Get notified about Tamas Blummer

Sign up to a email when Tamas Blummer publishes a new video

so welcome everybody this gentleman you
likely heard the buzzer on the
blockchain that is the popular name of
the technology behind bitcoin it might
look like a hype but I can assure you it
is not the technology is here to stay it
systems that are auditable and executing
business processes in a trust minimised
manner that is a bridge that is precious
especially when interacting with
corporations trading partners using
distributed Ledger's built on the
blockchain technology so what does it
promis-- I'd like to point you in this
presentation towards the technological
reasons of interest and that are in my
opinion auditability network wide
consensus privacy and compliance and
trust minimize execution finally smart
contracts the blockchain is a data
structure built for provable consistency
the consistency in a database
consistently check on this database is
cheap to prove however it is flagged on
different levels if somebody would
attempt to temper it tampering with it
is practically infeasible let's review
the database structures from the button
transaction log it substitutes the
ledger if you aggregate the transactions
by the owner of accounts which is the
owner of the outputs on the lowest level
we have transactions that relocate their
input to a set of outputs the
transactions are identified by their
cryptographic hash a hash function is
mathematically a trapdoor function it is
easy to calculate in one direction that
is the data to hash but it is quite
infeasible to do the reverse that is to
find the data that would produce the
or find an alternate data that would
transactions are collected then into
blocks blocks away so hashed with the
same cryptographic function to define
their identity and the blocks in the
block refers to the previous block using
this ID the references define the chain
of blocks and hence the name blockchain
since the order of transactions
influences the blocks hash their order
is also fixed fixed within the same
block and the global order of
transaction is given by the chain of
locks it is apparent that adding
altering or removing transactions would
change the hash of that block that is it
is contained in and with that with the
hash of all subsequent box the entire
chain would have to be recomputed from
the point of modification to the latest
lock to present a new consistent state
by adding further constraint on the
block hash they can imply a significant
work that needs to be performed or
blocks my might be required to have a
signature of early data or so the
verification is undeniable and the
transactions are hashed into the blocks
identity by ordering them into a binary
tree called a marquetry the value of
every level of this tree is the hash of
hashes on the lowest lower level this
structure offers a significant advantage
to audit since the existence of a
transaction can be shown in a compact
proof including the transaction and only
the aggregated hashes that entered the
three paths to the transaction the
compact proof saves bandwidth and also
allows for pruning of history for
transactions that are no longer relevant
since all the output is spent the
advantage against ordinary deletion in a
database is that an approved merkel tree
it preserves mid-states of the block ID
calculation hence it still proves that
oh transactions remained in the block
we're already there as the block was
a prune block is just as strong provo
proving evidence or existence of a
transaction as a full block this proof
of existence were is very exciting
applications a transaction may hold a
reference to external data and prove its
existence in a compact manner referring
to the blockchain since blocks have a
timestamp that is always a part of their
ID the transaction would prove the
existence of some external data at some
code some past time point note however
that only a fool block history can can
prove non-existence of a transaction to
a certain time point another very
important property of the technology is
a network wide consensus block chain is
typically not maintained by a single
server but it is held in identical
copies by nodes participating in a
network this makes the ledger is that it
substantiates the distributed ledger
which is the primary interest of
overwork a digital asset and achieving a
network-wide replication was trivial if
the nodes were trustworthy operated by
realistic assumption if especially if
they do replication across companies
execution on the other hand on the other
end cannot be fully trusted by because
every single actor could be hacked such
a failure should however not have a
network-wide effect the problem we face
here is what is known in mathematics the
Byzantine generous problem that is
roughly hard to achieve consensus in a
network in absence of trust between
actors there is a limit to for tolerance
that can be seen very easily with the
triangle chart on this sheet that shows
you again general with two lightens and
then the message communicated to the if
the general is folded in the message
communicated lightens is contradicting
and there is no no way to
solve this problem until the proportion
of food if the food proportion of 40
actors is below 12 at one-third while
above result is generally walid naku
satoshi nakamoto proposed an approach
that works well in practice adding an
assumption that mathematicians usually
avoid and that assumption is that there
is a computation of work without any
shortcuts that nobody can reduce and now
if we require that every actor who wants
to vote on a validity of a block
performs that work then we can compute a
majority opinion of the network a
brings the threshold for the honest
majority to one half instead of
one-third again measured with computer
computing capacity why both approaches
approaches the byzantine for tolerance
algorithms and Anaka moto consensus our
wallet for a practical solution there is
a trade-off between trust and computing
power the network with known actors
could be operated cheaply and
efficiently assuming less than one third
of four reactors by the network video de
sump shins can use proof of work at less
than one half of computing capacity
being dishonest the combination of both
approaches is a current subject of
research and product development
operating a shared ledger across
companies is obviously superior Aryan
performance if we compare with
proprietary systems connected wired
typical batch processes and
reconciliations having a shared ledger
however raises the concern of privacy
between trading partners the block chain
offers here a new security model the
solution is to keep identities that
engage in a transactions secret not
reusing Keys if implementing this
individual transactions reveal nothing
about trading predator trading partners
and not note the emphasis on individual
transactions sophisticated data mining
algorithms Harbor could cluster the
activity and gain some insight by time
especially if correlating the evidences
with our a bit correlating with
evidences available outside of the
blockchain for heightened alert privacy
there for so-called zero knowledge
proofs could be applied zero knowledge
grew in a zippy the zero no pude lodish
proof the transaction is partially
encrypted this disallows part is not
involved the use the economics of
transaction but they can still convince
themselves that the transaction is in
fact wallet I do its content is
partially unknown there might however
also be legitimate interest in trading
activity by authorized party such as
regulators arbitrators early or legal
actions these users can be addressed for
example by deterministic key generation
that allow across the ring of seemingly
unrelated in identities on the
blockchain the homomorphic property of
the elliptic curve key generation
suppose deterministic in identity
generation without inside to the
this can provide read-only view vial
ensuring that transaction can still only
be initiated by the owners of the
identities accounts of a distributed
ledger can also be owned jointly this
allows for support of arbitration or
even automated transfer to repo da shin
account note that the transaction
validation is executed by all nodes
autonomously and why they are update
their own copy of the ledger this means
that those validation algorithms have to
be executed with identical outcome on
our notes so they are database features
the enemy as animated in a consistency
check to remain in sync with others
nodes may not deviate from the rules
without losing consistency on key and
especially not validate this rule
case-by-case basis the described
execution environment is unique since no
one knows and the network is able to
override the rules therefore the trust
into a proper execution of the nodes is
minimized revising removing or adding a
transaction to the common history
quickly becomes infeasible even for a
bigger many minority of 40 notes this
new paradigm assay of execution is what
we call trust minimized and its
application is countless we are just
starting to explore this huge solution
space a network-wide consensus building
could be also too expensive for for
example high-frequency trading for
scenarios we are trading partners
repeatedly trade with each other we may
introduce netting channels they consist
of offsetting transactions that are
wallet to be connect committed to the
blockchain at the discretion of the
trading partners at any time they
however doesn't have to instead the
offsetting transactions could be updated
to reflect the new allocation of funds
on regular basis and then flush to the
ledger once the appropriate once it is
appropriate by a higher level business
function the challenge here is to
establish a protocol whereby the none of
the trading partners gains an advantage
by not following the rules or not
following through the protocol to its
end finally validation rules observed
that validation rules for a transaction
may include more than just signature
validation for the sources it can
require extemp ille several signatures
it can use a set of signatures out of a
set of joint owner the evolution rules
my avoid in the context of the
evaluation from a transaction to a block
there by introducing time-dependent
validity rules such rules an able for
example automated escrow agents even
external sources could provide into the
algorithms could be provided to the
so validation decisions could be
influenced by them note that however
introducing certain records the reduces
the trust into the source of that
information now the properties I
enumerated are really enormous we
observe the birth of a new technology
that might have first sight they are
technically complex and wasteful but if
compared to the opportunities it unlocks
an area that is really verte of
exploration let me finish with a quote
from our CEO blood masters who said you
should be taking this technology as
seriously as you should have been taking
the development of the internet in the
early 90s thank you i would agree
inventions yeah so do we have any
questions refinished quite early so you
have time yes can you cover some
examples of you know some grill aight
sounds where you could use this
technology i mentioned this presentation
for example proof of existence since the
blockchain creates an auditable heart
neil practically infeasible to
manipulate structure it is easy to
commit a document or an existence of
antecedent existence proof of a document
into the blockchain and this I I think
alone this idea could revolutionize a
whole range of current business
processes starting from patterns to even
eventually like registries yeah what we
currently work on our is settlement
processors settlement of financial
instrument a violent financial contracts
the blockchain itself is a settlement
system so it's very obvious to use for
that purpose
well I have a question yes two nodes
want to commit the transaction there are
different transactions right we
concluded reporters how do they how do
those things get sorted in the
blockchain how do you define you which
one goes first don't they both the
auction at the same time is there are
there any mechanism by the first
question is whether the true
transactions are are really related it
means whether one is spending the other
if they are unrelated then they could
hit the block chain boat and could hit
in any order since they are on rated it
doesn't really matter it so the block
chain defines an order of transactions
which is at hawk if they are basically
at houck if they are unrelated
transactions what is important is
disorder the transactions if they are
related because only if you order that
if you fix globally fixed order it for
transactions can you decide whether a
transaction was before another and means
whether the next transaction is an
attempt to double spend some previews a
transaction so I always thought that
when you have a transaction you have to
basically compute the hash of the entire
block to be able planet right wasn't
that well su if you compile a block then
you hash the transaction and this hash
enters through the market marquetry into
the hash of the block this is this is
of the Bitcoin network so this is the
process to include a transaction into
the block but this process has have only
have to be done once but checking that
is very cheap and can be done by anyone
else with consuming the blockchain
I'm greeting that you're doing something
called high molecular hyper ledger is a
company that was bought by digital asset
just like my company which was bits of
proof it's a different story I was not
I'm in the in the mean why certainly I'm
looking at at what they did but I'm I'm
the responsible for building the
distributed ledger and the technology
layer of the company and certainly the
ideas that we saw with hyper lecturers
are entering these considerations okay I
just wondering work like for this
division for high village and I was also
noting that it says that it doesn't have
a currency doesn't have a cryptocurrency
and I was kind of interested in that
choice well hyper Ledger was meant to be
a an implementation of Byzantine for
tolerant algorithm yeah this is an
alternative as I mentioned there are
there are several alternatives how you
achieve network wide consensus
practically be a Byzantine for tolerance
which they did is one of the
alternatives the question of having
having a current saving a token is
completely unrelated to this question of
how you build a consensus on a ledger in
case of Bitcoin that currency also is
used to incent device that work but
that's a different story then in high
quality how you intensifies the work if
you don't have a good well it's not the
question is not hyper ledger it's a
business interests to operate a
distributed ledger so there is an
interest to settle transactions that
that alone is sufficient to work with
the system you do not have to in it
issue a token to incentivize the this
because it's already incentivize by
doing that economic activity on a more
efficient laminar
the damn network is only operated by one
company can you trust the network know
that that's not the point it's we do not
think that these distributed Ledger's
would be operated by one company that
distributed ledger we are developing
software based on bits of proof my
previous software and i pologize ideas
and this software meant to be run by
different entities eventually be even
compatible or interoperable with
software developed by other companies
the point is we are creating a protocol
on a very similar to what bitcoin did
creating a protocol to achieve consensus
on a ledger and to maintain a structure
a blockchain structure with that very
high level of consistency properties and
auditable properties that I just
do you do you think it's going to be a
private network more in well I think
there is place for both so bitcoin is a
network that operates in an in a very
trust less environment so I mean where
you cannot trust environment that's
that's why it needs a huge computing
capacity to achieve consensus with proof
of work however if you would operate a
similar network of companies who know
each other and have a certain assumption
on the plausibility of collusion against
each other there you could operate a
network which is less depending on a
proof-of-work algorithm but more
depending on a Byzantine full Toronto
application how much the difference so
in terms of the we all know that you can
produce a lot of computing power right
now because it is operating this
environment what if what if we go to the
other environment where we have
companies that kind of trust each other
but they
might not he always hundred percent sure
they will be competing power that you
get it well I do not assume that the
private network would achieve consensus
by pre for proof of work at this point I
would assume that the private network
photo and replication proof of work
might also play a role if you consider
that if the blocks are in addition
protected with some level of proof of
work that also adds a very objective
level of security since consider that if
somebody would want to forge a ledger
usually do not you do not have like an
unlimited time to do this forgery but
you have a short level short time period
and through the aggregator through the
aggregation feature of the blockchain
that aggregates the work spanned and
you probably try to forge it means that
there is an unbalanced on how much is
therefore needed to forge and how much
is their fault needed to protect so I
assume that this kind of private
networks will use both technologies with
time and we are currently exploring what
is possible there and it so any more
questions if not let's just lift today
thank you thank you very much