American productivity is being saved by smart architecture-is your internal infrastructure up to par?
Remote work is a decidedly modern phenomena, especially in the scale it currently exists. It was not planned for. as black swan events cannot be. As COVID-19 swept the country and one by one governors and mayors ordered non-essential employees to go home and shelter in place. As a result, hundreds of companies scrambled to become enabled for remote work. America went from a country where only 45% of businesses had formal remote work policies to 100% remote in about two seconds flat. The transition could have been significantly rockier. While you might hear of companies who had to purchase millions of dollars of laptops, IT departments flooded with helpdesk tickets from people who had never accessed VPN, concerns about the security of video conferencing systems or the challenges of balancing a full house with a full time job. One thing you’re not hearing? Connectivity issues.
All of that silky smooth video showcasing the Zoom 1990’s Macy’s family picture backgrounds is thanks to an infrastructure that was designed over 50 years ago by Vinton Cerf. There were a few groans at the onset, but quickly global providers were able to seamlessly manage the shifting of loads and surges without seeing a major outage. In the US alone, peak traffic was up by near ⅓, with major metropolitan areas seeing surges as high as 60% above normal. Name one utility that could handle that level of increase without seeing massive outages. You can’t, because utilities were not designed to withstand nuclear war. The internet — surprisingly enough — was. However, nuclear war or pandemic are not required to take the lessons of Vinton Cerf and apply them to your own infrastructure choices. Doing that is perhaps more important than ever as the scramble continues to bring previously not digitized services — such as small business loan processing — online. Underlying protocols of the internet’s infrastructure adapt to shifting conditions, working around trouble spots to find efficient routes, and managing glitches in ways that make sure you can access your spreadsheet (or your cache of cat gifs) from the cloud while tens of millions of other people do the same.
To break it down further, if you want a system that is always (or almost always) up consider the following best practices inspired by Vinton Cerf in your architecture decisions:
- Architect for uncertainty: When building systems for scale, you can’t anticipate every eventuality. You have to bake uncertainty, resilience and redundancy into your architecture to account even for ‘acts of god’ like COVID-19. Netflix for example has a tool called Chaos Monkey that randomly shuts down services in production to constantly test system stability and resilience.
- Simple > Complex: A simple, stable system always outperforms a sprawling and brittle one. For enterprise and large scale systems it is sometimes challenging to boil them down to their core but that’s the job of the systems architects. Identify common processes/uses and abstract them out into simple, reusable services.
- Favor protocol over oversight: Humans break things. Always. When building systems, you have the opportunity to create your own tiny world that can be scrubbed free of human error (thank you QA!). These systems don’t live in a vacuum. They’re often used by and governed by humans. Humans who are breaking things right now even as you read this. If you take that into account and simplify human involvement in governance by creating both technical AND governance protocols you can reduce the instability of the system over time
- Consensus over mandate: The wisdom of the crowd is greater than the wisdom of individuals (see Wikipedia and the Internet…for the most part). Bringing the power of smart, invested people to solving challenges is immense but it is hard to do this with mandates. Fostering open discussion and driving for the best solution for all can be a powerful driver of system stability and resilience.
Evan Rose, Founder of Rose Digital, Evan Rose has been building web and mobile applications since 2009. His focus is on usable, performant application interfaces. He attended the Harvard University and graduated with a degree in Social Anthropology. Evan has launched two venture backed startups, is a board member and investment team member of NFTE Ventures and chairs the Harvard Club Tech and Entrepreneurship Panel. Most recently he was a Senior Presentation Layer Engineer and Mobile Application Architect at Razorﬁsh.