Bandwidth utilization comparison thread

sad bee noises
I got yesterday some requests, that pinned my upload at maximum for a few hours. But overall my stats are pretty consistent. Nothing special for me

But storage is increasing at a pretty good speed. This month I got about 500Gb more storage than in the beginning of this month.

I have that big egress, starting the same date, on 2 nodes out from 18. They are 2 years apart as age.

So for the last 10 months, I’m pretty leveled at ~110 TB stored on 18 nodes/10 IPs. I hope for the next 12, we will see an up trend, but… I don’t know. June was awful, showing only +580GB more than may.

2 Likes

My nodes seem to have stabilised at roughly 10TB per IP

1 Like

Up about 12.5% of storage space for June, or thereabouts, n’ stuff.

2 cents,
Julio

… and the bit goes on… :man_dancing:
(if we were back 4 years ago when egress was 20$… :star_struck:)

3 Likes

Wow, only one node does so much for you?
Even all nodes combined (plus egress and ingress combined) won’t achieve 3Tb for me. :o

Only those 2 nodes from last month. The other 16 are “normal”. They got some lucky pieces.

Ah ok, thought my nodes were performing really badly.
This is one of my nodes alone:

You definitely have some pretty princess data on those nodes. The egress exceeds my combined ingress and egress on all my nodes so far this month.

Holly molly, one of my (only 2 :frowning: ) nodes get crazy today (this node is quite new, <1TB of data yet):


I know those are funny numbers for big guys here.

2 Likes

I have one lucky Node, too

Network goes brrrr

oh I got my first magic data fragment yesterday on one node. normal egress is around 8GB. yesterday it was 146GB

If storj was my company I would move those segments to company owned nodes… :money_mouth_face:

With current cost $ per TB it’s few cents.

This would implicitly suggest that you either do not trust your own storage network or are considering betraying your SNOs. I would not run a node for your company. :thinking:

3 Likes

Exactly, let alone competing with its own vendors; this is something Microsoft would (and did) do. Don’t be Microsoft.

2 Likes

Surge nodes were exactly that, BTW. Nodes spun up because Storj Inc. didn’t trust the community network to handle elevated traffic.

3 Likes

I think it’s different – keeping a horde of on-demand nodes across the world in various datacenters is a very good idea to absorb spike demands, or perhaps to improve CDN usecase performance: if there is a file that is accessed millions of times from across the globe every day – it makes sense to replicate it to a horde of low latency small nodes – this will improve apparent performance for everyone, and cost storj less money. (Eventually this data can further migrate to regular nodes if it remains in demand, with perhaps signinificantly overkill erasure coding scheme, or kept on on-demand nodes only, that (either data or node) can be thrown away once demand passes.

There is no competition – nodes serve stability needs, on-demand nodes serve performance needs.

Then if demand persists long term – storj can facilitate more operators to join – which is a slow process, unlike spinning up VPSes from a disk image.

Not doing this would be reckless and detrimental to network stability.

1 Like