This week we spent time learning about how Google and how they handle the infrastructure for services on such a large scale. One of the solutions involved in managing this problem is by automating what you can. While some automations are simple, like executing a shell script or monitoring machines, automating scheduling becomes more involved and once you have multiple machines with different environments, you must create further strategies to manage processing at that scale. One concept that stood out to me in this chapter is the concept of containers. While I've used Docker for multiple classes now, I wouldn't say I have a firm grasp on the advantages of using it. For example, why couldn't you just provide me with the dependencies I need? In this current project, I've actually gotten a small taste of why having a contained environment would be useful. For example, when designing the CSS for this project, I just messaged my partners to tell them to install bootstrap and react-bootstrap for their projects. I can only imagine how complicated it would get with multiple engineers working on multiple processes that all require their own dependencies.
No comments:
Post a Comment