Means of State Replication

The means of developing replicated state machines.

Decentralized and permissionless state replication as well as distributed application specific state replication are defining the next era of computing. Replication of data on distributed systems using leaders and workers, distributed micro-services and containers, peer-to-peer blockchain protocols, UTXOs, and digitally enforceable agreements and agents. Blockchain is the next generation database architecture for trust minimized computing.

A database needs to do two things:

When you give it some data, it should store the data, and when you ask it again later (via a query language); it should give the data back to you.

Data Models: the Format in which you give the database your data.

  • SQL
  • key-value
  • NoSQL
  • Graph

Query Language: The mechanism by which you can ask for it again later.

How does network replicate it’s state so that:

  • the graph grows
  • it determines validity
  • it determines order, if needed
  • the history is shared
  • economically incentivized
  • logic is shareable, upgradeable

Trust is the key feature of blockchain protocols;

Because of this the protocol can mutate and replicate and incentivize. Is the process of mutation ever seemingly finished, code is not / never static. Why is this so accelerated? We are observing mutations on a much faster timeline and scale then we could imagine. The growth and replication of the networks are:

Replication of state is the means of which a blockchain can produce a copy of itself, with it’s governance, it’s liveness. A copy does not need to be produced in order to write. A net new copy needs to be produced in order for a mutation to take place. Forking a blockchain is by nature is change for:

Linked blocks and linked transactions are what a blockchain is comprised of, state is replicated and address space is shared.

Protocols change and improve. Data is linked. Means of replications to reproduce a series of state. Are replicated state machines necessary in order to achieve consensus. Why is there a need for the control back. What is it that is being unveiled and that we are converging over.

Why do we trade storage space for address state?

The distance between the space and the state changes? What is the difference between on-chain state and cold key-values static, dormant. One replicates state in order to achieve consensus.

Stateful databases have a persistence Layer. Does it matter what type of database is used to determine if it is a blockchain?

Lisk: mySQL, Corda: H2 vs. Key-Value ( Parity: Rocks DB,  Geth: Level DB)

To store and replicate state, blockchains have to have a persistence layer. The consensus mechanism and state are separated with blockchains.

If state isn’t replicated across all nodes, is it still considered to be a blockchain

  • OLTP – optimized for transaction processing
  • OLAP – optimized for analytics
  • OLTP
    • Log-structured – only permits appending to files and deleting obsolete files
    • But never updates a file that has been written.
      • Bitcask, SSTables, LSM-Trees, LevelDB, Cassandra, Hbase, Lucene
  • Update-in-place – treats the disk as a set of fixes-size pages that can be overwritten.
    • B-Trees
  • Data warehouses
    • Becomes important to encode data very compactly
      • To minimize the amount of data that the query needs
  • Column-oriented storage helps achieve this goal

Consensus

What types of decentralized transaction ledgers have the security guarantees of open replicated systems state across every node?

Means of transaction:

  • UTXOs
  • Accounts

Means of replication:

  • DAGS (Directed Acrylic Graphs)
  • Merkle Trees
  • CRDTs (Conflict-Free Replicated Datatypes)

Means of consensus; the consensus algorithm.

  • Asynchronous Byzantine Fault Tolerant
  • Byzantine Fault Tolerant (BFT)
  • Fault Tolerance (Raft)
  • Practical Byzantine Fault Tolerance (PBFT)
  • Proof-of-Work
  • Proof-of-Stake
  • Delegated Proof-of-Stake (dPOS)

avalanche

Data structures are replicated as state machines that are crypto-economic incentivized. Is there a more secure way to create that replicated state machines with no native token?

Confidence score creates, a determinism for state shared on a Merke DAG. If below certain value that the state flips. RSMs (Replicated state machines) are traversable, many-to-many state structures. Unconsumed state transactions that are replicated on isolated merkle dags. avalance DAGS are live replicated state graphs.

Distributed, Permissioned, and Application Specific

We can think of the ledger from each node’s point of view as the set of all the current (i.e. non-historic) states that it is aware of. State can be a shared or unshared. State can be an agreed upon fact. The fact is shared, the state is shared. Historically, a record of agreed about facts can be determined as agreed upon historically shared state. “This sequence of state replacements gives us a full view of the evolution of the shared fact over time.” UTXOs that are linked as historically state/facts.

Validity

Is a transaction valid or invalid. Is a transaction that is invalid need to be included in the ledger.

Atomicity

“a transaction is just a proposal to update the ledger. It represents the future state of the ledger that is desired by the transaction builder(s):”

 

  • The transaction’s inputs are marked as historic, and cannot be used in any future transactions
  • The transaction’s outputs become part of the current state of the ledger

 

“Outputs need to be valid to produce new inputs; to create liveness, layer historical state. Each transaction consumes a set of existing states, to produce a set of new states.”

Order

Is order of importance? Is the order of valid transactional state changes a prerequisite to achieve consensus. Is who owns the fact, the state to be shared of importance? Gossip Graphs; is it of importance who told the fact, where the fact came from?

 

Access

Doorman / Certificate Authority. How do we let you into the network.  How do we get you up to speed on what has happened, or do we not need to for you to join?

Logic across data structures;

The scope is about being able to program logic that is replicated across permissionless state machines.

Leave a Reply