How fast does your internet speed need to be compared to hash rate?

stoon

My internet is very slow now that my 3 rigs are running. Rig 1 is 103Mhs, rig 2 is 92Mhs, rig 3 57Mhs. All three are hard wired to a managed switch then to the modem/router. I tested before I mined and it was about 25 down and 6MBs up. Can someone enlighten me as to how much more internet speed I need to produce 400Mhs?

dlehenky

Your hash rate has very little to do with bandwidth usage. If you're on a pool, make sure you are using a stratum connection. If you're solo, which I doubt, the bandwidth you use is strictly 'geth' maintaining the blockchain, which is not much, once you are in sync (we're talking 10-20 Kbytes/sec, max).

work

You should need barely any bandwidth for pool mining. If you watch packets between your rig and the server, they make up a very small amount of data (human readable too, thanks to the use of json in stratum protocol).

thesmokingman

I noticed a slow down in my Internet connection as well and I have a 150MB Down/Up dedicated Fiber optic connection. I'm running a Cisco enterprise router and switches.

I have two ISP connections (Comcast DOCSIS 3.0) 50MB Down/ 15MB Up / (Verizon FiOS) 150MB Down/ 150MB Up and my Comcast ISP connection is faster than my FiOS connection only when I'm mining. I figured it was just so never said anything.

stoon

I changed the service provides Router/Modem to just a modem and my Time capsule for the router. I have the 3 rigs connected to the managed switch then to the router. Everything seems to be running a lot better now. I'm thinking the service providers Cisco Router/Modem combo that couldn't handle the traffic. It was weird because I could still get 70 down and 6 up but I would always get these problems, not internet connected or major delayed and lots of buffering.

bctopics

dlehenky said:

Your hash rate has very little to do with bandwidth usage. If you're on a pool, make sure you are using a stratum connection. If you're solo, which I doubt, the bandwidth you use is strictly 'geth' maintaining the blockchain, which is not much, once you are in sync (we're talking 10-20 Kbytes/sec, max).

Why is it beneficial to use a stratum connection when pool mining?

dlehenky

@bctopics It decreases the delay getting the work from the pool to the miner.

bctopics

dlehenky said:

@bctopics It decreases the delay getting the work from the pool to the miner.

So in turn does that do what? Increase hashrate? or?

work

Latency matters a bit; bandwidth not much at all.

nasa_de

I am new in mining,
1- My connection is 512 Kbps, is it enough for 2 rig? each rig have 4 R9 390 GPU.
2- How much Traffic i will need for that in solo and pool mining?

shutfu

@bctopics stratum sends the work to you instead of ethminer using the getwork mode constantly pinging the server for new work, which could result in the miner working on a job that has already been solved, resulting in stale/invalid/rejected shares. At least, this is my understanding of it.

GreaterNinja

So if latency is a key factor to receiving work and updating geth with requests, then wouldn't using stratum for solo mining on a lan be ideal as well to be more in synch?

If ethminer is setup to send requests for updates every 100-200ms, then that is already quite a bit of delay.
Thats to me is like playing a FPS games on dialup back in 1990s versus others with digital connections that ping 30-100ms. 200ms vs 30-100ms + actual packet latency to synchornize, share and propagate with other peer nodes.

1 day of mining on a 6 gpu rig yields the following results using ethminer build 110 by genoil. Keep in mind this is only 1 of my mining rigs.

[email protected]:~$ uptime
22:24:04 up 1 day, 1:14, 3 users, load average: 0.71, 0.48, 0.45
[email protected]:~$ ifconfig
eth0 Link encap:Ethernet HWaddr 888888888888
inet addr:192.168.1.101 Bcast:192.168.1.255 Mask:255.255.255.0
inet6 addr: 88888888888888/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:11231208 errors:0 dropped:0 overruns:0 frame:0
TX packets:16892797 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:1615950133 (1.6 GB) TX bytes:1996785098 (1.9 GB)

Now for a ethernet LAN, its not a lot of data.

My Geth node server has RX bytes:4757886550 (4.7 GB) TX bytes:9341870553 (9.3 GB) after 14.75 hours.

I'd say it becomes a lot of data eventually and depending on:
1. how many miners you have connected to your geth node.
2. how many peers you have connected to your node.