Would it be crazy to run Storj in S3?

I’m exploring a wild idea based on how well Storj is running with iSCSI…

Conceptually, would Storj have any problems running on the S3 datacenter storage platform?

The working theory is that S3 IOPS are better and allows for faster access to files, but it’s definitely more of an edge case right now to present the datastore in a way that Storj likes…

Why are you looking to lose money?

  • Customers pay about half for Tardigrade as they would pay for the equivalent on S3.
  • The network adds 2.7x overhead, so for every TB stored on Tardigrade, there is 2.7TB stored on nodes.
  • Storj takes a cut
  • Open Source partners take a cut

I’m sure you can make something work if you run the node itself on AWS as well, but that would add even more cost. If you run the node on your own hardware, no matter the solution, the latency to S3 will probably be way too much.

People already run though ISCSI but on there own network and if your planning on doing it though the internet this is going to be alot worse then if on your network. You will loose races constantly.

Egress cost, amirite :sweat_smile:

1 Like

Wasbi doesn’t have ingress/egress costs, so it’s a predictable and flat rate

I’ve got a free year of cloud compute + S3 Object storage like Wasabi with no transit fees

The concept is what I’m going for to mount S3 + Storj and seeing what happens for two optimized connections between the datacenters.

1 Like