Comments by "Biomorphic" (@biomorphic) on "IGotAnOffer: Engineering" channel.

  1. There is no need to store the files in the CDN. You do serve the files through the CDN, and the CDN (in case of AWS is CloudFront), requests the file to AWS S3. For the developer this is transparent. The CDN is distributed, so based on the user location, it will serve from the closest server farm.
    9
  2. ​ @AppleGameification  Because this is done automatically when you request the file the first time. It's done by CloudFront itself. You do request the file through the CloudFront, then CloudFront check its cache, and if the file is not there, it gets the file from the connected S3 bucket. And it caches the file. The CDN does that, you do not need to store in the cache yourself. This is done automatically connecting CloudFront (the CDN) to the S3 bucket. Your app doesn't even know that you are using S3, because the files are always served through the CDN. At the same time the CDN is auto managed, you do not need to worry about caching and zones.
    6
  3.  @Solo2121  indeed, that is what a CDN is for, and that is what a CDN does, it's called transparent caching.
    5
  4.  @deepspace9043  if you speak about a specific technology, CloudFront, then you need to know how it works, otherwise do not mention it. No one would ask you how to implement a CDN, that was not the purpose of the system he was asked to design. He made a mistake when he said he needed to cache the files. Just say: to serve the files I would use a CDN. Then, if you know one that you would use, mention it, and say how it works, to show that you know what you are talking about, otherwise do not mention it, and do not dive deep into details unless you have been asked to. To give you another example, the guy said he would have used PostgreSQL (if I remember correctly), but of course there are many other choices, especially if you are on AWS. As the interviewer I would have asked why PostgreSQL and not Aurora or DynamoDB. In the moment you become specific, you may be asked more detailed questions about your choices.
    2
  5.  @Solo2121  no, it doesn't work like that. You configure CloudFront (the CDN) to use the S3 bucket. Then you access to the files through CloudFront, e.g. https://d111111abcdef8.cloudfront.net/my-file.jpg There is no need to store the files in the CDN, because CloudFront will retrieve the file from the S3 bucket if the file has not been cached yet. Caching is done automatically by CloudFront. CloudFront is the CDN.
    1
  6.  @iFireender  If I was the interviewer (which happens quite often), I would have indeed asked a clarification when he said that you need to store the files in CloudFront, because that is a wrong assumption.
    1