Satellite vetting and traffic comparison

Hey guys I’m currently vetted on 2/5 sats and noticed something interesting.

I have one node that is 0/5 vetted and another that is 2/5 vetted which will be 4/5 later today or tomorrow. Saltlake was my first sat to vet and has absolutely not seen any difference in traffic when compared to the other node on my device (~1.5GB each node yesterday) whereas when my Asia East node vetted I noticed a 7x increase in traffic when comparing the unvetted to vetted nodes. Does that simply mean there is a 7x higher ratio of vetted to unvetted nodes on saltlake that the traffic distribution is diluted? I am eager to see if my saltlake traffic can get a 7x boost as it was my biggest traffic earner.

1 Like

Interested in insights too.

1 Like

When a SNO joins the network, they join all 5 SATs. What determines vetting is a bit more to how the particular SAT chooses to audit any data you have, which is somewhat random. If one SAT is vetting faster than another, it means your node keeps getting picked for audit on a chunk. Not a bad thing, just normal network operation.

I understand the point you made. The question I’m trying to ask is… Given that 95% of uploads go to vetted nodes and only 5% goes to the collective unvetted nodes of that satellite… Is it expected that a node vetted on a given satellite has no change in traffic?

It depends on the number of nodes being vetted on that satellite. For example if a satellite has 5000 nodes (/24 ipv4 subnets), a single node would get 95%/5000=0.019% of all traffic. If that satellite has only 100 unvetted nodes, a single unvetted node would get 5%/100=0.05% of all traffic. So in this example the unvetted nodes would even get more ingress traffic than a regular node.

The speed difference between vetting on different satellites is also explained by the difference in ingress amount. The more ingress a satellite has, the more data a node gets and therefore has a higher chance of getting audits earlier than on other satellites.
However, the amount of data on a satellite’s network plays an even bigger role. If one satellite holds 10x the amount of data compared to another satellite, vetting on that satellite will take a lot longer beause the chances of a pieces being audited is smaller as the node holds a smaller percentage of the total network storage.

3 Likes

also, lastly, you should take into account that saltlakes bandwidth has taken a deep drop this month…Take a look at the graph below from my longest running node:


Peak (1): was on Tuesday 10Nov and then is sharply fell to the bottom (2) on Friday 13Nov. It’s been super low for everyone. Also, just FYI the majority of that traffic in the graph is actually egress… here’s the ingress only graph:

Also, keep in mind that this is a test satellite, so the ups and downs of bandwidth is all primarily at the whim of Storj for what ever type of testing they may be working on. Seems to me that they’re just in a lull period for the time being.

1 Like

How do you know which are test satellites and which are production?

Because the “production” ones are listed for tardigrade:
https://tardigrade.io/login/

2 Likes

The biggest pointer I could give to any new (or relatively new) SNO, you just have to be patient. I haven’t been around nearly as long as some of these people on this forum, but having been a SNO for 16 months now, and patience is the name of the game.

I know I’ve used it before here in the forum, but I’ll say it again; “if you build it, they will come.”

My optimism and confidence in Storj’s long term future grows on almost a daily basis. If you keep an eye on their official blog on the website (Storj Blog) they continue to announce partnership after partnership, which only fortifies my confidence in their capability to grow this business, to provide real world value. Which in turn, although perhaps not as quickly as some might hope for, will increase the value of their business/product and grow their paying customer base, and in turn the value of the Storj token will grow without a doubt.

Most recently, their partnerships with Carnegie Mellon University, University of Maryland, Docker Registry, and not as recently with Github, to serve as backups and data repositories simply blows my mind. I can only image the huge amounts of data that they will soon be or already are uploading to the Storj network. I also, purely based on intuition, can only assume that other large research schools/organizations are going to be considering their own data centers or corporate contracts in search of a more affordable, reliable, and user friendly solution in the coming year(s).

So once again, I’d recommend that anyone who is less than 6 months as a SNO, don’t get discouraged and just hang in there, it will be worth it in the long run.

Also, for future reference, there’s a pretty solid thread already here on the forum, where a good handful of us post, more periodically these days, our daily/monthly stats for comparison and there have been some good discussions around trends/observations/theories. See below:

4 Likes