Status SWARM Orange Summit, 1 Aug 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Swarm APIs

Description

In this presentation from Day 2 of the #SwarmOrangeSummit, Swarm’s developer @acud gives a talk on Swarm APIs. Since last year Swarm has received an upgrade in the form of three new supported Bzz API schemes: list, hash and resource. In the presenation he walks the audience through the steps of how to run a Swarm node and how to use the new Bzz schemes.

A

So morning my name is al adhim. The latest addition to the swarm team started them a couple of months ago and I'm gonna talk about swarm api's today, so yeah, a little bit about myself, is come from in the last couple of years from more corporate web development background before that, I was doing all sorts of weird stuff like homeland security and embedded development and yeah. So let's talk about swarm API.

A

So basically in the last year, as you all know, so, the team has been working on the network rewrite and a lot of things changed. There's schemes that were added to our BCG protocol suit and yeah. So quite a lot changed. We introduced encryption implementation not so long ago.

A

No so yeah we introduced the encryption lately. So in the last few weeks this is counter mode encryption and it has. You could distinguish the results of encrypted content by double length, references which are different from our current double lengths in comparison to the existing references that you would get normally from Swan from unencrypted uploads, so yeah it has its own. The CLI on HTTP implementations so and thing you should know at the at the moment about our encryption is that it is non-deterministic non-id important as of such, because we generate random encryption keys.

A

Well, so if you upload the same content, basically it would get different hashes and the different references yeah. So is it supported through using busy, zero and or normal bzz slash encrypt when you're, when you upload or using the encrypts for like when you're uploading from a CLI yeah, so we have a year ago, so we had just busy busy zero, and this is immutable.

A

But these days we have three more basically schemes that are supported through the BTC suit, so we have list hash and resource, so I would yeah walk through the existing schemes and a bit about the new ones, so the BCC suit, basically, you can walk through, manifests and retrieve metadata about files and yeah. So this is the most basic scheme and you can also transparently get content which is behind manifests because manifests have a sort of like a default path behind them.

A

We will see that in a bit yeah busy zero is about working directly with content and manifests, but when you want to actually display them correctly in the browser, so you would need a a sort of a mime detection which is configured on manifests, and so we have busy list which is scheme which allows you to actually list content which is inside, manifests and Traverse manifests trees, and you could also retrieve metadata about files to it and BCC hash is meant to actually do ENS resolutions and retrieve hashes, which are pointed to by the NS records.

A

Bcc resources are mutable resource updates and they allow you to actually register content on register content which regularly updates on swarm issue resource updates to these resources.

A

We're still ironing out this feature so yeah, but I will show you a walkthrough through most of these, so quick one just so you would how many of you actually work with the businesses should know a bit about the api's. Please lift your hands. Okay, so I show so I. Guess, there's no objection. I.

B

A

A little bit so it would be a bit tricky, but okay, so.

A

A

For example, of usage of this easy to just, if you would do this.

A

So this is the swarm website, basically on on the that sits on swarm gateways and so on. This is what you get when you see the well get the website through bzz on the on your browser.

A

If you would like to see the manifest behind it, for example,.

A

Looks something like this, so common prefixes would be the directories and entries would be the actual entries inside the manifests. The files.

B

This is the manifest.

A

That's how a manifest looks like.

A

This is, of course, using Jason. You could actually render it also as HTML and, for example, BCC hash would show the actual, like the hash that is pointed to by the DNS record. So that would be this.

A

You could also get the hash of a certain file, which is pointed so this is.

A

What else did we have so you know so. Basically, these are the schemes that we support, oh and with vz0. You would actually also see on screen sorry.

A

B

So with busy 0 it would.

B

This is that you get the raw content behind a file or.

A

For example, also, this would actually point to the manifest, so you get the manifest directly so yeah. There are different use cases for each one of these schemes and I hope that we could in the afternoon. Maybe if you would like to try it out and to give feedback also and ask questions about how how to use each one of the schemes and I would be happy to assist. If somebody wants to do some hacking and to try out some examples, yeah so.

B

So back to our presentation, so we deprecated a couple of things: HTTP.

A

Puts are duplicated now and certain CLI flags are duplicated. We renamed busy R&I to BCC raw and immutable respectively.

A

And you could expect in the upcoming period to have fuse integration directly with the CLI, so you don't have to go through actually working with it through the JSON RPC, which would make things a bit easier and we're still ironing out some description. Cysts. We have an HTTP API, but it will be more coherent, I guess in the upcoming weeks, yeah. So integrating you should.

A

When you integrate with swarm, you should prefer HTTP API so sure for the basic IO that when you get and put content on swarm and but for sure that the RPC API is also available for whatever needs you have. Pss is right now supported only through through RPC and.

A

Swaps, we're in swindle will be available also through RPC, so yeah, and so to touch a few points about how you could actually run a swarm. Node so have a few few possibilities. One of them is to run a participating node on the main network chestnut that might require bootstrapping it with known addresses, which is all in the documentation.

A

Which was actually revamped and I will show you in a bit. You have also signal turn mode if you want just your local development just to test out stuff. So it's also fully documented on how to actually do all of this and the docker container. You can run a single container. Naturally, you can just build build the container and run it or you can run a fully fledged swarm with the Amish is repository which is on this address.

A

I will also send make sure that this presentation gets to you at the end of the day, but that you can actually get to these resources, but you are not built at terraform script that actually brings fully-fledged swarm and you can choose how to run it.

A

So that really helps if you want to stimulate your own swarm and there is lying somewhere light, PSS client for Android, but we need to get in touch with the guys from the etherion foundation brilliant to actually get to it, but it's it is somewhere and we have a branch of swarm which actually supports this so more info on this later on yeah.

A

So basically we are in the on the roadmap, so for POC 4 we are working on enriching our manifest with more data which is more developer facing, so you could actually utilize this much better and have more possibilities to make out more from what we have on Swan. Basically.

A

So this is very important and we're really working on this there's more things to come, like access, control and erasure coding, which will be as part of the POC for and yeah I, also as part of POC 4. So we aim to really standardize and make our suit of protocols much more solid and robust, so that people can really build and rely on these protocols and integrate them. They will be integrated also into mist and into ENS as part of the infrastructure.

A

So you could expect much more ease of use and higher level of integration yeah. So that's more or less its I'll be. If you have any comments on our documentation, which I actually wanted to show you yeah this one right here, so you can access the documentation on I, keep on losing the screen.

A

You can access a new documentation, draft.

A

This address so would be nice if you can have a look and also work with it when you actually try out the new release, candidate and I hope this will be happening in the next couple of days.

A

So yeah would be really happy to hear what you think about also documentation and how it actually how much coherent it is. With your experience with setting up a swarm note if you haven't run one before so, that will be really helpful and.

A

Regardless we are very much open to whatever you have to if you have any feedback about our APRs and how you would actually like to see them. If you find any discrepancies or points where you think we you could do better so yeah we're actually really open to hear that, so you can get in touch yeah and regardless we're working on a lot of things to actually make the experience much better in regard in any sense of usage.

A

So if it's through CLI HTTP, so there's a lot more to come there yeah, so that's more or less at Argus. If you have any questions.