Filecoin Filecoin Storage Providers, 17 Aug 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: HyperDrive Upgrade - Data Analysis of ProveCommitAggregate Usage by IPFSForce

Description

Steven Li from IPFSForce updates the Filecoin Community at the first Storage Providers Show and Tell. With a focus on the HyperDrive upgrade, we learn more about the data analysis of ProveCommitAggregate usage.

A

This is steven from ipfsforce community.

A

And today,, I want to share with you about data analysis of provecommitaggregate usage.

A

After hyperdrive upgraded in filecoin, network.

A

Here,, I'm not showing about the precommit,.

A

Because that one is quite easy.

A

That will, anyway, save your cost of gas,.

A

If you use the precommit batch.

A

Here, we're talking about the provecommit.

A

First,, I want to show the two pictures, which is from the filescan.io.

A

One is about half one year: the base fee, the price variations,.

A

And the bottom one is about three days.

A

As you can see,, if we see the last half year,.

A

The base fee variation is very huge,.

A

You could say it could be higher as five nanofil.

A

And after the hyperdrive upgraded,.

A

We can see it keeps in a very reasonable level,.

A

Is around 0.15 nanofil,.

A

Seldom exceed 0.20 nanofil.

A

It's quite stable., I think,. This is the first time for the filecoin network.

A

To have the gas base fee keep very reasonable.

A

And very stable level, and makes the service providers can really forecast their gas cost for the power growth,.

A

Which is very good.

A

from the planning point of view.

A

And finance point of view,: it is a very good news for service providers.

A

And for the potential providers.

A

To join the filecoin network.

A

so after the hyperdrive upgraded,.

A

A

The per day power growth is actually increased about 50%.

A

So why the base fee is around 0.15 nanofil?.

A

Actually,, this is by design..

A

If you go to the (indistinct),.

A

You could see that we have a power meter, here,.

A

Which is fixed,, which is called the batchbalancer,.

A

Which is 2 nanofil.

A

and, if we adopt the provecommitaggregate message,.

A

Your base gas fee actually.

A

Has to be equal or larger than this one.

A

This is an option for the service providers.

A

To consider if you'd like to take this approach.

A

Actually,, it depends on the status,.

A

The actual network base fee,.

A

And also your total power or your power growth, rate.

A

So let's come to this table.

A

This one, we actually take one example.

A

From the filecoin network,.

A

I want to say this is real node.

A

I'm not mentioning which node here,, but this is a real one,, which is the data about read, node.

A

If this node only use provecommit,.

A

Or only use provecommitaggregate...

A

This node is actually about 600 (indistinct)bytes.

A

Power growth, each day...

A

You can see this page, on the right,.

A

It means the orange line is actually,.

A

If this miner only do provecommitaggregate,.

A

So you could say that the orange line is almost keep the same,.

A

No matter they basically is higher or lower,.

A

It's almost keep about 64 or 65,. This level.

A

But if this provider only take provecommit,.

A

It's about the blue, line,.

A

As you can see,, if you only take the provecommit,.

A

While the base fee is lower,.

A

Gas cost for the power growth will be lower,.

A

It will be higher when the gas price is higher.

A

We can see there is a balance point,.

A

Which is about 0.13 nanofil.

A

For miner,, a reasonable decision will be.

A

If the price is higher than 0.13,.

A

It will take provecommitaggregate.

A

If it's lower than 0.13,, it will take provecommit.

A

So it's a reasonable decision for people.

A

So that's why we say that.

A

Base gas fee is kept the same.

A

because the system is actually designed by this.

A

And also it's a negative feedback system.

A

And the miners will adjust it and to keep the base fee like this.

A

By the way,, I want to mention this 64 gigabytes, sector-size provider,.

A

If it's a 32 gigabytes, sector-size provider,.

A

It will be a little bit different,.

A

It will be a little bit higher, about 0.15 to 0.19, something like this.

A

It's almost that big difference.

A

So let's come back to the network.

A

We have done some statistics.

A

In the whole network, currently,, we have about 2750 storage, providers.

A

All the providers, we say about only 770 providers.

A

Adopt provecommitaggregate messages.

A

So only 30% providers did this,.

A

Or other 70% providers not ever tried.

A

Aggregated provecommit yet.

A

We will also say it's a reasonable choice: actually,.

A

And will also say all the top miners.

A

Actually enabled aggregated provecommit,.

A

Which means all providers which power is higher than 10 petabytes.

A

Enabled aggregation.

A

This is because.

A

The miner, whose power growth is higher,.

A

They would like to take this because their gas cost is higher, too.

A

But I will say that this is only factor.

A

Most people consider right now.

A

But there's also another factor..

A

I think that is why the small providers.

A

Even didn't try this.

A

Because there is one factor we need to consider is about the return. so,.

A

If you use the procommitaggregate,, actually you delayed the return,, which means your power. Growth will be later proved.

A

So,, you will delay your block.

A

Because your power is at the due level,.

A

So your block return will at the due level.

A

So I had done some math here.

A

If you have one day, delay of a 64, giga sector,.

A

You will about lose 0.00234fil,.

A

Which means that one day delay is actually.

A

Likely equal to 0.05 nanofil base fee cost.

A

So for small and medium providers,.

A

They will choose not to aggregate, because their power growth is not so fast.

A

So if you use this,.

A

They will delay a longer time.

A

The loss about the block reward is also very high.

A

Comparing to their cost.

A

So this is also one thing: people need to consider. Here.

A

And another factor we need to consider is that.

A

If we can say, in long run,, long-term,.

A

Actually, we'll say that the initial pledge per sector.

A

Currently is going down,, but because we have the batchbalancer,.

A

They are fixed., so the gas burned keeps the same.

A

So if we keep this,.

A

The gas cost portion might need to be revisited.

A

And adjusted later., but it's not (indistinct).. Perhaps after half year or one year,.

A

We need to revisit this and adjust the batchbalancer.

A

This is one thing I want to mention: here.

A

And another thing: we capture from the filescan.io,.

A

We say that from all the messages,.

A

If you count all the messages and to calculate the provecommit sector,.

A

And provecommitaggregate, the two messages actually take about 95% gas,.

A

Which means that the proof actually.

A

Also consumed most of the initial bandwidth,.

A

Which is also very significant.

A

I will think that for long run,, we need to think about this and continue to try to reduce this proportion in our network.

A

And to leave bandwidth for real deal and real data, storage and retrieval.

A

So this is why the neutron project comes in.

A

So this may not be on very top priority right, now,.

A

But I would expect that this should be considered.

A

Perhaps in half year or in one year.

A

That's all I want to share here.

A

and I want to remind.

A

All the providers and all the developers from my point of view.

A

by the way, I'm from force,, we are a web3.0 infrastructure service provider.

A

And ipfs filecoin is our major area.

A

That we're working on and promote the whole ecosystem.

A

Thanks for you, time.