So just how much "IPFS" is in dtube at the moment?

Hello everyone,

I'm just merely starting to use Steemit and dtube and wrapping my head around the ideas and technology these platforms are using. Being a quite tech-savvy person, I'm digging the idea of a decentralized metadata and data repository in form of a blockchain and the IPFS, respectively.

To my understanding, uploading a video to dtube (with respect to storing and publishing the data) consists of at least two crucial parts: Appending the Steemit post into the blockchain, and adding the payload (i.e. the source video and possibly re-encoded versions) to IPFS. In my mind, the latter part is realized by means of dtube operating an IPFS node and providing hashes to the different versions of the video for everyone to consume.

I can certainly imagine that the amount of data being uploaded to dtube on a daily basis is huge and thus older or less popular contant (currently) needs to be purged at some point. For example, some videos that I uploaded a few months ago don't seem to be available anymore. After some thinking I'm actually okay with this at the current state in order to keep the platform running and new content coming.

Now I thought, if I wanted to keep the content I uploaded available for longer, maybe I could set up an IPFS node myself for my own content and keep it available as long as I feel fit. Since dtube provides me with the IPFS hashes of the original video, the re-encoded files, and the thumbnails, it should be easy for me to pull these files to my IPFS node and (re-)add them to the network such that in the future, people could still access these files from my node.

This is a fundamental idea of decentralized content, right?

However, this is where I'm stuck at the moment. I set up a basic IPFS node and started experimenting with pulling files from the IPFS network. This seems to work fine for the examples provided by IPFS itself, as well as for some thumbnail images. However, trying to pull the hashes of any video files using the hashes provided by dtube seems to be stuck forever, not pulling much data at all, even if I keep trying for some time.

Inspecting the URLs from which dtube videos are pulled on the dtube website itself, they always seem to be downloaded from "". And indeed, I can download the videos using the different hashes using this particular host just fine. However, using the same hashes on the "regular" IPFS network just doesn't seem to work.

Conversely, trying to access other files – such as the IPFS example files – using the "video.dtube.top" host results in empty responses.

So my basic and very fundamental questions are:

To what degree is the "video.dtube.top" domain running a regular IPFS gateway? Is it particularly tailored to hosting the videos uploaded to dtube (only)? And why do the videos not seem to be available on the regular IPFS network using the hashes provided by dtube? What are the possibilities to truly decentralize the content away from that particular dtube domain?

I could probably download the generated videos using the aforementioned host and re-add them to the network. Would that work to keep the videos available? However, I'm really wondering why I need to use that particular host to access the files in the first place.

I'm looking forward to some insightful responses!