So you have Cgnat. Your good then.
so given you pay arround 150 DKK for each ip, does it still make sense money wise?
and do i get it correct that each node has about 20 TB ?
i do, but with the setup my isp offers it does not matter, you could not tell from the outside so only i know… and everyone else that reads this
Absolutely! - if you don’t counter in the expenses for the hardware and the power it uses ![]()
Total # of nodes on this host ----------- : 132
Number of hashtable(s) ------------------ : 0/1056
Number of memtable(s) ------------------- : 1056/1056
Memory (used/available) on this host (MB) : 200953/3030202
Total used space - reported by SN ------- : 612.17 TB
Total available space - reported by SN -- : 2637.72 TB
Total amount of trash - reported by SN -- : 18.58 TB
Avg. used space per SN (without trash) 612,17TB / 132 = 4,63 TB
Avg. available space per SN 2637,72 / 132 = 19,98 TB
Total free space (2637,72 - 612,17 - 18,58) = 2006,97 TB
Th3Van.dk
2PB free ?!?!
Jumping zombie Jesus! ![]()
My point was about complying with the rules about multiple nodes on the same hardware. If I rent a VM in a datacenter, I can’t tknow if they put another node on the same host and have no way to prevent it.
It matters when it comes to rules. The goal is one thing, but the rules are slightly different. You have to appear fair with their enforcement, because unfair rules could make the company look bad.
For example - let’s say a node gets disqualified, the operator writes to support or here and asks why. The answer - multiple nodes on the same hardware. The operator was renting a VM in a datacenter. So, basically, doing so, becomes a lottery - you can do everything right (that is in your power), but still get the node disqualified becasue someone else did something that he too thought was OK.
This is why some nodes get reinstated, if it turns out that they were disqualified because of a node software bug.
You really want to avoid the situation where someone does everything in his power right (people would naturally assume that running a node in a datacenter would be better than at home, as a datacenter may be more reliable).
If the ISP provides internal IPs to everyone and uses CGNAT, it probably is easier to do a 1:1 NAT to the customer instead of routing that IP there.
I still think that it’s better that the protocol itself has a deal with it, there are changes in this regard, like “choice of n” rule in the node selection, there are others (geofencing, Select, Regional, Global, etc.).
So, once there would be implemented something to make suggested list of nodes to the customer is truly uncorrelated, like with the placement Global Collaboration, where pieces are distributed around the world, not only to the nearest nodes as for Regional or even worse - for geofencing, then it wouldn’t be so matter, how much nodes you run on the same hardware, they all will be selected to follow the pattern to make them uncorrelated.
Right now for the Select nodes, for example, we use the fact that nodes are in different DCs with 24/7 NOC, redundant (or a quick available replacement) hardware, redundant and wide network connection, redundant power, a personnel, available to solve any issue in a short amount of time, so the node selection is less strict regarding IP (there is no /24 rule), however, it’s still makes them uncorrelated as much as possible (there are other rules, which are more restrictive, if would be applied to the Global Storj).
yes but i think that is for 132 nodes.
so to get as close to what is required to selects i would have to solve that one…
ie be able to start solving a problem within like 15 minutes?
hardware wise i would prop go with getting the same hardware twice and just have one set running and the other set as a spare that i can replace in like 10 minutes or so
power…. with the way denmark has their power grid build we simply only have 1 power company, there are many sellers of power but they all use the same lines, but its the same company that work on those, i guess its the use of smart meters etc that makes it possible to have more than one seller
its on the inside of my house i can make the difference.. ie making sure that some fault in the house does not interrupt the power to a node and the network gear from the internets entry point to the node itself. and then double up on power supply and the ups’s
same for internet connection, i could trick hte utility company in to installing a 2nd fiber but i would have no control as to what location that comes from and what trench in the ground it is… it could even be in the same tube in the ground….
i assume that a step between the regular nodes and the select ones are not ever going to happen, but it would be nice for those of us that want to go the extra mile or half
Why would you want that? Select is a totally different story:
Hi Storj Team and Community, I’m currently an active Storj Node Operator and have been running a standard node on the public network for a while now. I’m very interested in understanding what it takes to become a Storj Select operator. Specifically: I have 100–500 TB of storage capacity ready to deploy. And can get hold of more if needed. I have access to a dedicated server room with stable power, cooling, etc. I’m wondering: What are the exact technical and operational requirements for …
Why would you want that?
to MAYBE do this:
a step between the regular nodes and the select ones
and why that? select gets paid more… or so they say
so say a “premium node” would be paid somewhere between select nodes and normal ones
ie be able to start solving a problem within like 15 minutes?
And have everything before that, starting with owning of DC (so, no individuals here yet, sorry) and having at least a 1PB of available space.
and why that? select gets paid more… or so they say
Why do you think so? Rates are lower actually.
Actually the difficulty would start when you will need a mandatory certifications, not only for your equipment, but also for you, as a provider.
so to get as close to what is required to selects i would have to solve that one…
ie be able to start solving a problem within like 15 minutes?
You need certificates. That’s the biggest difference.
Electricity - you can get power lines from different substations and then also have a generator. The grid in Denmark is probably the same as the grid in Lithuania, in that there is one grid. However, gettign power from different substations would increase reliability, since a substation is more likely to fail than the whole city (or the whole grid). And in case of a big failure you can have a generator.
Internet - you can get a second line or use mobile (or wifi) internet as the backup.
Multiple power suppies and multiple UPSs are easy, no problem there.
The big problem is then getting the various authorities to give you a certificate - ISO 27011, SOC2 (I think this is only for the US) and a bunch of others.
This would be unreasonably difficult and expensive just to run Storj nodes.
In theory, I could say that my house meets the requirements for Tier 2 datacenter - redundant internet connections (two different ISPs), redundant power (city and generator), redundant ACs, redundant UPSs and servers with redundant power supplies.
However, nobody would give me a piece of paper that said it. And the piece of paper is whats needed here.
You need certificates.
i never figured what good they would do other than draining my pocket faster
from different substations
i live 400meters from the waterline and we dont have many substations out here. we have one 100 meters from our house and that one was replaced about 6 months ago. and i was curious as F and asked them how it was connected. turned out it was in fact connected to 3 lines on the high side so that reliability was ensured. the reason why it was replaced was that one of the inputs was fried. they hooked up a big as generator that could take all the load, replaced the substation in 2 days, and then everything was back to normal
so yeah i dont think i can order a 2nd input from another station, not unless i was stupid rich, and then maybe even then they would refuse to take the order.
hence my idea of making sure that faults inside my house would not cut the power
redundant power (city and generator),
aha… you gave me an idea… a generator big enough to run nodes and network stuff and maybe fridge an freezer…
Internet - you can get a second line or use mobile (or wifi) internet as the backup.
never thought of that… so maybe 5G as the backup. the router i plan to replace my current one with has that option.
The big problem is then getting the various authorities to give you a certificat
yes and again what good would they bring me ? i can only think that they would provide some sort of evidence that i can run the stuff reliable…?
However, nobody would give me a piece of paper that said it
hmm…. i could have an electrician certify that. i would need him anyway to install the extra gfci breaker and the generator input. so to ask him to certify the install would not be that much extra….. but that would prob not be enough
select gets paid more… or so they say
Where do they say that?
What is the advantage of a “premium node” from a customer perspective?
The Select network serves a specific purpose where a customer needs regulatory compliance otherwise they would not choose Storj for data storage:
Hello Storj Community, We are so excited about the growth of customer data that we have been experiencing on the network. It has been quite the journey. It is amazing to realize that two years ago, we were excited by a customer with 10s of TB. Last year we were excited to see customers with 100s of TB. This year we saw a growing number of single digit PB customers. We just signed our first double-digit PB storage customer and we have many in the pipeline. Many of which look promising for us to …
You might want to read the annoncement more carefully. The customer decided last minute that the public network doesn’t work for them. We even explained to them that storj select is running the same code so there is no difference. I can say in this instance we did our best to push them using the public network but their response was no way we need storj select or we have to bail out.
aha… you gave me an idea… a generator big enough to run nodes and network stuff and maybe fridge an freezer…
Or use a renewable power sources like solar or wind.
select gets paid more… or so they say
I thought Storj charged clients a bit more for Select… but offered the commercial SNOs a bit lower $/TBm as they had the scale to still make money at the lower rate (and could fill faster without the /24 restriction). See: “Because of the scale at which they operate they typically accept much lower payout rates because of the efficiency of their operations” Select is a premium tier of service… so it makes sense that Storj can earn greater margins from it. Though in this case the 10PB customer …
never thought of that… so maybe 5G as the backup.
Will not work. They usually behind CGNAT.
You need a physical line, fast and reliable, at least two, maybe more. With a short SLA.
but offered the commercial SNOs a bit lower $/TB
so what allows the select nodes to make money is they do not have the /24 restriction ?
that changes things a lot
Where do they say that?
not anywhere…. i just assumed it
Will not work. They usually behind CGNAT.
true, thou some have an option of not having CGNAT but you pay extra for that.
normally for commercial purpose’s
The scale. They may got more data, than even Th3Van, but they have many restrictions and must follow rules, everything is much stricter there. One mistake = disqualification. It’s much more efforts than handle even 100+ nodes like @Vadim and more stricter rules. You put your money to be in this league.
However, there are benefits, if everything is fine.
It’s pretty attractive for DC owners, otherwise their capacity just idle (and have costs).