CDN is a distributed network of caching servers that can briefly keep some of your data closer to users.
If you have a popular website that you hosted on object storage, you can optionally connect a CDN. In this case, your customers can receive “hot” data, such as the main page of your website, not from the main server, but one of several caching ones. This will increase the speed of access to this data.
Unfortunately, CDN only works well with a small amount of “hot” data. For example, with the preview for pictures, HTML, and CSS files. If your clients start navigating to less popular pages or downloading large files, CDN servers will be forced to go to the main repository to receive data, transfer it to the user, and cached locally, and this will greatly slow down the downloading of content.
In general, a CDN could help you out in some situations, but this does not solve the problem that occurs when you use the first generation of object storage. CDN cannot guarantee the availability of necessary files near your clients; it can only, under certain conditions, cache part of your data closer to clients for a short time.
That is why classic object storage with a connected CDN cannot work as fast as a geographically distributed storage can.
We understand that the CDN cannot significantly increase the speed of our service, but JUCE still plans to launch its own CDN in the future. Giving the best possibility is the main principle of our work.