I’ve read that it is useless to have a node larger than 24Tb, as the ingress/sent to trash amount of data becomes the same. But sometimes i read of operators having large nodes with many hdds in raid-z etc. How large is the largers node in Stroj-network so far?
This limit was just an estimate based on typical traffic in the past. @BrightSilence has recently changed this estimate to ~75TB.
Besides, this is just an estimate for a single /24 IP block. If you have nodes within different /24 blocks (e.g. you have network connections from multiple ISPs at home), then you can cross this estimate easily.
does take a long time to get there…
got a 2 year old node that is like 6.5 TB
seems a bit small for its age…
had an 18.5 TB node that i GE’ed a few months back, it was like 3 years or so.
but had gotten the last push of the test data pre DCS launch.
the $ pr TBm wasn’t great on it either… due to low egress on the old test data.
needed to free up some space so got rid of it…
a bit sad i did…
the largest nodes currently in the network is around the 22-24TB
not to many of those around tho… i know @BrightSilence had a node with more data than mine.
i think that is about the max, because there was a network data reset 3 years ago, prelaunch…
so if a node has gotten most of the ingress since then, it would be among the largest.
That estimate was more useful when it was still realistic to get there in any reasonable amount of time. The earnings estimator shows the first 10 years and by the end of that you would have about 45TB if the network situation remains the same (which it won’t over 10 years). So sure… if you would have 75TB today, it would probably remain fairly stable and stop growing. But nobody has that on a single node. Right now the biggest nodes are probably around 23TB and those are still growing quite fast.
These nodes are definetely not a one hard drive nodes. Are they using raid or zraid?
All nodes behind the /24 subnet of Public IPs are treated as a one big node for uploads, so they can be easily one-disk nodes.
If they are single nodes, then probably yes. I personally don’t recommend that, but my largest node runs on a RAID6 (like) array as well, simply because it’s a multipurpose array and my node is using space not used for other purposes. I started just using this free space on existing hardware because it was the cheapest and easiest to set up. All my newer nodes run on individual HDDs as that is just more space efficient and the original node made enough money to pay for those.
But they are separate for uploads and their identities are unique.
only for downloads, uptime and audit checks. Uploads are distributed between them as if they will be a one big node.
All wordings from the customer point of view.