Pre-Dev-Update Thread! Yay! :D

Actually I mentioned that in another thread. I am currently looking into that. Not docker though, but you can install third-party programs on the DiskStation which is based on Linux Kernel. If that´s interesting for you and you have some understanding of how this could work out, feel free to PM me.

3 Likes

With messaging perhaps.

Probably a bit optimistic, but loved the post! :slight_smile:

1 Like

I’d triple that estimate… Christmas 2017, if some more funds are found.

3 Likes

Hey! This is interesting…

EDIT: I wasn’t repo watching, really; I just stumbled upon this. :wink:

2 Likes

Tell me this is not a heavy-handed way of ensuring we only run one vault per box/ADSL connection?

I’m sure there is a good reason for it…

1 Like

I haven’t looked into it enough to say just how limiting it is, whether “on the same LAN” simply means “on the same subnet”. They’re not the same thing. I found in my recent learning about Docker that the container I was running was on a separate subnet from my physical machines, although it could ping my main router.

Lots of tests at the moment. We know there are issues with running very very low resource requirement (in terms of cpu etc) vaults. So we have created massively low resource requirement vaults now that allow folk to run many 10’s or more (thousands even) per connection. Sorta cuts our own throats right now as an attacker could easily run thousands of nodes and kill them all easily, but that is not the problem. The area we are working on is allowing the network to find nodes that are simply too low in capability and disconnect from them. That’s not simple, as load is dynamic.

So to start we want to reduce the many nodes per small connection, at a cost where folks have huge connections and can run many vaults (including us).

So this upcoming test (probably not today) will be important for us to measure and find that correct way the network should respond to these under-resources nodes.

It’s a bit more complicated than this, not a huge problem, but it does require a very thoughtful mechanism to deal with it.

18 Likes

Makes perfect sense: the instability near the beginning of the test was probably due to someone running an awful lot of nodes that went down all at once.

For small test nets running the latest build, we’ll just have to find a slightly round-about way to have a quorum of nodes. :slight_smile:

6 Likes

Also caching was disabled I believe

2 Likes

I’m a slow node as well sometimes, even with a good connection. Some folks (like me :nerd:) like to use Usenet now and then. But that thing doesn’t really care about the rest of your connection. It’s just brutally downloading as much as it can. That must be a pain for every p2p-dev.

1 Like

Routing arrived at version 0.23.0

Crust arrived at version 0.16.0

7 Likes

There’s work done on launcher v0.5

Update API changes for v0.5 #165

Talk about the v0.5 RFC so let’s see what that’s about :wink:

Here’s the link to the RFC.

Motivation
New features to the existing API can improve handling large data volume efficiently using streaming APIs. This can have a significant improvement in the performance of the launcher and the safe-core ffi for handling large data volume. Incorporating the standards in the API will also improve the stability of the APIs and make it easier for third party applications to integrate.

12 Likes

That’s fantastic news! So close!

3 Likes

commits over commits on github :slight_smile: :hugging:

oooooh i’m so glad i’m back now (was on holidays last week - really was a tough one to get infos about the update because my mobile internet didn’t want to work there and i needed to get some wifi without speaking the local language Oo …) good to be back and more in the loop again :smiley:

ps: do you guys never sleep?!?! still new commits on github Oo

8 Likes

Ah’m all agog, so ah um

3 Likes

Routing arrived at version 0.23.1

6 Likes

A user that posted multiple times in this thread has been deleted, his posts and the replies from others are gone as well.

I saw people responding on this user and calling the member a troll, please don’t do that. We’ve got the flag feature for posts that require the attention of the moderators. The only result calling a misbehaving member out is that it derails the thread and thereby giving a troll what he wants, please use the flag :flag_black:

16 Likes

Nice to see many commitments and arrivals, but actually we have no idea what they mean and how far that is to finish.

1 Like