Learning DevOps from Nature and a 125-year Old Food-Delivery System

January 30, 2018 | By Zakir Hossain | Filed in: Technical Blog.

Nature has an advantage over us. It has infinite time to develop complex systems and make mistakes. So nature can go through the iterative process of trial and error and develop operations that are awe-inspiring. These complex systems are large in scale, yet they are based on simple rules. They are a DevOps team’s dream come true.

Our roads, our factory assembly lines, and our computer networks are reflections of the same iterative process. We have built these over long periods of time starting from simple principles. In DevOps, we are trying to create similar patterns. So the dabbawala business, India’s 125-year old food-delivery operation, gives us a unique opportunity to study a system that is built on simple rules.

Dabbawalas: Delivering Complexity With Simple Steps

In Mumbai, dabbawalas deliver home-cooked lunch or tiffin from their clients’ houses to their clients’ workplaces. They serve 200,000 customers every work day. The Harvard Business School has given them a “Six Sigma” rating because the dabbawalas make only 3.4 mistakes per million transactions.

People depend on dabbawalas for the following reasons:

  • Carrying a tiffin box on the train during the Mumbai morning rush hour is very challenging
  • Consumption of previous day’s food is considered unhealthy
  • Religious reasons prevent people from eating restaurant food

If you look at the numbers, 200,000 is not a large number in a city of 18 million plus people. But these dabbawalas navigate the vast and complicated labyrinth of Mumbai to deliver food from one point to another with the highest level of efficiency in the modern world.


The Six Sigma Dabbawalas of Mumbai

From the home of the client to the client’s office, a single tiffin box will go through a complex set of handovers and different modes of transport. The whole process is dependent on simple, chalk-scribbled codes on the tiffin boxes. Most dabbawalas are illiterate so they can recognize only a few letters. Yet with a two-tier management structure and no technology, the system works with accuracy and precision. The dabbawalas ecosystem reflects the ideas that DevOps teams want to implement – an efficient process without complicated tools.

Murmuration: When Birds Follow Simple Rules

The sight of a flock of starling creating complex patterns in the sky or murmuration is a great spectacle of nature. Each individual starling remembers these simple rules:

  • Stay close to the birds near me
  • Don’t bump into my neighbors
  • Fly at the same speed as my neighbors
  • Move towards the center of the group

devops-starling

A single bird following some simple rules can create a complex whole.

 

When you look at murmurations, it is difficult to fathom how you can design it from the outset. It will be an overwhelming task. If you try to come up with an efficient system that is fail-safe, you will probably end up with starlings clashing into each other and dropping from the sky.

 Why Should DevOps Look at Evolved Complex Systems?

It seems most engineering teams start at the point of complexity. If you don’t have simple steps at the start of your process, the latest shiniest technology will only create chaos. Studying natural complex systems and understanding their simplicity at the fundamental level can help companies build better workflows. Jenkins, Docker, VMWare, Ansible, Chef, Puppet and the Cloud are not what is going to save an organization. They can even hinder progress by introducing chaos.

In conclusion, DevOps is not an easy way out. Remember nature takes a lot of time to create complex systems and dabbawalas are efficient without any technology. If you want to create a long-term company, concentrate on simple steps. Complexity is not the answer. Be an ascetic. DevOps is not an out-of-the-box solution. It needs time, deep introspection, patience, and understanding. It is technology independent. DevOps is a philosophy to solve long-term problems. And it needs time to evolve.


Tags: ,

Comments are closed here.