Maybe it was the hustle and bustle of Times Square just within earshot. Maybe it was the smell of that legendary pizza on every corner. Or maybe it was having a front row seat in the financial capital of the world while the drama of a possible government default played itself out like a Broadway show. Whatever it was, Velocity New York felt markedly different than its west coast cousin.
But discussions on scaling the operations of the squishy side of the organization–developing healthy team interactions, talking through the realities of the ever-elusive DevOps culture, and what it all means to scaling not just web sites but full-fledged businesses–were more present than ever.
Operating on the edge
What was especially notable about this Velocity was how front-of-mind knowledge and lessons from various disciplines was: Dr. Richard Cook, an anesthesiologist, gave a keynote on operating on the edge of failure pointing out the future of our web systems is safety: as these systems move toward the center of the business, they become business critical systems. Whether they are directly responsible for safety–say as in a fire truck dispatch system that might have originally been intended for taxicabs, or a more indirect task, like being able to sign up for federally mandated health care–these systems eventually tend toward a critical focus on safety. More than ever, the web sites and infrastructure we build and operate every day have more in common with nuclear power plants and the National Airspace System than they did with the boxed-packaged-and-shipped software of the last century, and we need to prepare ourselves and our organizations for this shift.
Business is taking note of this too: Fred Wilson, a partner at Union Square Ventures, gave his keynote (which he continuously deployed on his blog as he gave the talk) on eight management-oriented rules for scaling technological systems. Some of his insights included: fear-driven organizations do not scale; sometimes the only solution to a continually problematic “component” is to rewrite or transition it out; and real time metrics collection and feedback mechanisms for the human side of organizations is critical to course-correcting and system-wide success.
Performance in real time
The theme of real time everything in the city that never sleeps was a palpable undercurrent of numerous presentations: Facebook’s Phil Dibowitz discussed how they scale configuration of their systems so it’s possible for developers to directly control the infrastructure of the social networking giant; folks from the venerable Stack Exchange described (and open sourced) their tools for real time monitoring to keep developers’ go-to sites running on an impressively modest set of hardware. CafeMom gave a cringe-inducing and awe-inspiring look into what happens when infrastructure meets a hurricane, providing a special perspective on reacting to truly real time events. Speaking of real time events and responses, Github’s Marc Imbriaco shocked the crowd with how they accomplish numerous tasks within the highly-geographically distributed team, and how Github trusts every engineer with access to everything from updating routing tables to changing storage array settings–often via their IRC chat bot, Hubot.
Many presentations reminded us all that success (or failure) is not just all about technology; being able to healthily react in real time to changes, data, and situations requires a culture built for doing just that. Dave Zwieback delved into the details of anti-fragile systems, and how failing fast is only a benefit when you’ve baked in a method to avoid brittleness into your infrastructure and organization. Andrew Shafer wrapped up the culture track at the end of the conference by reinforcing this idea: the organizations that win are those that are learning organizations. For those organizations, there isn’t a shortage of talent, because they cultivate it, and help their people grow with the company.
When you look at possible solutions to these problems across organizational boundaries, it becomes evident that what you need to be building are real time businesses. You need to tune your ability to respond to real time data, both from systems and humans; optimize the human factors related to real time deployment pipelines; focus on the dynamics of how your engineers respond, in real time, to outages and failures; deal with complex systems that evolve in real time, and work to integrate them and reduce their complexity whenever possible. Taking this holistic approach is the only lasting way to increase your velocity.
[Eds note: We'd be remiss if we didn't mention that the full video package for Velocity NY will be available soon here. Also, we're returning to NYC next year, so mark your calendars for Sept 15-17, 2014.]