stebalien changed the topic of #ipfs to: Heads Up: To talk, you need to register your nick! Announcements: go-ipfs 0.4.23 and js-ipfs 0.41 are out! Get them from dist.ipfs.io and npm respectively! | Also: #libp2p #ipfs-cluster #filecoin #ipfs-dev | IPFS, the InterPlanetary FileSystem: https://github.com/ipfs/ipfs | Logs: https://view.matrix.org/room/!yhqiEdqNjyPbxtUjzm:matrix.org/ | Forums: https://discuss.ipfs.io | Code of
mowcat has quit [Remote host closed the connection]
cheet has quit [Quit: ZNC 1.8.x-nightly-20190530-13450c7f - https://znc.in]
cheet has joined #ipfs
scde has joined #ipfs
ZaZ has joined #ipfs
mauz555 has quit []
ylp has quit [Quit: Leaving.]
xcm has quit [Remote host closed the connection]
xcm has joined #ipfs
aditsachde has quit [Ping timeout: 245 seconds]
androbot[m] has quit [Ping timeout: 245 seconds]
muyloco[m] has quit [Ping timeout: 245 seconds]
aditsachde has joined #ipfs
androbot[m] has joined #ipfs
muyloco[m] has joined #ipfs
KjetilVaagenDisc has quit [Ping timeout: 245 seconds]
KjetilVaagenDisc has joined #ipfs
bengates has quit [Remote host closed the connection]
Taoki has joined #ipfs
foxroot30[m] has quit [Ping timeout: 245 seconds]
foxroot30[m] has joined #ipfs
cr4z0r has quit [Ping timeout: 245 seconds]
cr4z0r has joined #ipfs
aaron[m] has quit [Ping timeout: 245 seconds]
aaron[m] has joined #ipfs
scde has quit [Ping timeout: 246 seconds]
mikkel has joined #ipfs
Rettroix[Elliot] has joined #ipfs
RingtailedFox has joined #ipfs
jonnycrunch has quit [Ping timeout: 246 seconds]
<Alch_Emi>
Does anyone know if IPFS has any system in place to prefer connections to peers that are closer nearby, either physically or by latency?
<Alch_Emi>
Like, if I'm in NYC, will my peer prefer to open a connection to another peer in NYC over one in say Germany?
<swedneck>
physically nearby wouldn't make much sense i think
<swedneck>
i feel like it prioritizes nodes that are close on the network though, just can't remember anything specific
<Alch_Emi>
But isn't network closesness bascially arbitrary based on your randomly generated pubkey?
<swedneck>
wym
<ripply[m]>
I think that is how the dht kbucket works
<Alch_Emi>
I'm thinking mostly for the interplanetary case, like, making sure that a peer on mars doesn't make half a dozen DHT connections to Earth before being able to get the content
kanejDiscord[m]1 has joined #ipfs
<Alch_Emi>
ANd yeah
<Alch_Emi>
Unless there's another definition of distance
<Alch_Emi>
but the only time I've seen network distance discussed is xor distance in determining how similar two keys are and how many DHT connections are left to make
<swedneck>
afaik distance is counted in hops
<swedneck>
from one device to another on the same LAN it's 1 hop i think, just from the origin to the destination (i don't think switches count)
fleeky has quit [Read error: Connection reset by peer]
<swedneck>
at least if you're talking about IP, which in the end is what everything uses
<Alch_Emi>
Oh cool! Do you know what stage of the protocol that's taken into consideration?
<Alch_Emi>
And also what section? Like, does it apply to the DHT process too, or just picking a peer from the already known swarm?
<swedneck>
as i said i just have a faint recollection that IPFS takes some sort of distance/latency into consideration
<Alch_Emi>
a huh
<swedneck>
i know bitswap (the protocol used to exchange data) is at least supposed to prioritize nodes that give it requested data
vquicksilver has quit [Ping timeout: 268 seconds]
<swedneck>
i.e. if you provide a lot of data to others, you get prioritized
<Alch_Emi>
Makes sense. So like if you had a better connection to one peer you'd ask them for more stuff?
<swedneck>
not sure i understand what you mean
<Alch_Emi>
i'm just trying to confirm that I understood you right. Like, do you mean that that would solve the problem of latency by prioritizing nearby (and therefor able to send more data) peers?
lufi has quit [Changing host]
lufi has joined #ipfs
<swedneck>
no, i'm saying that it just in some way prioritizes nodes that have provided requested data in the past
<Alch_Emi>
oh okay
<swedneck>
idk whether the priority is in terms of bandwidth or who it sends data to
<swedneck>
or something else, or a combination
<swedneck>
hopefully a dev can hop in and answer better
skye0 has left #ipfs [#ipfs]
vquicksilver has joined #ipfs
fleeky has joined #ipfs
fleeky has quit [Read error: Connection reset by peer]
lthibault3 has joined #ipfs
fleeky has joined #ipfs
lthibault3 has quit [Ping timeout: 246 seconds]
fleeky has quit [Ping timeout: 250 seconds]
yorucontactDisco has joined #ipfs
ZaZ has quit [Read error: Connection reset by peer]
XeoX538 has joined #ipfs
gimzmoe has quit [Ping timeout: 250 seconds]
fleeky has joined #ipfs
gimzmoe has joined #ipfs
azuline has joined #ipfs
Arwalk has quit [Read error: Connection reset by peer]
Arwalk has joined #ipfs
BenLubar has joined #ipfs
xcm has quit [Killed (livingstone.freenode.net (Nickname regained by services))]
<BenLubar>
is there a way to add a link to an IPNS path inside a folder?
<BenLubar>
that is, can I have /ipfs/[some hash]/foo point to /ipns/[some other hash]
xcm has joined #ipfs
<Alch_Emi>
Like, an HTML link?
<Alch_Emi>
or linking the actual nodes of the tree?
<BenLubar>
like a symlink
<BenLubar>
the latter
scde has joined #ipfs
<Alch_Emi>
a that is out of my (shallow) knowledge level
<swedneck>
afaik no
<swedneck>
to my knowledge ipfs doesn't handle any kind of symlinking
rendar has quit [Quit: echi = moron]
<ShokuninDiscord4>
I'd guess no as well. Maybe <span class="d-mention d-user">bonedaddy</span> would know
scde has quit [Ping timeout: 246 seconds]
DiasDiscord[m] has joined #ipfs
mauz555 has joined #ipfs
<swedneck>
huh, adding a single large file seems to be a lot faster than adding the same size of multiple small files
<swedneck>
it's taking me 6 minutes to add a 35 gb file
jzelinskie has joined #ipfs
<bonedaddyDiscor4>
Emi: IPFS doesn't do proximity based routing. When people say "closest peer" that means the closest peer in the kad dht routing table, whether or not thats close to your proximity no one knows.
<bonedaddyDiscor4>
BenLubar: Yes you can, you would simply add the IPNS hash as a link within the unixfs dag structure
<bonedaddyDiscor4>
likely will require some low-level usage of `go-unixfs` i dont think theres anything at present which would allow you to do that
probono has quit [Read error: Connection reset by peer]
probono has joined #ipfs
<bonedaddyDiscor4>
swedneck: i would reckon thats because to add a single file requires less seeking on your drive than adding multiple small files
<Alch_Emi>
<bonedaddyDiscor4 "Emi: IPFS doesn't do proximity b"> Is there any plan for something like this? It seems like it might be kinda important for making IPFS truely interplanetary, and even just increasing practicality in terms of content delivery speeds. Obviously it's still great as far as a tool for creating a censorship resistant, redundant, and distributed web, but it was kinda cool to me that it could be used in that way
<Alch_Emi>
<bonedaddyDiscor4 "BenLubar: Yes you can, you would"> Wait, does that mean that could create a recursive structure in the D"A"G
<swedneck>
<bonedaddyDiscor4 "swedneck: i would reckon thats b"> makes sense
<bonedaddyDiscor4>
There might be some kind of spec on an ipfs github repo somewhere, but by and large I dont think that's anywhere on the roadmap for the next 1-2 years
<Alch_Emi>
aw bummer
<bonedaddyDiscor4>
THere's a lot of issues to sort out with IPFS before proximity based routing would even make a noticeable difference
<swedneck>
bonedaddy (Discord): think i could just init a new node and copy over my blockstore?
<swedneck>
or would that break things
t4skforce has joined #ipfs
stoopkid has quit [Quit: Connection closed for inactivity]
BenLubar has left #ipfs [#ipfs]
<swedneck>
hmm nope, didn't fix it
<swedneck>
also didn't break, so that's good
KempfCreative has quit [Ping timeout: 256 seconds]
<bonedaddyDiscor4>
swedneckL hm you might, its possible though you're hitting a bug with MFS
<swedneck>
any way to just reset the mfs?
scde has joined #ipfs
D_ has quit [Quit: No Ping reply in 180 seconds.]
fleeky has quit [Read error: Connection reset by peer]
D_ has joined #ipfs
verin0x3 has joined #ipfs
lthibault3 has quit [Ping timeout: 250 seconds]
verin0x has quit [Ping timeout: 250 seconds]
howaboutyes has joined #ipfs
fleeky has joined #ipfs
lthibault3 has joined #ipfs
howaboutyes has quit [Quit: Leaving]
ygrek has quit [Ping timeout: 246 seconds]
XeoX538 has joined #ipfs
scde has quit [Ping timeout: 246 seconds]
_jrjsmrtn has quit [Ping timeout: 256 seconds]
XeoX538 has quit [Quit: XeoX538]
__jrjsmrtn__ has joined #ipfs
lthibault3 has quit [Ping timeout: 246 seconds]
lthibault3 has joined #ipfs
gimzmoe has quit [Ping timeout: 246 seconds]
gimzmoe has joined #ipfs
fleeky has quit [Read error: Connection reset by peer]
plexigras has quit [Ping timeout: 256 seconds]
fleeky has joined #ipfs
<postables[m]1>
not sure
<postables[m]1>
i think the only way would be to wipe out your ipfs node and start from scratch, or just not use MFS> If you're using MFS to save on space, you can import your files to go-ipfs and then remove the original copy
<postables[m]1>
so you dont take up a lot of space
<swedneck>
i'm just using mfs to organize things and keep track of what i have on my node