Intercoin Technology: Background

With credit-based money, A can owe B, but can also owe other people as well. B may not know all the creditors of A, and takes a risk accepting A’s debt. (The original Ripplepay service by Ryan Fugger was based around Trustlines.)

With value-based money, A can pay B, and a third party ledger keeps track of information that lets B know whether A actually owns the money it’s paying, and updates to keep track of who owns what. Typically, this ledger was maintained by a centralized third party, like a bank or a social network. In fact, in 2018 this is still the dominant paradigm, as WeChat, Facebook, and others maintain centralized server farms to process payments on their own giant networks.

The crypto currency revolution has been about decentralizing this third party ledger. One of the biggest challenges with that is the double-spend problem. A bad actor, Mallory, could legitimately own a token T at some point in time, and transfer it to Alice, but then later “forget” about it and also transfer it to Bob. Alice and Bob need to make sure that the third party ledger knows that Alice now owns T, and is able to furnish this information to anyone who asks. A decentralized third party ledger consists of many different actors, and we need a scheme that makes it very hard to “fork” the history of a token.

Bitcoin solves this problem with “proof of work”. All transactions go on one, giant, ever-growing ledger. Moreover, tokens can split, forming an ever-growing set of Unspent Transaction Outputs (UTXOs). The system is secured by an ever-growing hash rate of machines competing to be the next miner. Each transaction in the world needs to be transmitted to this miner in order to get into the next block.

Algorithms based on proof-of-work, proof-of-stake, or something similar are called leader-based consensus, where one leader (e.g. a bitcoin “miner”) validates all the transactions for the next block of the blockchain. They typically exhibit massive scalability problems: because one blockchain is used for all transactions on the network, thus all transactions have to go through this leader (this is the root cause of all the bottlenecks). If the leader is known in advance, then it can be attacked, bribed, etc.. If the leader is not known in advance (such as with bitcoin), then every transaction in the world has to be sent to every potential leader, which can quickly get unwieldy.

It’s hard to architect a scalable payment system around leader-based consensus, which is why 10 years after Bitcoin has been introduced, crypto is still not being used for everyday payments. If 1 million people buy coffee at the same time, this will cause 1 million transactions to be put into one global ledger.

The centralized solutions at WeChat, Facebook, etc. are just much more faster and scalable. A new approach is needed to build an open and byzantine-fault-tolerant system that can scale while maintaining the trust people and institutions need in order to transact confidently. That’s why Intercoin is designed differently. It is designed to be embarassingly parallel so as to support a nearly unlimited amount of transactions per second.

Next Topics:
Intercoin Technology: Architecture
Intercoin Technology: The Tokens
Intercoin Technology: The Ledger
Intercoin Technology: Consensus
Intercoin Technology: Mitigation
Intercoin Technology: Recovery


What methods do you plan to utilize to better solve the issue of transaction speeds as they currently present themselves and the possibility of even higher speeds in the future?


Tokens don’t split into UTXOs - in Bitcoin, only the UTXOs are represented, and bitcoin balances are a sum of UTXOs in a given wallet. 1 bitcoin could be represented by 1 or more UTXOs. UTXOs can be consolidated, as has been happening in early 2018. It seems that higher prices and more usage will increase the size of the UTXOs set, but not necessarily beyond what many current nodes can handle.
Also, a bitcoin miner doesn’t have to include transactions in blocks (but it is good for the network to include the transactions, as well as adding the miner fees.
There are actually multiple scaling problems. See this article by Nick Szabo:

1 Like