The Ebb and Flow of Traffic

Node Operators
1

I understand that network traffic is driven by real human customers. However, traffic sometimes falls off a cliff into oblivion:

2019-11-11T00:54:53.254Z	orders.118UW...	{"count": 582}
2019-11-11T01:54:53.252Z	orders.118UW...	{"count": 511}
2019-11-11T02:54:53.251Z	orders.118UW...	{"count": 627}
2019-11-11T03:54:53.259Z	orders.118UW...	{"count": 601}
2019-11-11T04:54:53.255Z	orders.118UW...	{"count": 573}
2019-11-11T05:54:53.253Z	orders.118UW...	{"count": 518}
2019-11-11T06:54:53.255Z	orders.118UW...	{"count": 527}
2019-11-11T07:54:53.255Z	orders.118UW...	{"count": 501}
2019-11-11T08:54:53.252Z	orders.118UW...	{"count": 521}
2019-11-11T09:54:53.255Z	orders.118UW...	{"count": 538}
2019-11-11T10:54:53.255Z	orders.118UW...	{"count": 573}
2019-11-11T11:54:53.260Z	orders.118UW...	{"count": 591}
2019-11-11T12:54:53.257Z	orders.118UW...	{"count": 571}
2019-11-11T13:54:53.257Z	orders.118UW...	{"count": 593}
2019-11-11T14:54:53.256Z	orders.118UW...	{"count": 548}
2019-11-11T15:54:53.256Z	orders.118UW...	{"count": 643}
2019-11-11T16:54:53.257Z	orders.118UW...	{"count": 552}
2019-11-11T17:54:53.252Z	orders.118UW...	{"count": 351}
2019-11-11T18:54:53.247Z	orders.118UW...	{"count": 46}
2019-11-11T19:54:53.247Z	orders.118UW...	{"count": 1}
2019-11-11T20:54:53.244Z	orders.118UW...	{"count": 3}
2019-11-11T21:54:53.245Z	orders.118UW...	{"count": 1}
2019-11-11T22:54:53.245Z	orders.118UW...	{"count": 1}
2019-11-11T23:54:53.243Z	orders.118UW...	{"count": 1}
2019-11-12T00:54:53.245Z	orders.118UW...	{"count": 2}
2019-11-12T02:54:53.244Z	orders.118UW...	{"count": 2}
2019-11-12T03:54:53.243Z	orders.118UW...	{"count": 1}
2019-11-12T04:54:53.243Z	orders.118UW...	{"count": 2}
2019-11-12T06:54:53.243Z	orders.118UW...	{"count": 1}
2019-11-12T07:54:53.245Z	orders.118UW...	{"count": 1}
2019-11-12T09:54:53.244Z	orders.118UW...	{"count": 1}
2019-11-12T11:54:53.242Z	orders.118UW...	{"count": 1}
2

That’s because one of those human “customers” sends test traffic mostly from that satellite and actual customers are mostly on other satellites. There will be fluctuations, this isn’t strange behavior.

3

Great obversation. I don’t see any issues here so whats the point in creating a topic for normal behavior?

4

It’s something to write about I suppose…

There’s also something to be said about what normal network behavior looks like. At the moment it’s difficult to determine if any particular behavior is normal or abnormal, due to testing, or based on a particular SN attribute.

So, I’m simply noting this normal behavior for my particular node… Place the post in the “possible expected normal behaviors” category.

1 Like
5

Because, we, as Storage Node Operators, don’t know what “normal” behavior looks like all the time. Sometimes we are saturated with tons of errors; at other times, not. Sometimes we receive heavy traffic; sometimes, not. We are always looking for clues that we aren’t having a malfunction on our end. It’s good to share observations.

6

i for example am gratefull that beast posted that, coz every thime something suspicious happens i come here first to check if im the only one, and should i panic, and how hard i should panic eventually;> BTW, please turn the heat on, this month will look extremally shallow, after that sugre went off, it would be the first month for me im not sure is it storj paying me, or i will pay storj, because electricity cost may slighly overpower the node income (in my case, i dont run pc 24h for my purposes, i run extra hours every day for storj, so it can be 24/7/365, so no pretensions, but still a sad vision for this month) Like 1 TB egress per month would be satisfying for me, and u dont have to srge anything.

2 Likes
7

Please do not expect to have electricity costs covered by income from running a Storj node. You should be using existing equipment that is already online and running anyway, not hardware only dedicated to running a node on our network. We cannot guarantee you to always have a traffic coming to your node. This is not mining.

8

That’s the whole point. Thanks @Ruskiem.
I think I’ll look at it for another month, then I’ll look for another alternative. I thought the concept was great and would support it, only at the moment is it going to pay for the network. It would be a pity. To strengthen myself in press conferences and compete against Amazon, I find it a bit over the top. :wink:

9

I’m glad that 1 TB agrees good for you me on the other hand I had 1.97 PB hours last month and my income less than 20.00 us and my logs show more egress traffic than ingress

10

I see today 100mb download and 0b upload today!!! :stuck_out_tongue_winking_eye:

11

That was thru for V2 where the unused disk space could be shared when your CPU was up. In V3 when you need a 97%+ up-time and can’t shrink the Storj disk space usage you need dedicated servers even a NAS is switched off by any earth preservers when not one use it.
We the stations can build any configuration you could need from a PI on ADSL to a Petabyte SSD system but at least it has to cover the running cost.
The PI node electric cost is between $.50 and $5/month depending were you live, I guess you won’t get too many stations in Hawaii.
If Storj doesn’t try to match the customer demand to the node supply, or least allows transparent data, like we had with the third party site, we are going to see a lot of nodes coming-up and going-down, giving-up on an unprofitable system. Not very good the system stability.
At least they should put some throughput characteristics online so I don’t have to rush to my node and figure out what’s wrong.

2 Likes