IPFS Virtual Meetups, 28 Jan 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Introducing Humanode, with Victor Smirnov

Description

Paradigm CIO Victor Smirnov joined us to introduce the community to Humanode, a distributed consensus mechanism based on private biometrically encrypted equal human nodes, and built with IPFS.

For more information on IPFS
- visit the project website: https://ipfs.io
- or follow IPFS on Twitter: https://twitter.com/IPFS

Join your local IPFS meetup to attend our next event: https://www.meetup.com/pro/ipfs/

Sign up to get IPFS news, including releases, ecosystem updates, and community announcements in your inbox, each Tuesday: http://eepurl.com/gL2Pi5

A

Hey everyone: my name is victor. I'm a core member of the humanoid team, we're building a distributed financial system where one human equals one node like or, as we say in ipfs community, one human equals one replica. Why? Well every public permissionless now their blockchain networked. They are based their security on uh capital.

A

They base security, either on state tokens or on equipment of miners and well these this approach is susceptible to external attacks, like when you have a network with higher capitalization that this one it can easily attract it, either directly buying out assets or it will create a secondary market outside the target network and basically manipulate the price of the asset in eventually buying out the target network asset and attacking this permissionless network and what we're doing in humanoid. Our security is based on collective verification of human existence and all people who became humanoids.

A

uh I have to say frankly we had that idea like almost for three years, but we had no idea how to implement that and it became possible only in 2020 thanks to advances in biometrics, especially leaveness detection, uh thanks to advanced and matching of fully homomorphic encrypted biometric data, uh thanks to the private computations in permissionless manner. By uh guys from enigma and, of course, well huge kudos to the protocol labs team uh for inventing and implementing the merkle certity cluster.

A

Well, basically, when we have a network where one human equals one node that is capable of financial transfers, that means that the fees that the network that the protocol collects will be divided equally among every human node, as there is no block reward because again we're using the ipfs invention, which is uh miracle certainty.

A

Any monetary system can be built upon without these block rewards and, moreover, it opens doors and the way to the one human one vote, decentralized organizations and, as you have seen, the like state tokens tao are just terrible at the moment.

A

Well, uh thanks to the blockless nature of merkel, certit we've got pretty cool features there compared to blockchain, first less redundant, metadata and the second flexible transaction persistence time.

A

uh That means that, uh if you're sending like ten dollars to your friend- and you don't really want to pay the cost of this for this transaction, for this data to be kept forever, uh what we can do with miracle strategy, we can unpin the data from ipfs, not hold it anymore, and basically the costs become lower and lower.

A

Of course, if you want to keep the transactions feel the heat death of the universe, you're always welcome.

A

Now. How do we implement ipfs here and uh where do we use it basically everywhere uh and uh let's start from the like creation of humanoid? Okay, uh what we do we ask people to scan their biometric modalities?

A

Yes, first, we're using like external biometric modalities, uh face recognition, fingerprint and iris recognition are among the most advanced, for example, uh but you know later, if we find out that that's not enough for security, uh we're also going to introduce in internal biometrics. uh That means neurosignature with brain computer interfaces or dna.

A

But let's go to the slide, which explains: how can you create a human node and like how the system works? We have three actors here: it's a client device, public ipfs cluster and a private smart contract.

A

So this public ipfs cluster, it actually gets uh well, it's public.

A

You will ask how we can put biometrics such as sensitive data there. Well, we first encrypt it with fully homomorphic encryption right on the client device, and only then we send it to the ipfs cluster.

A

Basically, when a potential human node submits its encrypted feature vector from the biometric modalities, the public fs cluster performs one-to-many computations over that data to find out whether this person was a humanoid before or he's really a new one.

A

Well, basically, after that, we get the score, we get the score that tells us.

A

Has he been there before or not, and uh this core is then sent through the client device to private smart contract. This core is again encrypted. We have to decrypt it right there in the private, smart contract and the private smart contract understands what this core means, and then he authorizes the public key of the client to have a validator key or not to have it.

A

So now we like have permission for a new human node to finally join the network. Let's look at how the network looks like uh yeah. Here we came up with a nice figure for miracle certity, but I'll jump straight to the tech.

A

uh We reuse existing tools from the ipfs project to implement networking and dental replication level. Here uh we use miracle strategy for the state duplication and it becomes possible now to implement a key value store, which looks like miracle trees in avl from cosmos.

A

So, thanks to the protocol labs team, we can leverage data store component that provides a unified interface to access data, uh we're picking a distributed. Go data store implementation using merkle certity.

A

Well, it is basically based on the badger db as permanent storage. It uses uh leapd pops up and the provided pops up broadcaster as a broadcaster component, a user defined dac synchro component to publish and retrieve merkle dax to the network and for the interface, the node initializes, an ipfs, lite client.

A

So thank very much ipfs team and protocol labs. You were a great inspiration for us.

A

Basically, when all the external components are properly mounted, the data store is ready to be used by an old instance.

A

Since it's a key value store, we use prefixes to define separate data, buckets and well here for the application logic, which represents transaction processing and validation logic. We defined two sets of buckets.

A

The first set includes epochs, which we use instead of blocks in classic blockchains, and each epoch has a number as a key that points to a list of transactions and on application level. Again. Transactions, which is a bucket holding set of confirmed transactions and consensus participants here, decide whether to include or not to include transaction into that set.

A

If the transaction is confirmed, it's cid content, identifier and validator. Signatures are added to the set and the body will be replaced through dac sinker.

A

Well, basically, the second bucket represents uh accounts and validators well, this is this is where we like, combined the knowledge that we have from the public, blockchains uh cosmos, mainly and ipfs, and this is how we'll build the humanoid network.

A