In September 2015, Google engineering manager Rachel Potvin revealed details about Google's software code at an engineering conference. She revealed that the entire Google codebase, which spans every single service it develops, consists of over 2 billion lines of code. All that code is stored on a code repository available to all 25,000 Google engineers, and the code is regularly copied and updated on 10 Google data centers. To keep control, Potvin said Google has built its own "version control system", called "Piper", and that "when you start a new project, you have a wealth of libraries already available to you. Almost everything has already been done." Engineers can make a single code change and deploy it on all services at the same time. The only major exceptions are that the PageRank search results algorithm is stored separately with only specific employee access, and the code for the Android operating system and the Google Chrome browser are also stored separately, as they don't run on the Internet. The "Piper" system spans 85 TB of data. Google engineers make 25,000 changes to the code each day, and on a weekly basis change approximately 15 million lines of code across 250,000 files. With that much code, automated bots have to help. Potvin reported, "You need to make a concerted effort to maintain code health. And this is not just humans maintaining code health, but robots too.” Bots aren't writing code, but generating a lot of the data and configuration files needed to run the company’s software. "Not only is the size of the repository increasing," Potvin explained, "but the rate of change is also increasing. This is an exponential curve.