Filecoin Filecoin Virtual Meetups, 10 Aug 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Filecoin Virtual Community Meetup - August 2021

Description

We’re excited to gather the Filecoin community for another monthly meetup! Our community meetups are an opportunity to meet the people behind the many tools and projects being built in the Filecoin ecosystem. We have a wonderful lineup of speakers in store:
00:00 - Intro- Stefaan Vervaet (Protocol Labs)
00:12 - Storage Provider Branding - Stefaan Vervaet (Protocol Labs)
05:50 - HyperDrive Best Practices / 1475 Sight: Ecological Development after Hyperdrive Upgrade- Tim Zhu (1475)
15:18 - HyperDrive for Ecosystem Development - Charles Cao (FilSwan Team)
24:53 - Data Analysis of ProveCommitAggregate Usage after Hyperdrive upgraded - Steven Li (IPFSForce)
41:30 - The Optimisation Analysis of GAS By Athena Fool - Peng Weng (RRmine)

THIS EVENT ABIDES BY THE FILECOIN COMMUNITY CODE OF CONDUCT: https://github.com/filecoin-project/community/blob/master/CODE_OF_CONDUCT.md

A

All right welcome everyone on another great meetup session today, from from the community to the community. uh The agenda of today is uh is existing of five topics, even though, if you break it down, we basically have two main uh opportunities here to to talk about branding and hyperdrive.

A

As most of you know, uh more than a month ago now we have actually released a hyperdrive improvement to the network uh which accelerated the ingest um and today, we'll we'll have uh four storage providers or miners with us that will be sharing their best practices, their observations uh with the community, but first and foremost, we'll talk about uh storage provider, rebranding, um something that the community has been asking for for a while and so we'll dive into that fifth proposal in a second and after that, we'll have tim zhu from uh 1475., charles from phil swann, stephen from ipfsforce and penguin from rrmind, get us or present us on their findings all right.

A

Let's start! First of all, we love uh miners. um It's no doubt to you. Guys builds the foundation of the falcon network, and uh it's thanks to you that we can actually um fulfill our long-term vision of storing humanity's most important theater and um now. What the problem is, though, is that when we start talking about mining- or you know us being miners- that the traditional web 2 clients or the enterprise customers, don't always understand what it means to be a miner, they think proof of work. Energy consuming defy only no utility.

A

Now we all know, that's not correct uh fall coin is a delivers, real value, real utility is green, is uh proof of state stake, and so many other uh opportunities and and businesses are being built on our network. So what have we done in the last couple of months? We actually, we actually talked to a ton of miners, and it started off as an initial request in the uh community from the community um to rebrand the word uh miner and, and so the reason there was to better connect with our actual end customers.

A

Our clients um two is to convey the network, usefulness or the utility of falcoin as a whole and then three um be a leader in web 3 world when it comes to you know, provisioning or providing the decentralized storage infrastructure that will support humanity's most important information, and so that's exactly what we've done um in the last couple of weeks. We have actually built a proposal and we now love not just miners, but we love our storage providers.

A

So really the proposal is a fip and we'll talk about that in a second, the details, um but this is currently um in last call.

A

Basically, if you look at what we did is we we rebranded the store, the the minor terminology to storage provider terminology, because what we essentially do uh with falcoin is we connect clients to storage providers. We provide a certain service, we provide a storage infrastructure and and and of course, so many other services on top of that are being built. So fifth 18 is currently listed on github it's in last call stage: we've gotten some really good positive feedback. So far, I recommend everyone to have a look at it.

A

The url is, uh is right here on the bottom right, look for fit fit 18 and you'll see a table of the terminology, changes that we are recommending. You'll see that you know we are moving from miner to storage provider and we're moving from retrieval miner to retrieval provider. Of course, over time this is a journey that we are taking together over time.

A

This could we can improve this, and our goal is the is to be more aligned with what our end customers are looking for, that are storing data into our um into our filecoin network. Also, a couple of things that we've adjusted is instead of mining block awards. We focus on block, rewards, minor pool storage pool. We are focusing and emphasizing the storage capability that we provide to our end.

A

Customers have a look, provide us with your feedback, so far, very positive, positive feedback and if everything goes well, we'll uh finalize this on the 20th of august um oops. So what is the ask? uh Once approved? uh We asked the falcon community to adopt and help communicate these new changes.

A

um This is a request from the community, um and so we are asking uh everyone in the community to help us with integrating this terminology in your ux, your tools, your command line, marketing assets, anything that can help us to promote this new terminology would be fantastic as we build new marketing assets and we go after uh those and enterprise customers all right.

A

So with that again, my ask as well: let's have a look at the github proposal, provide us with your feedback, we're here to listen um and work with you on what the best recommendation is moving forward. So what is next, as I mentioned, we have four exciting partners today for storage providers. Now that we talked about the new uh messaging, that will tell us a little bit more about uh their experiences with hyperdrive and their recommendations, and first up is um uh tim tim zoo from 1475.

A

I'll pass the ball to you. You can share your screen out and I will go in the background.

B

Cool cool. Thank you thank you. So this is ting. um I watch my screen. Okay, cool hi, everyone. This is ting from ford 1475, I'm very happy to be here to bring you the presentation about bitcoin ecological development after how to drive upgrading okay. So, let's see the network power grows. First, from june to august we see consistent uh growth uh on network power and the growth rate is still accelerating.

B

This is pretty cool and how about the biscuits b, we can figure out that this test fee dropped greatly and remains stable within limited space after after hard drive upgrade. There are three main reasons. Firstly, the time limit between prove commit and pre-commit messages has been adjusted from one day to 30 days, so is no longer hurry to send the proof commits messages after pre-commitments.

B

Secondly, pre-comments and proof commits can be benched into one message which compresses the gas used required force for selling sectors. The third reason is proof command. Aggregate messages has added an aggregation fee, so in the economic sense the biscuits fee is locked in a certain range of about 0.13 nanofare to 0.2 nanofara. I will show the details later: okay, so let's move to fffip8 pre-commit sector bench effect on reducing gas.

B

So from my from our calculation, the pre separate, the pre-bench type of limit of packed sector's number is up to the maximum is 266.56.

B

And the bench efficiency of reducing gas is from five times to 18 times.

B

I suppose sap certain the proof commit aggregates messengers if we take a 10 10 per byte node for an example, we assume it's a sector size is 64 gigabyte. um We can see, we can see the trust set, the gas rate so get the efficiency of gas reduction and the base gas fee. Break-Even point both depends on the number of proofs matched the more some more uh some more proofs. You benched um some, the more efficiency you can reduce your gas and um everything is the difference.

B

Is it's tiny if, if, if you're, um but if your sector size is 32 gigabyte, okay, so next I will show you the network maximum capacity of the buy coin uh pipeline network after the hard drive, as we can see,.

B

Since each tip site contains an average of five blocks and the gas length of each block is 10 billion, the cup, the target gas limit of the whole network is 72 trillion and for now the gas use is about 50 trillion a day. So there are.

B

It means that at least 70 percent of the gas limit has been used, that is to say, we have two potentials to um to speed our network capacity, so one is more manners using pre-bench proof aggregate messages to see sectors. The second is the increase of gas used on the network. Okay, I will do a simple calculation.

B

It assumes that everyone use a sector size of 64 gigabytes. We can. We can see that we can see the single sector um pre-commence gas here about 200.

B

To too many excuse me, um we can see the single sector proof companies get here about uh about 1 million, so we add we added them up. We can have the whole gas for single sector selling it's about 3 million, so so the maximum number of sectors sealed in a single day for the entire network can be up to 9 million, which means that the maximum daily network power growth is up to 573.

B

Pentabytes, which is very amazing, cool, so to conclude, the hard drive upgrade is based on the optimization of the code of serial knowledge proof which not only greatly reduce the gas use for sector steering, but also significantly lowers base gas fee without changing the block gas limit. In a word, it has made great achievement in cutting down both quantity and price of gas, so the guessing of providing distributed storage services on the firecoin network is one step of significant improvement.

B

We believe that the firecoin network is on the road to real data expansion, ritual mining, smart contracts and the decentralized storage network for the most important information on the earth.

B

Okay, so as for 14 19 1475 uh team, we have we have been dedicated to the construction of firecoin eco ecosystem for a long time.

B

For a long time, we were honored to be one of the north race as the first group, and we have approved 15 clans, with a total of 100 terry bytes, real data, and in may, and in may 2021, we we we was, we were granted one pentabyte quarter and was re-elected as the notre in china again.

B

In addition, our fort 1475 team published rail data related fip 16. for now cc sectors can only seal garbage data and fail to up upload real data by through this vip it always the waste of storage space, as well as to improve storage resources utilization with subject with subsequent proposals in the future. It can even convert the cc sectors with real data into verified their sector, so that so that the sector can enjoy 10 times of power.

B

That's a huge improvement!

B

Okay thumbs up. Thank you for listening. If you have any questions, you can scan the qr code of richard to talk with me after the meeting cool.

B

Thank you thank.

A

You thank you so much then. I really appreciate those uh great numbers. um Definitely yeah. The calculation was very helpful. Anyone who has questions, please point, um please add them in the chat, uh we'll see how we can respond to those questions um as the presenters present um great now. Next one up is charles and thank you so much tim, as always um charles from phil swann ready to there you go.

A

Thank you so much take it away.

C

Hey everyone- and this is uh charles from plus one team, so as euro would bring to everyone about how the hyperdrive impacts the ecosystem development. So, as everybody knows that first one team is dedicated to the first coin: ecosystem development, especially on the application side. So we are dedicated to build a hyper storage and the computing solution on web three based on fire point. So this is the gut speed changes over the last year.

C

You can see that we have a high level at around five nano and now at a very long time, we have average on three. Nano recently is around. 0.2 basically means that we have reduced 90 percent to 95 percent of gas feed.

C

So for the given side, what does that mean when you have so much gas fee decrease? So this is the current network, which I just take a screenshot few minutes before the meeting start.

C

You can see that the average price is close to zero per year and uh the current minimum size of five is four gigabytes actually, because the gas price becomes even lower, which means the perfect become lower. People are more willing to send the small deals instead of large deals same as a minor premise like miner doesn't want to accept any deals smaller than 80.

C

So with this decrease um application, developers can use much more small data, so you can, you can retrieve faster, you can upload faster and you give you much more space to control your application yeah. This is a two data which is really very interesting from the last six months. First, one keep on watching the internet work price of average miner. We have a system sign system with over 100 miners worldwide we recording their daily price changes.

C

So this is what we can find is the average non-verified year price is dropped from 0.65 per year to 0.3. Basically, you see 50 percent of your cost where you save data, another site is a verified idea, even though the miner doesn't charge anything from verified ideas, but it still causes the burn fee. So for now the profit for the miner is dropped from 0.2 per year.

A

C

C

Which is great for miners so, basically now we don't really need to worry about when we take the template data and we maybe lost some phase due to the gospel which you give which will transfer those cost savings to our applicant developers so who will benefit from those changes? There are two group of people who have been benefited directly while in the screen shutters as everybody knows that every one or two months we have selling shorter competitions in the friday community.

C

Currently we are on the pro 2.5 selling chart and you can see this is. Overall. Storage has been saved on the network yeah, so in this frame we have about 700 terabytes. Previously we have 21 petabyte globally. Imagine how much burnt gas fee has been spent on those 21 petabytes, that's huge number of firepoints, but now it's becoming easier and those other applications. Potentially we are using those uh storage service on the pricing network. This screenshot is taken from the frequent community collective discussions. You can find it on the github link.

C

If you found your application is not there, you can submit it to the github and get your logo. Probably shown here, those are some categories of services. We found very interesting for storage providers and we just put some typical application who's working on it.

C

uh The first, the most famous one, is ipf as a frequent storage. Those ones is directly connected to the frypond permanent storage, which means that the reduce was industry where directed those applications.

C

Textile is very famous in the community people using it a lot in the selling shot and the recently they have a straw and slate, which is allowed to upload single file and images to the network. They have a bucket design, so they do it a schedule based their bucket to the storage network with the gasket decrease. Now it become much cheaper recently. They also have to develop a pro-life new service, called web3 dot storage allowed to upload a single file to the final network and the good news that that storage is free forever.

C

The reason is because using the verified data miners willing to selling the data with zero price, because gas gas is also very low, it can be ignored.

C

The difference and the payment part become interesting as well, because when you're doing the d5 currently like a frequent doesn't have a smart card moment, even though I know developers are actually working on it for the vm.

C

So when you're doing the transfers, basically like you're doing something like uh payment harps, so you need a lock some file to uh then on the other side, you wrap it to a instagram token or other blockchain token. So we know that some projects, like codify, secure finance and raptor token, so when you're doing those sending files uh adding so those transactions, it does cause boom free as well.

C

uh If you have very high fee on this end, then the total payment- you maybe have five percent or six percent loss previously, but now because the gas is very low, you can keep 99 percent of value, but still on the other side, the ethereum chain, maybe cost you a lot of gas v. But it's a consumer problem. That's not a fragile comfortable.

C

Another interesting part is sitting access because I know milo team is building a cdn solution for different.

C

Users in the global wides can faster retrieve their storage, while you're doing the storage retrieval. You also causes your burn free to initial the transactions so with this kind of decrease of the gas speed, which means that the trigger price will be the total cost of the retrieval website as well. It helps the management to building uh even stronger, distributed signal network globally.

C

Another two interesting currently we are working on is the first one is working on the hpv solutions to bring the computing data close uh computing resource closing to the data leave pl doing something related to the video transcoding. While we are doing those works, we need to retrieve all the data and upload the data from the frank network. If the gas space is too high, it will cost us a lot of fun. At the end, the salespeople become the cut for the users.

C

Imagine some video transcoding, maybe just uh one or two bucks, and your gas fee also costs the one or two bucks, which means that double the cost. But now we don't have this issue so just speaking that with the gas fit decreased by the hyperdrive, both applications. Now become much more useful and we can see that it make us more close to real world and especially the storage part compared to aws.

C

When we're using verified data, we can see that we almost has unlimited advantage compared to aws yeah thanks.

A

C

And this is the practice by the brass one team. If you have any questions you can email to us or you can find us on the slide, and you can also put us your questions on the channel or on the chat room we can answer later. Thank you.

A

Perfect, thank you so much charles uh next one went up uh stephen lee from ipfs force. Thank you for joining us today, I'll uh pass on the ball to you.

D

A

D

Yes, sign steven and hello everyone um yeah. Let me share my screen.

D

B

D

Songs, okay, yeah. Let me play this slides, yeah and hello. Everyone- and this is stephen from uh ipps force community and today I want to share with you about the um data analysis of proof, commit and proof commit aggregate usage after hyperdrive upgraded in falcon network um yeah. Here, I'm not showing about the um the pre commit and yeah, because um yeah that one is quite easy that will either way yeah save uh your cost of gas and yeah.

D

If you use that um the pre-company batch and here we're talking about the um improved, commit yeah, I want to show you first. I want to show you the two pictures which is from the filescan.io.

D

One is about half one year, the basically the price variations and the bottom one is about three days as you can see. Yeah. If we see the last half year yeah, we see um the base bay.

D

Vibration is very, very huge and you could say it could be higher as five ladder file and after they hyper that upgraded. We yeah, we can see. um It's yeah keeps in a very reasonable level, yeah it's around one, that is one 0.15 ladder file and um yeah silder. You said when uh they're about to oh a lot of file, it's quite stable.

D

uh I said this is the first time for the falcon network uh to have the um yeah and to have the gas base phase, keep a very reasonable and very stable level and to make the service providers can really forecast their um their gas cast for the power growth, which is very good and um yeah. From the um the planning point of view and the finance point of view, it is a very good news for their service providers and for uh the potential, um the pro their providers to join the falcon network.

D

So and after the hyperdrive upgrade. We see that the um yeah per day power growth is and actually increased about, 50 percent um yeah, so why the basically is around 0.5 largo file. Actually, this is by design um yeah. If you go to the 5th 13th, you could say that we have a the parameter here, which is fixed, which is called the batch balancer, which is too narrow, fail and um yeah.

D

If we adopt uh the the proof coming aggregate message and your batch gathering is actually has to be, and it has to be equal or larger than this one so um yeah. This is an option yeah for the service providers and to consider, if you'd like, to take this approach and actually depends on the status and the um actual network.

D

Basically, and also your yeah, your total part or your power growth suite.

D

And yeah here so um let's come to this table um this one. We actually take one example yeah from the falcon notebook. um I would.

A

Say this is a zero load.

D

And yeah, and I'm not mentioning which node here, but this is a real one. We choose the data about zero load and to say um yeah. If this load uh only use proof commit or yeah, it only use the proof commit aggregate, and because this load is actually about um 600, theta bytes um progress each day and yeah. If we see this one um yeah you you.

A

Can say next page and.

D

On the right, um it means, and they an orange line is actually yeah. If this miner only do proof count aggregate, so you could say that the original is almost capable yeah the same and no matter the base state is uh higher or lower right yeah.

D

It's almost cape about um yeah, six, four or 65 at this level, but if uh this the provider only take proof commit and yeah it's about the blue line, as you can see, yeah, if you only take the proof and commit um well, the base rate is lower and the and and yeah the gas cost for the poor growth will be lower and yeah. It will be higher when the gas price is higher, so um we can say, there's a um yeah.

D

There is a bonus point which is about what that 1.13 another file, so um yeah for a miner as well nobodizing will be. If the price is higher than 0.13, it will take the proof, commit and aggregate yeah. If it's lower than zero point wesley, it will take uh yeah, they prove yeah commit. So it's a reasonable decision for um yeah for people.

D

So that's why we say that they.

D

Yeah, they can't say um the base cluster escapes. They say yes, because the system is actually designed by this, and also it's a negative feedback system and the miners it will adjust it and to keep the baseline exists and also um by the way I want to mention yeah. This is the 64 gigabytes, six size provider. If it's a search, two gigabytes, the second size provider, it will be a little bit different yeah.

D

It will be a little bit higher in about um 0.15 to one nine, something like this yeah, it's almost um not that big and that big difference. So, let's come back to the network.

D

We have done some statistics so in the whole network. Currently we have about 2 000 and 2 650 storage providers and after and yeah scan all the providers yeah, we say about only 770.

D

The providers um and adopt approve, commit and aggregate messages, so only 30 percent of providers did it and all other 70 percent providers not ever tried obligating their proof committee yet, um and we will also say yeah. It has reasonable choice actually, um and we also see all the top manners and actually enabled aggregate net proof commit, which means all providers which power is higher than 10 petabytes, enable an aggregation.

D

This is because, um yes, the miner, whose power growth is higher, they would like to take this because their um gas class is higher too.

D

um Yeah, but I will say that this is only factor um most people considered very loud, but there is also another factor. I think that is the way the small providers um yeah. Even that didn't try. This is because there is one factor we need to consider is about the return.

D

So if you delay, if you use the proof coming aggregate, actually your dated return, which means you you yeah your poor growth- will be um yeah will be later proved so yeah, you will be, um you will delay you broke.

D

Because the power is not at the new level, so you probably return what your wallet at the dual level. So I have done some math here. If you have one day delay of a six to four gig sector, you will about lose um 0.00234.

D

Fail, which means that one day dna is actually um likely equal to zero point zero five line, nano fail-based, fake cost, so yeah for small and medium providers they would choose not to aggregate is because their progress is not so fast. So if you use it, you use this yeah. They will delay a longer time and yeah. The rules about the block reward is also um very yeah, very high for them they are comparing to their um yeah, their cost and the other cost.

D

So this is also one thing. um People need to consider here, and another factor we need to consider is that the um yeah, if we can say in number one uh non-term uh actually, we will say that the initial plate per sector currently is going down, but because we have the batch manager, they are fixed. So, yes, the gas burned keeps the same.

D

So if we do this yeah, if we keep this, the gas cast portion might need to be visible.

D

Yeah you to be revisited and adjusted later yeah, but it's not very numb yeah, perhaps and after half a year or one year, we need to revisit this and, as just adjust the bus manager yeah. This is one thing I want to mention here.

D

And another thing um we captured from the file and from the file scan.io yeah. We said that from all the messages yeah, if we yeah- if we, if you count all the messages and to calculate the proof, commit sector and the proof commit aggregate.

D

The true message is actually take about 95 percent gas, which means that the proof actually also consumed the both both of the um network, binaries um yeah, which is also very significant.

D

um I would say that yeah for long run. We need to think about this and continue to try to reduce this and the proportion in our network and to new bandwidth for zero deal and for real data storage energy. So this is why the neutral project comes in. So um this may not be a very top priority right now, but I would expect that this should be considered, perhaps a half year or in one year.

D

Yeah, okay, that's all uh I want to share here and uh uh I want to remind uh all the providers and all the developers about yeah from my point of view, um yeah by the way I'm from force. uh We are a web 3, an infrastructure, yeah service provider and the uh ips file coin is our main major area that we're working on and promotes the whole ecosystem.

D

Sense for your hand, and if you have any questions yeah, you can and you can put it in the chat room, and I can answer this there. um Okay and after my presentation, uh our next uh yeah, our next presenter, um it will be penguin.

D

He is from the top r minor.

D

E

E

D

E

E

E

E

E

E

E

E

A

All right, thank you so much and uh again appreciate everyone uh helping us out this weekend and taking on the hyperdrive upgrades some great insights. It's also great to see um all the sessions right after each other as uh new things myself included from each and one of you so uh again appreciate it and um looking forward to the next meetup.

A

If you have any other questions, suggest or upcoming topics or something you would like to hear more about or do a deep dive on, please let us know go to uh the github, uh follow coin ecosystem page and uh just submit your questions there. uh We are monitoring, uh obviously on a constant basis, our github page and also our slack channel. So with that, that's it for this week. Thank you. Everyone and hope to see you next.

E

D