Single point of failure on Tardigrade


The big cloud storage providers run on a few data centers. Tardigrade runs on a globally distributed network, meaning there’s no single point of failure - which is why making the switch makes sense.

Is there really no single point of failure. What is with the satellite?

I misunderstood this myself…

However, it seems the satellites function like bookkeepers not as data providers or distributors. The data pieces are sent directly to the decentralized Storage Nodes.

But when the sattelite is down nobody knows where the data is. So single point of failure


Correct, Storj still has a single point of failure (I think?). However, since the satellites only store the metadata it is very easy to make backups and run multiple redundant instances. (Compared to needing to have datacenters with a lot of disks in a lot of regions).

It is not a trustless system like Bitcoin. Contrary to currencies, decentralized data storage is pretty much impossible to do without centralized servers, at least not in a reliable and efficient way.

interesting, we shall see, I hope that satellite owners have redundant hardware available to quickly recover if need be.

I would be interested to run a satellite myself, does anyone know the hardware requirements in terms of IOPS, CPU, RAM, SSD space, etc?

If you are going to run a satellite there are other things you should be more concerned with; you will need to manage payments, get clients and also convince SNOs to add your satellite to their storage node(s).