Trias: Can We Trust Machines ?

Michael: What’s up everyone and welcome to Boxmining today we’re gonna take a look at a very different project in the crypto
space which is Trias. And Trias the project is about having trust in
machines.. When I first saw this I’m like dude is it Skynet coming is it
that soon do we need to trust the machine and the AI algorithms. The
simplest way to describe it if you’re a gamer is if you play video games there’s
gonna be cheaters so essentially they’re hacking their machine to gain an unfair
advantage in a video game we can’t trust their client their
machine and this isn’t just in gaming in a lot of enterprise applications in IOT
devices and in AI you need to make sure you trust the Machine and a code that’s
operating and this is why trusted computing is a huge field and I
interview one of the specialists Dr. Ruan AnBang he’s done his doctor of
philosophy on this subject and he’s also the CEO of Trias. He solved this
problem using an approach that we all kind of are familiar with the legal
system is separated into three different components each of them with limited
powers and he’s also dividing his blockchain into three components and we
talk about why this is important the industry it applies to because not
everyone wants to bring everything on to Dapp they want to run on their own
software and code as well and this is why Trias is there to make sure that
this is legitimate enough to use and to share between everyone in the network I
hope you guys enjoy this video and learn something from it
so Dr. Ruan can you tell us a little bit about what Trias. Trias is about trustworthy machines so
we use it in the separation of powers way so by special power we need first to
define the executive which is a machine which can run the actual correct code and
secondary we need legislature which is a law defining the actual behavior of
Records so this is the second system and the third one is judiciary which
identify who does what who will get what and if anyone does anything bad they
will get punished. So we use this philosophy in the cyberspace – to allow people to identify what is really going
on inside the cyberspace right so so you’ve been looking at the
idea of trusted computing what does trust mean when it comes to
computing you want to trust someone a guy this means that he can actually do
you actually do work he said tell you to do so by trusting a machine for example
we open the application on a machine when you send data to it youyou are
actually expecting that machine to process at a specific way and
in return the data for you but actually a machine up to now is a black box you
don’t really know what was going on inside it and we all know that the
machine is driven by applications so basically we need to solve three kinds
of problem one thing is that we need to know well this is a machine can actually
load and run the given application a piece of given application. There are situations that would not give an application to a machine and because the
machine was hacked so not much actually hacks my application by adding some my
no catches or infamously malicious behavior which is just like a virus
so right the first thing is very critical can can we actually trust one
machine can actually run the applications that is given by us human
beings yeah I was thinking about this earlier the examples I kind of thought
off was when you’re playing a video game when you play games there especially
competitive games there are some people who choose to have the game and
introduce something like a program that can help them aim or even see through
walls to give them an unfair advantage so what you’re talking about here is
that you’re trusting the environment that the application will run ads as
promised kind of as you’re trusting that the application is free from external
hacks right regarding the patient’s one things that when we or someone else when
I get the application the binary data of the application we might have the
incentive to change that binary a bit so that we can shoot inside again also maybe some of other people can cheat so in
order to detect what a machine will behave we somehow need to change your
application yeah so the basic way of Trias is what we called the
Leviatom network is actually a trusted computing network by trusted computing
we mean it to be an specific field research field in the operating system
security research area yeah so we use some small chip which is called the TPN
and currently we also used the Intel SGX capability and 2x SMT
so with this kind of chip so we can actually know what is actually what
kinds of application testing loaded inside a target machine
so with this capability onboard we’ll be able to tell whether your
machine is running a given application so this is the first layer the first
thing we’ve been solved a trusted computing environment run a given application. So that’s the first layer and now what is the second and a third
layer because you’re Trias so you expect three layers right think about the other
question that’s given an application if because of the application is develop is
developed by a third party. how can we tell that the application is generally in doing the
right thing for example if we got a correct machine but that application
itself contains malicious behavior so the behavior will be correctly
enforced then yeah we still got cheating so the second layer is all about the
telling the origin of a piece of application yes so we we use an
acknowledged technology called Prometh we use blockchain to record
the entire lifecycle for developing a piece of software it’s
just like if we can have a recorder to record a human baby’s entire life then you might be able to tell piece of track characters from
from their past behaviors so it’s just like we put any camera on to the actual
development process of entire different process of application then for example
we can recall the identification of any piece of source code and we can
tell this source codes have been compiled into these kind of binary
objects and then assemble scatter through these applications so with this
capability enforced we will be able to tell whether least piece of applications
contains malicious behavior because if we agree that all behaviors have coded inside the application So layer one is about the the Machine layer – Layer 2 it’s about the application and now what’s the final layer a lot a lot of services
service we need co-ordinate trusted machines with trusted applications we
need to know that this has to be sent in their application running on these kind
of machines so layer is actually our smartphone channel yet and using these
kind of small function languages to write decentralized software as a
service it is a little bit bigger than the sensor because you know Software as a service is a combination of different applications and then work together to
implement a service. Now all this is kind of like completing the Trias cycle
so you have to trust the Machine trust this software and then your software as
a service now what kind of problems does this solve over all so you built this
amazing machine what is kind of the target solutions. In our point of view, we want to enable trustworthy enterprise
applications so what can really solve this said you can help enterprise to
prove its trustworthiness to their customers
eg. When we send data to Facebook we actually don’t know where our data is stored and how its analyzed and who saw our data so if we got a way to build the
entire Facebook infrastucture as a blockchain smart contract
imagine that entire Facebook infrastructure as an actual smart contract so
we will be able to tell any piece of what work has been going on inside that
big application who has all data so right now that’s
that’s kind of scary part we passively trust a lot of companies whereas
Facebook or app or Google we trust that they’re not doing something really a lot
of spying on our data but what you’re saying that will try is is that once if
they migrate their services on to triose we can know for a certainty that there
are certain the execution is gonna be clean and that there’s nothing no funny
business when it comes to our data and even consumer and enterprise users you
can just trust that execution right ok so you’re talking about this trust and
you’re adding the blockchain element to it now one thing about blocking is that
it’s all public so how do you kind of make the execution more private or
privacy preserving so we don’t need to include all the royalty sounds were
watching you know so identification ability I can be put on and the
identification of a piece of software piece of vacation can also be put on
instruction and then the relationship can can can be recorded it’s like if we
if we check this big ledger we can actually know that which application has
process which takes a but we actually don’t know the actual size our only
application we love some some privately held key or something of that yeah so
some projects is Sophie and the auditing problem but we actually keep the most
critical critical and so the data application layer
separate something that we kind of discussed here is we have this kind of
try a system that you have set up but how is it kind of different from the
traditional blockchain something like etherium that has Nakamoto consensus you
know how is this three layer system different from what we saw seach
classically application application I’m making pigeons we can see as like like
the Microsoft Office the entire application so so in order to do this I
think it might be really hard to to reprogram to recall all in the office
source code and to just come and then we caesareans
not change erase the other way of doing this is that we if we can just voice on
the rebuild loss application and recall all the critical paths and put into
production so that we can we can send later time identify where malicious
behaviors have the increments so this is one things and then for a second if we
we can have a need to know where the applications below it then weekend on
the other hand make sure that this application is being actually executed
so so regarding the differences between us and the other branch an aside we we
don’t try to invent a new kind of machine from any language for every
underlying software components we try to may well make all the sauce goes
relatively stronger and relatively trustworthy and then we use yeah
language is to assemble in two bills and transfer see application right so the
difference here is that the client-side applications you want them to be
trustworthy via your separation of power the trial
system and then they can run natively on a smart device or on IOT devices and
then you’re using the blockchain element here to make sure and record the
commission of critical parts to make sure that there is no untrustworthy
execution but you’re not running every single bit of code on something like the
EVM you’re running parts of it so that’s how the whole trial system will work
it’s really for allowing client code to also run exactly exactly so and
something that was kind of interesting when I went read your white paper
what’s the idea that you’re looking for this – one layer so instead of scaling
by layer – you’re trying to build below that layer to make sure that you give
trust to all the nodes on the network exactly so so basically our first layer
also the first component bite and as I explained it contains two layers so so
the bottom layer is actually they make one on there – one and Leia is maybe we
can call it Leo as all the other projects so we did one layer below and
go deep inside the machine so we’d be gonna trust the hardware to bet it on
the machine so we use this hardware to record all the software behavior on that
machine so so most of their behavior actually be consensus parts for the
other project so we actually allow the other proteins they are known to
identify whether their counterparts are running a crack code so if we can make
sure that all joining knows are running crack poems I believe that it might be
much easier for them to build a consensus because most of the knows okay
right it’s all about being fault-tolerant or having malicious nodes
right so the idea is that you’re kind of weeding out malicious nodes before they
even start drawing and sending bad messages one of the concerns that
brought up here is that because Bitcoin you can have multiple
client versions multiple developers you can even develop your own variant of the
Bitcoin by a technical white paper you can write your own so in that case how
do you it’s like kind of white listing a client do your white list clients
because if it that that’s the pot if you do that and it doesn’t that bring on
centralization issues as well so we we ask our community forum in order to tell
whether an application should be put on white list we can actually use a promise
system so if we see a new application and then if we can tell how this
application is implemented from another way of approaching so so if it is a case
then the community will have the capability to identify the potential
behavior often layer one note itself you just like solarization power model they
instead each other is so for example the diverter network is transposing building
transversely computing systems and it actually helps the transit building
system which is the second part so the processing building system actually
builds the applications phony underlying divider network so it’s
just like because the small child or the first friend of sorcery and then allow
them to circulate and the range of this system I don’t know so so it’s kind of
relying on the second layer right the second layer about the client and the
development part so if you can follow the default part and record the
execution of the clients then you can know which nodes are trustworthy exactly
exactly let’s go a little bit deeper into the
use cases so we were kind of talking about how there’s this idea of cloud
computing and extending beyond compiled computing into the triad system so what
are some kind of key industries that you’re
looking at to start with the first adoption of triose so our distance in
China so because our first layer is actually a negative one so so we can
actually help other projects so we can we can we can accelerate at the factory
and Saturday is rayon so we can pack them together to build a construction
chain solution so our current strategies we sell mist and at least construction
chain to the local PCC China so we we currently be about around like 16 16
uses in air of the use cases using our underlying technology to make sure that
their data has not been tampered base or I know it has to be Tendo is and on the
other hand to make sure that unauthorized access to their data as
being recorded so can you give us a particular example of an industry that
you’re targeting and how this can really help them and also why they would choose
to use this over a say a just a traditional centralized database so our
concern is that we use to provide a trusted data layer to solve the local
pcs issue so when when will they are when they are companies find are trying
hard to trust our company’s data we can move their application to approaching
system so we can build an entire ocean fill in by doing so we can actually earn
money we can so we are currently expecting around like 7 million US
dollars revenue this year building these large use case for around 16 16 used
previous cases and we are moving their information system through test systems
one by one one of what is the current roadmap so how ready is this network
right now I was the current status are you a main
net or a test net face yeah so we have finished our first layer the
battle network the first one of the battle network and our testing is
actually up and running so currently we got research team in Peckham University
and research in from for our university they are testing our test net so so
currently isn’t in the internal testing phase and we’re actually planning an
hackathon with targeting at the research group in from from different
universities and our goal is to to release all publicly available testing
it by the end of this month hmm and yeah in q4 Rock you for this
year will be able to release all magnet and at that time I think we’ll be able
to sink in China we are migrating the videos case one by one through our
underlying system so I believe that we have got our magnet up and running we’ll
be able to migrate all those big use cases to our main net and it’s fine
yes six things I think the instant is it’s a little bit of an chicken and egg
problem it’s like if Lily the broken your kitchen has no new spaces how can
we attract developers to come here to give applications yeah so so our
strategy surgeries that we build a large use case for construction chain and we
earn money and then we microwave those chains through our chance
so when minutes on we got quite a lot of large use cases and we will then welcome
already positive to develop application follow stick use cases well then what is
the use case of the token so you got your own try talkin on what is the use
case yes so the try is actually eliminated a
currency of our entire ecosystem so when when like a change in the international
rain wants to build software as a service and just like maybe like like
like letter weeks of vacation like Facebook or the
other kind of CIA movie are key solution so they will need to acquire underlying
trust with a community resource manager while underlying trustworthy software
components so they may need to acquire some other services allow them to
combine its software and hardware to build it computer service so they
need TRY Token to acquire all these underlying new resources and at the
same time they can still write a smart contract to implement this big service and
issue their own token so Trias you guys gonna go to do an IEO and then launching
the mainnet towards the end of this year so if you guys want to know a
little bit more about Trias how can we learn a little bit more. Join our Telegraom and log onto website. I wish you the best of the mainnet deployment and I hope I can follow up with you once
the mainnet is launched thank you so much Dr. Ruan

29 thoughts on “Trias: Can We Trust Machines ?”

  1. Seriously this is an example where a dev is not the best person to explain the technology for the public ! Talk about being lost in mumbo jumbo and this is the state of crypto to be mass adopted , no way

  2. Short answer. People developed the start of a.i. and people can not be trusted so machines can not be trusted.

  3. How much did Trias pay you for this video!? all sponsored video's by law need to disclose amounts payed….

Leave a Reply

Your email address will not be published. Required fields are marked *