Fluence Labs Community, 8 Aug 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Developer Community Call #16

Description

Upcoming online & in-person events and hackathons
DOSA team presenting the project built with Fluence
TrustGraph and Registry docs
Aqua metrics and log collection, VSCode advancements
AquaVm performance analyzer

Useful links

DOSA project https://github.com/iqbalbaharum/dosa_project
Hacking On Fluence https://fluencenetwork.notion.site/Hacking-On-Fluence-Primer-28a87754397048e1bec72e3bfc91fd9b
Community Calls wiki https://www.notion.so/fluencenetwork/August-4-2022-Community-call-16-AquaVM-performance-analyzer-b83e6ac649f245a8830690198b475957
Fluence Documentation https://doc.fluence.dev/docs/
Aqua Book https://doc.fluence.dev/aqua-book/
Discord https://fluence.chat

A

All right, all right, hey welcome everybody. Another edition of the fluence technical community call get a full day. Today we get ella, giving an overview of upcoming events and uh hackathons.

A

We've got full legs with stuff, including trust, graph and registry documentation, our new aqua metrics log collection and vs code advancements, and we have uh community members. uh I think, are you here we're looking for sorghum and that was a team okay, so we got that all right so uh and then, with evan who's, gonna give a deep dive with the aqua vm performance analyzer and I've had a preview of it. So uh if you're in uh in aqua micro vm definitely stick around all right, let's start with anna.

B

Right, so there are some of the upcoming events that we are participating and also hackathons. So, first of all, just to remind everyone. We have community calls it's every month now and the next one is september 1st, then october, 6 november 3rd, etc, etc is every first thursday of each month. We get together here in discord, we're not using zoom and recording is going to be in our youtube.

B

um Then the first fluence meetup in nigeria, built by one of our community members active community members purity. So it's happening um actually on friday, the information in our discord. um Then we probably is gonna one. One of our team members is probably planning to participate in the mata cartel conference in denver. So if you want to get to know more about, dows feel free to join that, then there is berlin week the massive blockchain week september, 12 to 18.

B

uh adapcon, the conference for um ethereum developers and developers who want to build with decentralized stack. We, our team, is planning to be there so very welcome, let's meet and um and chat, and the um well-known ith berlin september 16th to 18th so full week of events.

B

If berlin is going to be only offline, I think only in person, so you can apply now um token2049. That's a big crypto conference in singapore uh running through the september 28 to 29th. We are having m4 stand there, so you can see it right here. So there are. A lot of companies are planning to be there and us as well and they're. Also, um okay, I'll tell a bit later that field community is planning to have conference as well, and that one is all things open in raleigh um us october, 3rd october.

B

Sorry, um I keep cutting out. I think that my uh I don't know what to do. Actually with that.

B

I think that's because I assure you well.

C

B

So it's fine okay! Okay, maybe I should speak louder or something I think, because I'm recording the same time. It's um not writing, but anyway, that's the conference for all open source um uh about the tools about the um like technologies for the developers.

B

Then, lisbon week again, east lisbon, hackathon offline in person um the week of events around it um file coin event, the big file coin conference. You can see here some of the dates uh compute on data community, meetup and solana, breakpoint conference and a lot of other conferences, just um watch, twitter and they're going to be all the updates and I'll post this. um This pdf in our discord and in our community calls wiki and then some of the hackathons that running right.

B

Right now, you can still take part in web 3, infinity hackathon run by a local singapore community, the deadline, I think it's august, 17th or something and hackathon, is running until august 23rd, it's via death post um check out. You can build with fluence check out the rules there and submit your project. We have youtube. We have a workshop published in youtube already and they're going to be in person conference. As I mentioned uh just before, talking 2049 in singapore, uh the big falcon singapore summit also very very welcome.

B

I think it's going to be also probably online live stream and uh polygon build it. Hackathon. The total bounty pool is over 500 000 and deadline. The extended deadline deadline is september, 5, so very welcome. We also have published our workshop, and this is the website you can still take part.

B

These are more of the upcoming hackathons. We haven't decided if we are participating.

B

Sorry, um yeah! Okay! That's it from my side! Thank you. So much.

A

All right, very cool you want to hack. This is plenty of opportunity. uh Okay, so next we get the the dosa team from hack fest. Why don't you tell us a little bit about your project and uh congratulations by the way you did very well and uh yeah. It's the! Let us know what you did, how you did it and what are you gonna do.

A

D

Very much actually just to be honest. I it took us three weeks to understand, influence and build it right. I mean for today we're going to share. Let me share the video.

E

F

E

Now, when the rocket is ready, we can then please.

E

We also support multiplayer.

E

Now the second part we call it the dosa workshop. This is the place where users can attach a rocket ship onto the current energy. We call it a consumable anything.

E

We reconnected the drugstore workshop.

E

Now we can see the skin exchange, we can go back to the plate, can go back to the game and play again.

E

D

All right, that's the gist of what we are building for the hackathon, but I assume that it doesn't explain much about what's happening around it.

A

D

All right, let me share my screen here, all right so for.

E

D

Part, it will actually help you to prepare for the some slight information, how we actually work with fluence in terms of getting what you saw inside the video just now all right. So I'm sorry.

E

D

I actually this project has been built by me and my friend it's actually he also joe earlier at sip as well. So three hours is the one actually working on the project. So what we're actually trying to build is what we call uh decentralized infrastructure we immediately.

D

So what you can see now is that what you see just like, I see how we combine two nst metadata to one metadata right.

D

So basically, our background is that we actually came from uh what we call a game industry right. So we actually build a blockchain games, so they actually have a lot of limitations in building what we call a crypto native game. So, most of the time the implementation is actually more towards. um You need to combine a lot of what we call a web 2 and a website technology together to win a single game right.

D

So one one, the very high- and I mean the most important part- is actually that in game we we need a lot a lot of what we call high throughput like in term of what we need to streaming the game and we need to stream about the information and also in terms of uh changing a blockchain itself to become a database all right.

D

So this is actually the infrastructure that we built at the back, actually one of it that you can actually check out in our. uh I will share it afterwards. This is actually uh the the the codes that we publish for the hackathons, so what we built is actually we actually work with two blockchain for the project. One is a polygon blockchain.

D

Another one is a bigchaindb whereby is actually a is a blockchain database, but big jbp is mostly used a and 10 min as the base that polygon as it is so how the transaction be happening is actually that when you meet when we actually okay, I will explain about the detail, but on the word down below so as you can see that we're actually using also the graph in term of of indexing, the data of the smart contract and we are using the service on a what we call a cron job, all right, we're running on the background service of every 100 seconds.

D

To achieve it, one thing that we actually missed out on this. We actually plan to do a consensus. It means that we are reading it for different, different different resources, we're actually planning to use a moralist at the graph, but because of the limit of time we just stick with the in the graph. Everything managed to do the processes, part of the fluence right.

D

So once they read the information from the graph, then only the certain information actually going for to remain the the what we call to remain the the industrial agreement right to mean the nfp to the user wallet with a randomized stats, and it also be used also for a couple of things, so they actually in general, we have a couple of services that you can see here that we run it. We have the.

D

uh But to sum it up is that they have two main things like: basically, the dosamin service and dosage game service to service is the one that interacting a lot on uh these two blockchain, whereby the dosage is the one that interacting to the game right uh so, but we're having a problem when we try to implement the service directly into a phaser 3 game engine right, something uh I'm not sure something that is when we try to compile it. It cannot read the fluence js right so to to solve it out.

D

We actually built a circuit. I o right. So it means that this, okay, the so can I become the communication channel between the game services to the game itself. So and on top of.

E

That the game service also is.

D

Actually, inclusive with we actually tried on this hackathon, but we didn't manage to complete we. What we do is we're building an anti-cheat system, wherever we are sending the information directly from the game, a certain interval for every hit right for every account and the information also been uh oh yeah. Another thing is that uh the data are being stored also, but this is also it's a bit incomplete that we actually try to pull everything together to the ipfs for any of the services, the broadcaster to the ipfs.

D

So this is how uh that the information looks. Basically the oracle is the part that the service every time you do, the main you do the relay here. Then we actually need the nft to the user. So then, after that the information are being stopped inside the bigchaindb right. So for the service itself, we acted for the key for the pitching debate information.

D

So it means that the service will be the what we call the point of contact or the trust to actually update the picture db of a certain metadata all right so.

D

D

So the one that you see actually for what we built one uh one thing that we try to build is what we call a consumable nft. So most of it like the biggest problem in the game, industry is like when you're trying to compare multiple nft to a single nft. So most of the time, if you have two nft, you need to burn through nft and build and create another nft whereby it's actually uh it's not efficient in terms of the gas and also for the process. It's not seamless, so what we built.

D

So what we do is actually the service one of the services right by the dosamin service will be the one that actually responsible to do the check so yeah. What we try to focus to do is actually that, because we are actually interacting for two blockchain so to produce a seamless and decentralized oracle to deciding in terms of merging between the base and also for the slot and ft right. So what we call the slot?

D

The scheme that you saw in the video just like it's what we call the slots, so it means that anyone can create a skin right now. So the idea is that anyone can create the skin and the skin that you can send a request to throw through a smart contract, especially smart contract, and it can merge directly to what we call the base and ft whereby it's the base rocket that you saw just now and he's updating the metadata directly from that right, so um yeah. So that's! Basically, I guess.

D

C

Side, it means that.

D

Every time we do the merge all that we are embedding the base metadata in term of adding the what we call the nft, the nft information that the slot and the slottable nft information back to the base, nft right and so from there then. So it means that we generate what we call we we plan to do more, what we call a created economy and displacer using this, our cost removal, nft.

A

Okay, very cool, very cool anybody have any questions. uh It was a great great. It was fun to judge.

A

Nobody all right, I have a couple questions for you. It took you three weeks to figure out what influence was doing, uh what tricks you up and uh what made you present beer to uh to actually try and play with it.

D

I mean both three of us. They, we actually came from what we call a web 2 project developer. So it means that our background has always been building a micro services. Api res mp, all this right, so one thing is that structures is actually the the infrastructure is not.

D

We have been too through, like a couple of hackathons right. The services that have been provided, the things are not really to the things that we are looking at, because most of it is not scalable whereby actually at the first, we all didn't know about the fluence right. But when we attend your your your session, actually every I mean when we entered the session.

D

Also, we actually like still playing in much of the things, but one thing that triggers this is actually the the idea that we can actually build a more decentralized oracle system right, because if you, if you want to build up a communication between these two, what we call these two blockchain together right so most of the implementationality, is using what we call a centralized solution, deploying aws and so on right in the cloud, so it triggers that we want this to be like as our what we call a landing ground, a playground.

D

We don't really mind that we lost also at that point like we just want like, because I we are are attracted with the flaws technology by that point right, so we said okay, we need to try. So I think I didn't ask a lot of questions inside the group and the technical at that time.

D

I know some of it is actually a quite a stupid question at that point, but because, like we try to understand, what's going on that, this part like like it is, and- and that's what make us like, we need to sleep with it. Doesn't matter because, at that time we have one week before the submission we said we don't mind, we don't really care at that point like we just need to complete it. We just need to have this, because this is to be honest.

D

It's not trying to to put you out it's just that, like what you feel. This is basically the solution that can solve a lot of problem, so we need this doesn't matter for the hackathon. At that point, we just need it for our projects and ongoing things that we are doing so we just get through it, but luckily we need to complete it, but even though we cannot complete all of it, but we managed to complete at least enough for the presentation. I guess.

A

Very cool, very cool: well, again, congratulations and uh definitely stay in touch and uh we've got a few pull outs and I think uh folic is going to talk a little bit about trust, graph and registry documentation, which I believe will be helpful for your future endeavors anyway. Congratulations again well done! Thank you very much.

C

A

You're good to go.

C

Sure so let me shirt: do you see the infinity?

C

uh So what should I start with okay? So I want to show you that we implemented log collection for for our network. So now we can like uh better debug this stuff when people come to us with problems and soon you know sooner or later, I hope sooner will have a thing. So anyone who runs their own uh friends account will be able to send their logs to this system and be able to just use our infrastructure for debugging and for observability into their own networks if they want it or as well.

C

We provide like all this is open source. So all this infrastructure as code is open source. So you can just run your own log collection performance pretty easily and we will have a documentation over it, but right now it gives us the actual, like look into the system. What we're like uh what's happening, uh for example, I can see that here is a particle. It came at this time and it went for through all these nodes. Then it did something in that time and I can like debug further further further and yeah.

C

So that slope collection, that's pretty pretty awesome, yeah and we're constantly working on metrics that I usually show a lot so, but not much changes and soon will change a lot okay. So that's log collection, as bernard said: yeah, we'll talk a little bit about register and trust grab documentation.

C

um We have a huge effort in writing and rewriting editing and re-editing the readme and documentation for registry. That's our service for service discovery. That's our basic tool for service discovery, because now nice pictures that hopefully explain what happens.

C

If you find something here, confusing whether you find some mistakes or anything at all, you can always come to us to telegram discord or just create an issue here and we will discuss it and fix, but yeah.

C

I think those features they can give a good overview of what registry is it's like a thing that you can uh that allows you to store routing information for services. So if you like, deploy a bunch of services like 10 of them or some something, then you want uh the services to be discoverable by other people. You just share with them a single uh identification string and they can uh find their services in network registry yeah. So that's, basically a decentralized content.

C

Discovery like service discovery, computation, discovery, thing yeah, so that's pretty awesome, um but guys did there. Also huge effort was into the trust group documentation. It has similar structure.

C

This is our distributed public infrastructure thing that allows you to assign different weights to different services. Soon we will use it for resource management, so, like more trusted, uh services will get more resources, resource locations- stuff like that, so it all will be programmable and flexible and yeah it's already there. You can already start to to learn about it and soon we will integrate it into registry uh completely and it will be possible to discover services uh judging by their weights in transgraph, so yeah, okay.

C

Another thing that we have um another thing that we have uh is: where is it? Okay is we we have implemented? Let me make it a little bit bigger. We have implemented uh metrics uh that are available from within the aqua, so uh it's a service matrix.

C

So if you, for example, have some service, for example, here I have a built-in service called registry that we just saw documentation for and like has some functions, uh and these functions you you might want to know how often they called uh how many calls were there, which calls like failed, which calls took what time on average or uh hero show like average time call time and last five calls, and also we show the change in memory consumption by the service.

C

So you here is zero, because the function merge is stateless and it doesn't consume any memory, but if it was, there would be different numbers and you can track like allocation of memory in your service. That's very useful when you deploying your own services and you want to know how they perform like how which functions are most used like if you think about it as an api.

C

That's basically metrics for your api for its call time for performance, for memory usage and for popularity, should I say yeah for how how many requests are there so yeah and I hope at some point we will uh drop some dashboard for this metric, so it would be possible to like explore the network and see which services perform, how and see charts on that stuff, so yeah that would be. That would be awesome, but that will take time.

C

um What else would I was? I yeah and also we've been working hard on the base code, integration so now aqua give aqua. So we have a programming language for distributed algorithms and we have a vs code plugin for it, and it supports different things now, for example, it supports go into definition, so you can go control, click on the function name and it will go to the service where we'll show you like, the definition of that function. Same works for like services themselves and same works for uh data structures.

C

So if you're like point and click uh on this structure, you will see what fields fields it has and where is it defined? Also, a very important thing is that if you have some imports, you can now click on them and actually see their api before that. That would pretty hard. You have to go to github and read code there, but now you can just control click on imports and you will see like what.

C

What's the uh what kind of built-in functions through these protocols provide, for example, you can concatenate arrays, you can like uh connect to create services and stuff like that, and it's all documented all here so yeah, that's pretty a pretty big change in quality of life for developers, and I guess that's it.

A

Yeah, okay! Well, in that case, elon's gonna give us an overview of aquam performance analyzer. Take it away.

F

uh One of uh cancer concerns for our team in previous month was the performance and performance of aqua vm as one of key components in a perfect world.

F

One would just profile there was binary but surround currently profiling tools for watson are very limited, and instead we introduced tracing crate, common rust tracing trade from the tokyo project.

F

uh What allowed us to measure uh the execution time of different parts of the aqua vm tracing, of course tracing is not, is not good for estimating performance as a proper profile. However, for coarse grained parts it works very well and to run the aqua vm locally one needs some tool.

F

It wasn't possible to run it locally before so air trace utility was introduced.

F

It has two subcommands run to run air, scripting, aquarium and stats to process some statistics gathered. First of all, we need to build the aqua vm.

F

We need to rebuild it with marine just standard operation, however, make sure that you build it with real in release mode, because moving performance of debug mode is close to useless. Okay. Let's do it.

F

The air chase utility it needs to to ignore the pass this binary. However, if you run the utility from the project root, it will will use default value that is going just work.

F

The runs of command needs interpreter. You see the default path. Some flags that affect marine lawson run time, tracing params, it's just a standard string for tracing trade. It can be info debug trace.

F

It can be worn, for example, to complete completely and disable it it you'll see. Why do you might need it so, and it has two subcommands plain and anomaly with plain: you define input data for the aqua game, execution data.

F

Different ids, you define it explicitly: loading from different files, peer id, for example, initial script, timestamp, and so on.

F

uh You define it explicitly and with anomaly you define just single anomaly file and only files are collected by our.

F

Fluence node, or by influence node and just a single file stored for particle executed for more than half seconds so for playing the only required parameter is data.

F

Other parameters have same defaults and I have just some large data. Let's see how it works.

F

Oh I'm sorry, I need to provide the script. One can provide script as a as a file or by default. It just reads from standard input.

F

I provide thing placed the airscape possible. Let's run so you see the execution traced of default granularity and, for example,.

F

F

For example, uh loading of this data took five milliseconds.

F

uh Other pages are.

F

Not significant here because we had a very simple script, but for as an example, it would work and it printed the outcome of the execution.

F

So one can use also.

F

Json trace it. This is the same choice, but is output in json form format for future analysis with air trace or even with gq? Let's start.

F

Pay attention that you need to forward standard standard error so now comes to play. The air trace statistics.

F

uh The air trace the air starts, of course, has.

F

Two main features: it just pretty prints the same choice that is stored in json format, and it provides cumulative statistics on execution of different spends.

F

If you provide provide no arguments like here, it will just do both if you provide any or all it will do what you requested one another important.

F

Execution option is native, it runs aqua, vm compiled in native mode, and it is useful for two things. First, you can run it with real profiling. While data will be different, it will may provide some hints and also you can debug.

F

You can run aqua vm with debugger and see what is happening inside and what is not happening.

F

F

And that's basically, all.

F

Right and you can repeat, because execution aquarium usually executes very quickly together, proper profile and statistics, you may repeat it many times.

F

F

Even 100 repetition takes less than one second.

F

Okay, six seconds enough together some statistics, some profile info. Thank you.

A

Thank you. Anybody have any questions.

C

Come on you say, the the statistics was cumulative. You mean like that. It's the sum sum of every.

F

More detailed tracing levels, uh there are some operations that are uh executed repeatedly, and this wrapper just sounds same spans.

F

A

A

All right- and I have some questions uh first, how does a developer, who wants to get more information, actually start using aqua vm trace? What do they have to do? How do they set it up? How do they prepare the data that needs to be processed like you had in that file? If you could maybe speak a little bit to that.

F

First data is a key conception of the aqua vm and just one just need to understand it, to get it somewhere or to be able to to write manually that I did it's a json, big json or small json file of particular form format and as of installation.

F

F

Is stored in two city lights, air trays, the and it's just.

F

Ordinary rust binary so.

F

Common installation, I'm sorry.

F

Common rust installation procedure will do.

F

Or one can even run it with um cargo run within uh micro, vm project.

A

Well, thanks everybody for tuning in once, more congratulations for the dosa team. Stay in touch did a great job, looking forward to see what you can do with it and push the envelope and we'll see you in four weeks. Good bye,.