I was headed west for a quick trip to speak on an analyst and industry panel about the future of cloud networks in Santa Clara.
The note I sent was mostly logistical. I would be in town Monday and Tuesday, and I wanted to see if a friend in venture was around for a drink. That combination felt normal by 2014: fly across the country to talk infrastructure, then use the small open space around the event to reconnect with somebody in the startup world.
The panel topic was broad, but the work behind my point of view was not theoretical.
At Shutterstock, we were running virtual machines through OpenStack, connecting cloud and physical environments, automating network changes, and building private paths between offices, data centers, and a DDoS mitigation provider. The previous few months had made one thing especially clear: a cloud network was not only about creating subnets through an API. It was about where traffic entered, how systems failed over, which services stayed public, and what happened when somebody deliberately tried to exhaust the path.
I had talked about our OpenStack system in Hong Kong the previous fall. Santa Clara was another step in learning to turn operating experience into a useful public argument.
That translation is harder than it looks. Inside the company, the conversation can depend on the latest graph, the specific hardware, and a pile of shared history. On a panel, the useful part is the pattern. What changed as we scaled? Which assumption failed? What should another company consider before copying the design?
The travel was fast enough that I described it as headed out west for Monday and Tuesday. There was no long tour around it. Speak, listen to the other panelists, compare notes with people working on adjacent problems, grab a drink if schedules aligned, and get back to New York.
I liked that rhythm. The company work gave me something real to contribute, and the industry conversation brought back new questions for the company work.