-
Cut that negativity.
-
Chia guys can’t handle it,
They got scared as soon as the HDD is 100% loaded
Only pro SNOs like us, can do the job like a man
and we will. got 1 x 16TB ready to add, its offline now.
2 Likes
Just for discourse’s sake: How would it look if storj is citing “exponential growth quarter on quarter”, having multiple prospective clients, asking the SNOs if they have any available capacity to be brought online, benchmarking the network to see the min sustained limits (“bandwidth benchmarks will continue for a while”), panic-deleting all data that can be deleted…and in the end…the network grows by 1PB stored data a year?
Honestly? Way worse than missing a client’s estimates.
1 Like
I guess in the end our varied opinions here won’t matter: Storj said they’re going to have SLC shovel data into the network regardless. So they’ll see if more capacity is brought online as more nodes fill. Don’t need opinions for stuff that can be measured 
They didn’t say that. They said they will shovel data with a time-to-live (TTL). So I’m playing Devil’s advocate here. We don’t know the TTL. I’ll say it’s less than a month, based on a few replies back that capacity should scale up faster than 1 month it takes for someone to notice and order new drives.
So they have two scenarios: Either upload TTL’d data and have them deleted in a month, which brings us to the exact same spot we are today, or they keep adding data as old data is purged. How is this better for storj? keep wasting money on their side (sustain the current size) and wasting our money (expanding, electricity, ISP)? Why not get to grips with the reality that the network is at X PB today and SNOs have said (and also showed, so proving) that they will add capacity as things go from X to X+1?
I could totally see them having an internal estimate of the next-3-months of expected capacity expansion… and making sure they’re always reserving that space with SLC data. If it’s not too much of a financial burden it makes a lot of sense. Because even 3-month-TTL data can still be deleted early if they needed to.
I can’t wait to see what happens!
I have no intention of defending anyone, just sharing my opinion, but for discourse’s sake:
My take in your points are the following:
- “Exponential growth” - Total used capacity, from the info available in storjstats.info and ignoring current deletions, has had a steady growth. More importantly, the latest Town Hall does show that customer data seems to be increasing exponentially. Im not saying this trend will go on, just saying that up to now, growth of customer data, as shared by the Storj team, appears good.
- Asking SNOs for more capacity: Here this is a complicated topic. I personally think it is better to allow the network to grow naturally without having to request additional capacity, mostly because then Storj is in a position where if they fail to meet expectations, they will lose trust with the community. On the other hand, Storj is also trying to maximize growth, and for that they need capacity (if the growth comes, is a different matter). Also keep in mind that the request for more capacity has been said as replies in the forum (if I remember correctly), so far there wasnt been a topic asking everyone to add more capacity now.
- Benchmarking the network: I see absolutely no downside to this. Storj must understand every single aspect of the network to optimize it and make the best possible deals. If it can be done better or could have been done better, for that I have no idea as I have not followed it closely enough.
- “Panic-deleting” data: I have already stated that in my opinion the deleting of data is not ideal short term for SNOs, but if what you care about are long term prospects, then its fine.
I completly understand your point, that if even with all of this, growth does not meet expectations, well lets be honest, SNOs would be dissapointed. The Storj team would be dissapointed. Is it better than not meeting the requirements of a customer? In my opinon, Yes. Failing to meet customer expectations and, even more importantly, requirements, will in turn stagger future growth, which is more important than the growth right now.
Not saying I would be happy, or that I would even continue being an SNO if this happened (I have not considered what I would do), but for a company I believe it is still best to ensure customer requirements are met.
With topics as complex as these the final opinion will obviously change for everyone, depending on their goals, and general view.
1 Like
Here is a graph from EU1. The scales are very important (and why I don’t use storjstats):
I don’t see any hint of the network having any issues whatsoever keeping up with demand. In fact, I don’t see any issues with the network even coming close to overturning the added data:added capacity tipping point. If we are expecting that green line there to go vertical upwards, then our expectations are unrealistic.
Ok lets postpone that question and revisit it later. I am curios to see where we meet somewhere in the middle. I get your point. Storage nodes will add capacity. But you might also understand my point later down the road. Adding capacity requires time and we have no data yet on how fast the storage nodes can scale up. Sure you will now say it only takes days. That might be correct for your and my node but not average joe that will find this thread next month wondering why the payout is outside expectation.
I’ll try meeting halfway. Let’s say we are adding 20TB drives. How fast would you want to see these drives added? I’m not being aggressive in any way, I’m trying to balance investment (ie inventory) vs demand.
To put things into perspective: There are 141 SNOs that clear payment no matter what the L1 fees are (recent month payout report, couple of months back?). I’ll add a bit of spice on that and say that 41 SNOs were waiting for their undistributed payout from the previous month, coupled with their (then) current payout, so they cleared payment. That brings us to 100 big SNOs (the ones saying “we’ll keep adding capacity if it’s used”).
How fast would you want those 100 SNOs adding 20TB drives? Ordering 10x20TB drives and having them sit on the shelf for a month, or ordering 1 drive every month? That’s 100x20TB drives per month = 2000TB/month or 2PB/month.
2 Likes
all these TBs in trash due to saltlake cleaning should be released in 1-2 days, right?
Can add ~300TBs within 24 hours (weekends excluded) and additional 1.7PBs in the same timeframe if someone would pay for the drives. Otherwise at a rate of ~40TBs a month. It would be an insanity not to expand the money printing machine if there would be such opportunity.
4 Likes
Thanks to all the cleanup work you shouldn’t have to add any extra drives. For now I would say please don’t order new hard drives and only use what you already have connected and on standby. Next week all the data that was moved into the trash folder this week will be available for new uploads. I am trying to solve this question with math but there are to many variables. I guess we need to wait for the actual uploads to hit our nodes. At that point we should be able to calculate how big our nodes will grow. All we need for that is some data about the TTL distribution.
5 Likes
Didn’t they always say don’t buy hardware for storj? 
2 Likes
20 characters and more.
4 Likes
So I think you’re saying the slowest the network would expand would be 2PB/month if only the most-invested SNOs added one 20TB/month. And only if customers went above that rate would Storj need to consider what the other/average SNOs would do.
That seems like a reasonable way of looking at things: good estimate @Mitsos !
3 Likes
This however might approach those bandwidth limits if only 100 SNOs would expand. But hard to tell, as we don’t have the information about how bandwidth intensive the uploads will be, if it will be sustained or peaks only etc.
3 Likes
My limit is only available internet speed and I guess this applies to many SNOs.
5 Likes
That 2PB/month is just the big SNOs. There are still 4300 smaller SNOs that would love to add more. I’m not saying that someone running a 4TB drive would go out and buy a 20TB drive, but I can speculate that at least 1000 of those would get an 8TB drive to go with their 4TB one.
Just so we don’t have any misunderstandings by either side: that 2PB/month is raw expansion. It doesn’t take RS into account. It’s just a scale of how fast things can move if need be. At 2 drives per month (per big SNO), that point becomes mute anyway.
(I wonder if I can summon him?)
Hey @Th3Van ! If Storj cranks up their capacity-reservation ingress, and your setup starts to gain 100TB/month… would you continue to add 5 20TB HDDs every month to keep up? Or would you let your current disks fill and be content?
(I think I saw the last trash wave took over 200TB from his disks: perhaps he’s no longer a true believer? 
)