Byzantine General Problem
·
Due to the decentralized nature of the blockchain,
o
it becomes difficult to reach conclusions regarding
certain activities on the network.
·
For the computation of any data that the blockchain is processing,
o
there needs to be a common conclusion regarding
the state of data among all participants.
·
This conclusion, known as consensus,
o
is one of the most important parts of
the blockchain architecture.
·
However, due to its inherent distributed nature,
o
Participants on the blockchain will
be functioning according to what incentivizes them.
· The
decisions they take –
o
will also be based on the various kinds of information
that they have on hand.
·
To ensure the frictionless functioning of the
blockchain in a network,
o
where there are multiple distinct actors, consensus
needs to be reached consistently.
This
is done by solving a long-standing
problem known as
the Byzantine Generals Problem.
·
The Byzantine Generals Problem is a term used in computing to denote a situation:
o wherein certain components of a system may
fail if participants don’t agree on a ‘joint
strategy’ to deal with the problem.
·
The problem assumes that some of the participants are corrupt, spreading misinformation or unreliable in some way.
The problem can also be described as such.
·
If there survives a city
o
which can only be taken over only by the concerted attack of
multiple armies,
o then the armies are
required to collaborate and attack the city.
·
However, in a situation where there are an odd number of
armies, common consensus must be reached
o
among all the participants in order to
successfully mount the attack.
·
However, certain generals of some armies choose to disagree on attacking the
city,
o
leading to a critical system failure.
The required task is not undertaken due to
disagreement by unreliable actors.
·
This failure is known as a Byzantine Fault,
o and represents a
situation in computing where it is unclear whether a component in a network is
working properly or not.
·
This is generally extended to blockchains as each node is
considered a ‘general’
who contributes to the consensus of a network.
·
There are multiple solutions that were proposed
before the implementation of a true Byzantine Fault-tolerant consensus
mechanism in the form of Bitcoin’s Proof-of-Work algorithm.
·
One of the first BFT systems is only possible if the
number of malicious actors is not more than 1/3rd of the total number of actors
on the network.
·
Another BFT system also adds unforgettable message signatures
to ensure that messaging is verified. This can be conducted through public-private
key encryption.
This led to the creation of many BFT
algorithms such as the Practical BFT (pBFT) algorithm
proposed by Miguel Castro and Barbara Liskov in 1999.
The pBFT algorithm created multiple other BFT
algorithms that improved on it in many ways.