[POLL] Haiku User Location

Community
1

I need some information on our current user base’s geographic distribution. Please choose the option which is closest to your physical location.

Ideally you should only participate if you run Haiku daily or plan to in the near future :slight_smile:

Where are you located?
  • United States
  • Canada
  • Germany
  • Russia
  • Japan
  • China
  • France
  • Italy
  • United Kingdom
  • India
  • North America (other)
  • South America (other)
  • Europe (other)
  • Asia (other)
  • Africa (other)
  • Australia
  • Antarctica

0 voters

4 Likes
2

Looking at the results I’m guessing you are missing some European countries (like Belgium for me) in there :wink:

3

Or Sweden for two of us :slight_smile:

1 Like
4

There are a lot of countries in the world :slight_smile:

Overall, just looking for guidance on where “most of our users are” for repository mirroring purposes. :slight_smile:

At the moment, looking like one European mirror (Germany likely based on the split), one US mirror.

5

I suspect that a country like Spain would have more users than Antarctica.

4 Likes
6

Thanks for the data!

This was related to my ongoing IPFS experiments for distributing packages. Essentially:

  • First dedicated IPFS gateway node will be in Germany
  • Potential second dedicated IPFS gateway node will be in the US
  • IPFS updates will occur between 22:00-00:00 daily (10pm - 12am CST, 5:00am - 7:00am CEST)

This can all be potentially expanded in the future if it works out… also anyone can standup an IPFS gateway on their local network or publically and mirror (or host access) to our IPFS repositories via /ipns/hpkg.haiku-os.org

1 Like
7

Is there something like: IPFS for dummies?
Still don’t know how to join!

8

That’s the big downside… it’s complected :slight_smile:

Here’s my overview of the basics (super non-technical… general concepts not exact implementation):

  • All IPFS nodes are peer-to-peer and host data
  • Content and directory indexes are broken up into chunks and identified by their hash (called a CID)
  • “CID A” may be a hash of a directory with three files
    • {CIDA: “unixfs”, “/filea”: CIDB, “/fileb”: CIDG}
    • {CIDB: “file”, CIDC, CIDD, CIDE, CIDF, etc}
    • {CIDC: “data”, ‘datadatadata’}
    • {CIDD: “data”, ‘datadatadata’}
  • Nodes can “pin” and store CID chunks… or go out and serve them from the network if they don’t have the chunk of data when requested.
  • Every ipfs peer can be a gateway and get data over http
    • Navigate to /ipfs/CIDA and get a directory of files
    • Navigate to /ipfs/CIDB and get a file named CIDB
    • Navigate to /ipfs/CIDC and get a small chunk of random data

Bonus
/ipns is a way to reference CID’s via “static names”

  • /ipns/egfg34534teryret34t43t4t is a public key pointing to a single CID. You need the private key to update it.
  • /ipns/hpkg.haiku-os.org is a “dnslink” record. We have a dnslink record on hpkg.haiku-os.org in DNS:
dig _dnslink.hpkg.haiku-os.org TXT
_dnslink.hpkg.haiku-os.org. 3600 IN	TXT	"dnslink=/ipns/k51qzi5uqu5dgnsdtttsbgsu7rbpky52k7lwi55mk8eywf2o0ds749syqiq6g9"

I have a quickstart guide for mirroring our stuff here:
https://github.com/haiku/infrastructure/blob/master/docs/ipfs-pinning.md

Please keep in mind this is all an experiment. No guarantee it’ll work or we will keep it around.

1 Like
9

Ah, wow, thats nice to read. But hard to understand… I will just keep on reading and try to understand it…
Thanks anyway, I will get it somedays.