Three times as old as your daughter, did I get it right?
3 Likes
You are in - all you need to do now is buy the requisite quantity of beer,
3 Likes
Definitely, and for the general public using 1/2 of ones bandwidth on nodes is not even a good idea. It should be less than 1/4 of ones bandwidth for the general public if we want smooth operations both with the network and user satisfaction with their internet experiences.
I am also wondering this morning if something is wrong with my port forwarding attempt I started last night with PUTS and GETS happening but no quotes or earnings over night.
My figure was just giving a baseline for current circumstances and always meant to be just that.
Although as the network grows the load on each node should reduce and this maybe the argument to increase from 5 replicas to like 7 or even as high as 9 so that the load on each node is even less.
Damn if I could get that I would, but that is the way I am, been in comms since the start of my career with 110 baud devices and even at 100 Mbps has my 1st harddrive I interfaced to my computer transferred in less than a second. And 5 Gbps would be amazing to have. I could upload that 70-100TB I have in much faster time than the pathetic 40Mbps upload I have now (max before crazy cost business plans)
But should you get it, just remember that it also allows for you to upload your files to Autonomi fast as well
3 Likes
Yes it is very tempting considering that it is only $30 more per month but the issue I have is they bring it to the the house and then everything needs upgrading from that point on.
When the tech comes on Monday I am going to chat with him and tell him that I am most likely going to switch to 5 soon and can he please just lay the groundwork for it instead of me calling him out again, then they can just flip the switch when I decide.
The ONT is about 30m away from my office, I want him to either lay fiber to the office where the router will be (it can accept fiber directly) or a decent solid copper ethernet that will handle 5gigs. The ONT will need a upgrade too, so really I just want to see what they will do for my needs.
The only reason I would get it though is if it allowed more nodes from home, but with how these nodes are taxing my cpuâs right now 2gig is way more than my node running capability allows.
I wonder if people with wicked fast upload bandwdith will offer a service like AWS Snowball? Ship them a drive to be uploaded and the keys to a wallet with an amount of tokens that will cover the cost and compensation for the uploaderâs time and effort.
Then in reply the data owner gets a list of the addresses. Problem being that if the data owner doesnât want the files to be public the uploader will have the addresses and the keys to download them. Someone clever with cryptography might be able to see a way round that but I canât.
10Gbit switches are getting cheaper, I even bought one 5 port to be able to connect 2.5G gear via a 2.5/10G switch to my NAS and the old & new NAS together to transfer files a tad faster at 10Gb
When you want to utilise it for Autonomi then the SBCs might be the way to go, at 2.5Gb into a 10Gb switch allows full speed for each over the 10Gb switch into the 5Gb fibre. For me it would be 4 SBCs into the 2.5G/10G switch (4x2.5G + 2x10G copper) into the fibre router or into a 10Gb switch then to the fibre router
1 Like
You noticed some cpu spikes on node-manager right? I had 35 nodes run for hours utilizing ~90%, I hoped it would settle, it never did.
I just started 35 without node-manager on the same machine and cpu is way better.
Going to start another 35 with node-manager again on another too see if I can confirm,
1 Like
I found no difference and on a SBC (4 core) I was running 40 nodes started by node-manager and was not cpu bound, highish at 40% but not bound. I am guessing I could have had another 40 before reaching the excessive cpu usage of 80+%. But my upload speeds was the issues and others not liking the increased lag time of their internet experience
In theory the node-manager is simply running the nodes for you doing the cli and filling in the options for you. Then it has no effect on the nodes as its not running. It does RPC calls to do any work
1 Like
Yeah, I wonder why the difference then. 
Location of log files? Logging level? Perceptions
My 40 node run was running with much lower CPU for first hours after initial bout of CPU usage and then the next morning it was high at the around 40% usage and network usage was up too
I am getting my KVM switch today so I can fire up my second Odyssey and have them running side by side and compare things. Not on the same machine but very close in most aspects and by swapping the node-manager / native I should be able to get good comparisons in real time. So ISP conditions wonât be a factor
4 Likes
I made no adjustments to log level, but something is different somewhere, I had 3 different runs over the last 2 days trying only 35 each time which was never a issue for me. cpu was a massive pita.
This is the fourth try and and immediate difference. I much prefer using the manager so I would really like to figure out what is making the difference,
3 Likes
Well idk WTH. I started 35 each on 3 different machines now.
1st just plain old safenode with port forwarding.
2nd using safenode and --home-network
and a 3rd using node-manager with port forwarding.
Non have the cpu issues that plagued me for 2 days on 3 other tries.
I did nothing different so I can only guess it was network related.
4 Likes
Thanks, need to get used to hanging out there more
1 Like
Can you share the full command you used for 1,2 and 3 please Josh?
3 Likes
Maybe in the early network with few nodes, data moves around more or has to be kept in memory more? Idk just speculation.
3 Likes
It wasnât really early network though, all of this has happened over the last 3 days.
It was so bad with my first attempts that I was sure that I would wake up to glowing cpuâs this morning but no it is all good accross the board.
But yes the only thing that could have changed because nothing did on my end was network conditions.
2 Likes
My equipment upgrade is delayed till later on this week. Apparently, a lot of prep work required to go 5Gbps (both on the ISP and at home). I canât imagine 4700Mbps off continuous traffic per month would mean nearly 1.5 Petabytes off data in and out per month (ooof).
All for SAFE! All for SAFE! 
.
11 Likes
here is the robbery going on in ONtario Canada with Cogeco, the service that I have at home, pretty typical in fact for many living in smaller towns and cities.
4 Likes
This is typical pricing for the more rural areas that Iâve always lived in. Economies of scale just says itâs not worth it to bring us the good stuff yet. 
3 Likes
Thing is the fiber follows the railway lines. Easy to lay on already cleared paths.
Those lines go through little towns with repeaters roughly every 30 miles apart.
Its not unreasonable to provide decent internet in many rural areas where they just donât.
2 Likes