Step On It

How to drive really fast on the Autobahn and still arrive safely at your destination.

popularity

By Nithya Ruff
On a recent trip to Germany, I was a passenger in a car with my colleague Tom De Schutter on the German Autobahn. We had just landed in Frankfurt and were driving to Aachen, Germany. The anticipation of the potential speed at which we would drive created both excitement and anxiety in me. What I learned about driving on the Autobahn was eye opening with how Germany allows for speed, but also enforces rules of the road to ensure safer driving. You need to have a decent car, and a driver who has nerves of steel and is trained with the rules of the road. It does help that the roads are in good shape and you are driving with others who have been trained from childhood to drive on the Autobahn. I don’t claim to be an expert and won’t go into the details of the rules, but will point you to this article on how stuff works on the Autobahn.

It did get me thinking about the parallels of driving on the Autobahn and how we are driving faster and faster in the tech world with getting products to market. The whole product life cycle is shrinking while the complexity is increasing. Lifecycles of 12 months or less are quite common in mobile and consumer, and even automotive and industrial cycles are shrinking. So how do you go fast and still remain sane? You clearly cannot do the same thing you did before and just attempt to do it faster. You need to systematically consider all aspects of the process to adjust for this speed. Going fast without training, rules of the road, better vehicles and better roads can be disastrous. Changes are needed.

Using yet another comparison, innovation in product without innovation in process and methodology, team dynamics and organization, is like driving a Maserati without any training in driving a fast car. So why would you sign up for developing your new product in a much shorter time with the same process as before? Everything from social dynamics—how teams work together, the HW and SW development process, the testing process, the training process and rollout to the field and customers—are all fair game for examination. When I wrote this article, that’s what I had in mind. De-risking a project and working with speed and yet quality is about changing the way we develop:

  1. Not just using PowerPoint, spreadsheets and discussions to make design decisions but actually simulating the products and doing what-if analysis.
  2. Not waiting till the end of the project to do 50% of the solution development, i.e. software, but instead starting early, incrementally developing and testing against targets available in each phase. For example, starting with models and moving to FPGA boards and sample boards and reserving only the final tests for the final boards. Why risk doing all your development on the final boards? Why not limit use of real hardware for the final test phase?
  3. Even HW development can be agile when you have multiple feedback iterations from running actual SW on the models even before RTL is created or committed to. Why depend on SW to work around any HW quirks? Why not change HW design early so both can be optimal?
  4. Even when all SW teams are not available to work on the project, start with the sub-groups that are available. Enable them with portion of the prototype that they need to get their work done.
  5. Why not do ‘what if’ analysis with models so you can fine-tune performance and power utilization? Why wait till you get complaints from customers to make the adjustments? The more we wait, the more expensive changes become.
  6. Multicore designs are one way to get more horsepower into designs. But if you have horsepower without knowing how to use it, that can be a huge waste. Developing software for multicore designs is still an art, and having more runways to do and test it with full visibility into the behavior of different cores is critical to its success.

More than anything, as I get older and smarter I realize that I don’t want to stress my mind and work whenever I don’t need to. Gone are the days when I crammed for an exam the night before and stayed up all night. I plan and spread out my work over the days I have to minimize the risk, stress and wear and tear. It has resulted in better quality work and a happier life. I also am happy to say that I survived the Autobahn and have a healthy respect for speed, and realize that it can be done safely with some planning. I wish all of you speed, safety, happiness and a stress-free holiday season.


Tags: