golub had found a home with Docker as CEO, following a successful run leading Gluster, the storage company that sold to Red Hat.
Today, Golub is now the interim CEO at Storj and Hykes just a few weeks ago left Docker. Their departures are a discussion for another day. But they took the lead in opening the market, and since 2013, we do know that containers have been widely adopted. The question now is about application-oriented architectures and the resources that they run on in an ever-distributed manner. That is a topic that transcends discussions about containers and into topics that orient on application infrastructure and the influence of data architectures for building new predictive-modeling systems. These are event-driven platforms that run on resources configured for the workloads, largely with containers running on Kubernetes or Kubernetes-style architectures. In many respects, the biggest shift in five years is the abstraction from the machine itself that was on the verge of happening when Golub, Hykes and I had lunch at the Doug Fir.
Storj reflects a shift in conceptual thinking about abstracting the data from machines. It addresses a demand for unused resources. Networks are increasingly widely distributed, but the utilization of resources unused on people’s devices is mostly untapped. The resources that can be drawn from devices is what Stor makes available for developers and service providers. The differentiator, Golub will say, is in how Blockchain is integrated into the platform.
“It is Airbnb for disk drives,” Golub said.
It’s funny talking to Golub as CEO of a new startup. There was just an energy about that OSCON summer show where we met with Ben. Maybe it was the warmth, the sun, out on a bike in Portland with so many geeks in town. But there was something else. Application architecture thinking had shifted with APIs, and there were new platforms to drive use of cloud resources, in particular, Amazon Web Services. There was a puzzle about it all — how did the resources fit with application architectures that were elastic and could scale? The pieces of the puzzle were showing patterns — there were threads starting to develop. But it was still not easy. Build a LAMP stack on cloud resources? No thanks — there was something that could be different that developers would adopt. They had all these resources available — how could they be more easily used?
Now we know how things have changed since then. But look at us now. The resources are ever distributed. And Golub is back in the storage game.
In this Edition:
3:39: What is Storj?
8:42: How does the architecture achieve the economics?
12:28: What’s the blockchain architecture that makes this different?
15:16: Edge computing, parallelism, and storage
18:32: How is this a platform that will evolve?
22:51: What are the architecture challenges you have in thinking about building a scaled-out, distributed storage network?
Feature image via Pixabay.