Google's secret to staying technologically competitive: global-scale engineering first

[[341436]]

This article is based on my personal experience as a software engineer at Google for six years and my observations of Google's extensive engineering operations. I hope this article can provide a high-level overview of Google's global-scale engineering operations.

I had the opportunity to work at Google on a variety of projects over the six years from August 2013 to April 2019. Most recently, I moved to another startup in Los Angeles as a founding engineer.

During my time at Google, I was deeply impressed by the tech giant’s ability to build and operate software at global scale.

Although I cannot reveal too many details, I believe that those who are interested have already felt the development and great advantages of distributed systems in the past decade. Hadoop itself has developed into a billion-dollar industry; in fact, similar systems within Google are not only more well-designed, but also more prosperous. It is the mutual cooperation of nearly 100 systems that supports Google's huge and prosperous business system.

Google considers global engineering as one of its core business values. Take the cloud gaming platform Stadia as an example. Although similar small-scale products have been launched many years ago, only Google has the ability to expand it into a truly global platform for users from all over the world to use at will.

Virtually every product at Google is designed to scale globally. Google has been working hard to invest heavily in the world’s most challenging technical problems, with Waymo (now an independent company under Alphebet) being the most prominent example. It’s clear that Google was willing to invest heavily in self-driving cars when they were still in the research stage. Accordingly, any project that can’t serve a global customer base will be downgraded or even killed. A recent example is Google+, a social networking service that Google had been trying to get to customers, but was shut down (for non-commercial customers) after realizing that it wasn’t gaining traction with the audience.

Google's global engineering capabilities are reflected in several key areas:

People management: Global-scale engineering necessarily requires a global-scale engineering team. Google has more than 40,000 world-class software engineers and an equal number of non-technical people who are also indispensable in daily operations. They have offices in dozens of countries/regions around the world, and it is this global business system that allows Google to attract many top talents. They have very different personal and professional backgrounds and join Google with different beliefs. Generally speaking, giant companies of the same size as Google tend to establish their own global offices to handle relatively minor tasks such as corporate performance and revenue. But Google's situation is completely different: Take its Borg team as an example. In addition to Google's maib campus, Bog has several important members in Warsaw responsible for development and maintenance, and the SRE team is in Zurich. Since its establishment in 2003, the various groups under Borg have maintained this distant but closely cooperating collaborative approach.

Technology: Technology is the foundation of everything, responsible for providing tools and guarantees for people to collaborate, optimize operations, create business opportunities, and realize innovation. Global engineering organizations cannot rely on third-party suppliers. This kind of technical achievements is very difficult to build, but once completed, it is enough to become a core element for enterprises to gain a competitive advantage. Google provides engineers with a complete set of developer tools to help them unleash their talents and inspirations. Google has also achieved many breakthroughs in distributed systems, AI/ML, and security, so that its software always maintains high performance, high reliability, and high security. Google has a number of world-class blockbuster software products, which have an unparalleled strong position in terms of scale and quality. Statistics show that Google has a fantastic performance in building and running large software systems, demonstrating advanced technical expansion capabilities that are unrivaled by its peers. More importantly, Google not only built such a system, but also proved the rationality of its technology investment through excellent market performance and the hands of a global user audience.

Operations: How can we make full use of technical infrastructure? How to properly address the goals and risks in short-term and long-term projects? For this reason, Google created SRE. Unlike traditional operations teams, SRE needs to devote at least 50% of their working time every day to writing daily operations automation software. This can not only ensure continuous investment in operational excellence, but also ensure that infrastructure and applications have excellent quality levels. In addition, data-driven decision-making has long become an important part of every Google employee's daily work.

Business development: The capabilities achieved by integrating these factors must ultimately be reflected in products with actual business value. Since the author is not familiar with this topic, I regret not being able to provide a detailed introduction.

Looking at the entire infrastructure system, Google has a comprehensive technology stack in its products. From data centers, transoceanic fiber links, customized servers, programming languages ​​and developer tools, to multiple operating systems, basic cluster services, middleware, application frameworks, big data, AI/ML frameworks and tools, and even consumer-grade hardware, Google has always firmly held its position in the world's first echelon, and has even achieved the undisputed highest level in many fields.