I'm thinking of implementing Yggdrasil as part of secured off-site backup solution and wanted to test the speed when rsyncing large files through the network. I was pleasantly surprised.
So, this was a pretty basic test where I was pulling a 5GB test file from one server to another. The first test was between two servers, one in Finland and one in the Netherlands. The test was ran over Yggdrasil and over IPv4 (The Finnish server doesn't have native IPv6 capability).
The second test was between two servers, one in Dallas, Texas, USA and the other one being the same Netherlands server as above. With native IPv6 in both locations, I was able to test that as well as Yggdrasil and IPv4.
Anyway, the results:
Yggdrasil - Netherlands pulling file from Finland
root@ygg-test:~# rsync --progress -avz -e "ssh -p 8305" root@[200:23af:xxxx:xxxx:xxxx:xx:xxxx:xxxx]:/root/yggspeed.test /root/yggspeed.test
receiving incremental file list
yggspeed.test
5,368,709,120 100% 126.39MB/s 0:00:40 (xfr#1, to-chk=0/1)
sent 43 bytes received 5,221,101 bytes 125,810.70 bytes/sec
total size is 5,368,709,120 speedup is 1,028.26
Above is pulling the 5GB test file over the Yggdrasil network with speeds of 126.39 MB/s. Not bad! Especially when compared to the clearnet/IPv4 speeds shown below:
IPv4 - Netherlands pulling file from Finland
root@ygg-test:~# rsync --progress -avz -e "ssh -p 8305" root@xx.xx.xx.xx:/root/yggspeed.test /root/yggspeed.test-v4
receiving incremental file list
yggspeed.test
5,368,709,120 100% 129.91MB/s 0:00:39 (xfr#1, to-chk=0/1)
sent 43 bytes received 5,221,101 bytes 120,026.30 bytes/sec
total size is 5,368,709,120 speedup is 1,028.26
So, as you can see, the clearnet speed was barely faster, coming in at 129.91MB/s, a difference of 3.52MB/s.
This isn't a very scientific test. Geographically, both these servers are relatively close to one another (Finland and Netherlands) so that may have something to do with the performance being similar.
Texas to Netherlands
Yggdrasil - Netherlands pulling file from Texas, USA
root@ygg-test:~# rsync --progress -avz -e "ssh -p 8305" root@[203:51cb:xxx:xxx:xxxx:xxxx:xxxx:xxxx]:/root/tx-to-nl.ygg.test /root/tx-to-nl.ygg.test
receiving incremental file list
tx-to-nl.ygg.test
5,368,709,120 100% 137.33MB/s 0:00:37 (xfr#1, to-chk=0/1)
sent 43 bytes received 5,221,106 bytes 128,917.26 bytes/sec
total size is 5,368,709,120 speedup is 1,028.26
At 137.33MB/s, we're faster than we were moving the same 5GB test file between the Netherlands and Finland. Interesting.
Let's try moving the same file over traditional IPv6, which is usually noticeably faster than IPv4 over longer routes:
IPv6 - Netherlands pulling file from Texas, USA
root@ygg-test:~# rsync --progress -avz -e "ssh -p 8305" root@[2604:fbc0:x:xx:xxx:xx:xxx:xxx]:/root/tx-to-nl.ygg.test /root/tx-to-nl.ygg.test-v6
receiving incremental file list
tx-to-nl.ygg.test
5,368,709,120 100% 200.56MB/s 0:00:25 (xfr#1, to-chk=0/1)
sent 43 bytes received 5,221,106 bytes 171,185.21 bytes/sec
total size is 5,368,709,120 speedup is 1,028.26
Wow, so that is a big difference in speed between Yggdrasil's v6 and the clearnet v6.
Let's see how clearnet v6 compares to traditional IPv4:
IPv4 - Netherlands pulling file from Texas, USA
root@ygg-test:~# rsync --progress -avz -e "ssh -p 8305" root@[xxx.xxx.xxx.xxx]:/root/tx-to-nl.ygg.test /root/tx-to-nl.ygg.test-v4
receiving incremental file list
tx-to-nl.ygg.test
5,368,709,120 100% 196.98MB/s 0:00:25 (xfr#1, to-chk=0/1)
sent 43 bytes received 5,221,106 bytes 165,750.76 bytes/sec
total size is 5,368,709,120 speedup is 1,028.26
Okay, so still a pretty big difference but Yggdrasil is still incredibly fast.
Pulling 5GB Test File from Finland to the Netherlands:
- IPv4: 129.91MB/s
- IPv6: (Unsupported on Finland server, so no test ran)
- Yggdrasil: 126.39MB/s
Pulling 5GB Test File from Dallas, Texas to the Netherlands:
- IPv4: 196.98MB/s
- IPv6: 200.56MB/s
- Yggdrasil: 137.33MB/s
BlueHat wrote
Good to now. I am currently researching about yggdrasil. It seems like great technology, and actually fits most of my use cases. I was getting tired of I2P's slow speeds, so I began digging deeper into alternative P2P networks that are less on anonymity and more on decentralization.