The following
diagram shows how, after the first attempts at decentralization in
the 1990’s, the Distributed Ledger Technology has progressed in the
last few years, by using crypto-networks’ blockchain technology and
different consensus models:
Examples
and classification
The following
are examples of blockchain crypto-networks and how their consensus
protocols can be categorized, referring to the two key questions in
the previous diagram:
Who maintains
the integrity of the ledger?
P:
Permissioned: Trusted owners/agents:
Ripple-based (a
group of actors): Tillit Ripple,
Tembusu,
Ripple Lab.
Decentralized
group of trusted agents: Hyperledger, Clearmatics,
Eris.
Bitcoin-inspired
blockchain, oracle-based: MultiChain, Cryptocorp.
Flexible, Proof
of Stake (PoS): Tezos.
Sharding:
Disledger.
Trusted
node-to-node paths: Trustlines.
Federated:
CREDITS,
IOTA
(Cooperative
foundation),
Stellar (Flexible
Trust, quorum).
U:
Unpermissioned:
Untrusted agents (Practical
BFT):
Who
verifies and assembles
a new
block?
O:
One agent,
among a “proven” set, verifies
AND broadcasts
the new
block.
First
generation crypto-networks:
Proof of...
-
Work: Bitcoin, Ethereum and similar: Litecoin, Namecoin, Darkcoin,
Dash, NimbleCoin, and others. -
Stake: (random group of validators): Tendermint, ALGORAND, Peercoin,
Blackcoin, NXT, Ethereum (2018). -
Elapsed Time, in a trusted environment: Intel.
-
Activity (PoW + PoS): No implementation
-
Capacity (Pay to play): Burstcoin.
-
Burn (Pay to play): Slimcoin.
C:
Cooperative:
All
active
nodes
participate
in
block
verification and equalization.
Second
generation crypto-networks:
Majority
agreement:
MARPLE
(By
Gorbyte
Inc.,
Heuristic,
Logical Environs).
Note
that many
classic fault tolerant solutions (BFT, XFT, non-BFT),
such as Paxos,
Raft,
etc., and Directed Acyclic Graph solutions, such as IOTA
Tangle,
and Dagcoin,
are not designed for, or
implemented on,
blockchain technology.
MARPLE,
by
Gorbyte Inc.TM
MARPLE,
a new cooperative consensus protocol, is the key component of
Gorbyte, a
second generation crypto-network.
It
distinguishes itself in the decentralization aspect of block assembly
and equalization.
Since all nodes are already broadcast every transaction, MARPLE reaches agreement by equalizing blocks within Logical Environs.
Gorbyte’s security against DoS and majority attacks is also enhanced by its BRUD device architecture, allowing the detection and prevention of such attacks.
Since all nodes are already broadcast every transaction, MARPLE reaches agreement by equalizing blocks within Logical Environs.
Gorbyte’s security against DoS and majority attacks is also enhanced by its BRUD device architecture, allowing the detection and prevention of such attacks.
This
combination
allows
for a
higher throughput (from current
crypto-networks
3-15 tps to 200-6000 tps), lower cost, a free service model, and
scalability: every user device can participate as a full network
node.
Properties
of consensus (MARPLE):
-
Termination: Every correct active node eventually decides.
-
Convergence: A timeout can be set so that an arbitrary percentage of the correct active nodes terminate within the allotted block period.
-
Agreement: No two correct active nodes decide different values.
-
Validity: If the required majority threshold propose the same block B, then no value different from B can be decided.
Replication
requirements (Gorbyte):
-
Each node must commit to the same block within a certain time (synchronous,
no forks). -
In most cases, transactions are broadcast only once. In the remaining cases only minimum exchanges are required, within a node environs.
MARPLE’s
restrictions
and
assumptions:
-
Randomization of sessions: Sessions established by a node at startup time connect to physically dispersed responding nodes (ASML protocol).
-
Environs optimization: At node startup time, a minimum and maximum length of loops are allowed with network peers (FID protocol).
-
Order restriction: Predefined rules for block normalization. This is essentially a restriction on the input values of the consensus process.
-
Convergence timeout: Parametrically adjusted for an arbitrary minimum threshold of agreement. Late, disagreeable, or faulty nodes have to re-initialize.
-
Synchrony assumptions: all nodes maintain the same timedate. The block collection period is predefined and synchronized.
-
Cooperative assumption: Most nodes are cooperative, as they contribute their device’s resources to use free services. An attacker cannot replicate multiple nodes, because of the uniqueness imposed by the BRUD (virtual) device associated to each node.
MARPLE
Attributes
MARPLE (Majority
Agreement Recursive Protocol based on Logical Environs) is
characterized by the following attributes:
-
Deterministic: Consensus is irreversible – no forks are possible.
-
Synchronous period: Nodes and block periods are synchronized, to increase the probability of similar blocks and to enforce time-outs.
-
Verification - Consensus independence: Transaction processing and verification are independent from the consensus process and block composition.
-
Decentralized: There is no master node, no set of miners, verifiers or generals, but each node participates in the consensus.
-
Heuristic: Each node cooperates with a random set of peer nodes.
-
Cooperative: Consensus is achieved by majority agreement, not by a set of competing agents (miners or verifiers).
-
Unpermissioned: Nodes do not need to be given special permissions by an authority.
-
High throughput: Most transactions are broadcast only once. Blocks are never broadcast. Most protocol messages are within node’s environs, and happen in parallel.
Reference
List of Networks Using Original Consensus Protocols
Abstract
(Azyzzyva,
Aliph,
R-Aliph, ZLight)
https://infoscience.epfl.ch/record/208030/files/The_next_700bft_prot_a12-aublin.pdf
https://infoscience.epfl.ch/record/208030/files/The_next_700bft_prot_a12-aublin.pdf
ALGORAND
https://arxiv.org/pdf/1607.01341.pdf
https://arxiv.org/pdf/1607.01341.pdf
CREDITS
https://credits.com/
https://credits.com/
IOTA
Tangle
https://iota.org/IOTA_Whitepaper.pdf
https://iota.org/IOTA_Whitepaper.pdf
Paxos
https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/The-Part-Time-Parliament.pdf
https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/The-Part-Time-Parliament.pdf