Implementation of node classes depending on their performance

Hi! Does Storj assume adding a data type breakdown that is stored on the network?

I believe that if someone (for example) wants to store in Tardigrade a backup from their computer that they will need only if their equipment in the home fails, this data does not need to be stored on nodes with high performance and excellent connection.
This type of data can be stored on slower nodes (for example, those that are running on Raspberry Pi).
It’s something like S3 Glacier

In contrast, files that are regularly downloaded or modified should be stored on high-performance nodes capable of transmitting large amounts of data at the same time.
It’s something like the S3 Standard.

Tardigrade users could choose in which layer (hot or cold) they want to store their files.


Low powered nodes on RPi’s are actually surprisingly capable of being part of high speed transfers. The storagenode software doesn’t require that many system resources to perform well and because all transfers are distributed, individual nodes don’t need to have a massively high transfer speed.

So while I like the concept, I think these kind of tiers simply don’t seem to be necessary. Besides, having all nodes take part in all kinds of transfers is better for decentralization as well. I also think people running on RPi’s wouldn’t be happy with a change like that. They shouldn’t have to pay for other SNO’s running on massively overpowered hardware.


There is also difference in uplink speeds - some people have 100mbps, some have 1gbps and some have whatever the minimum allowed is.

1 Like

I know I wouldn’t be happy, running 2 RPIs and just upgraded my ISP to be able to contribute to production :slight_smile:

For little bit of thinking I’m completely agree with you. Thank you for explaining.

1 Like

I think upload speeds would be more of a bottleneck than the hardware. I have only access to DSL in my area, which has terrible (upload) speeds.