A detailed operational understanding of blockchain is not necessary to
follow the work in this paper, but an overview is provided here.
Each participant in a blockchain (each node) keeps a copy of all the historical transactions
that have been added to the ledger, and by comparing to the other nodes’ copies, each is
kept synchronized through a consensus process. Unlike in a traditional ledger system, there
is no node with special rights to edit or delete transactions – in fact there is no central party at all, which is one of the reasons that blockchains can be useful when a trusted central party is either unavailable or too expensive.
The idea of having a ledger that propagates to all its participants has been around for some
time, but there were some serious barriers to overcome. The most important of these was
the ordering of transactions and the ‘double spend problem’. In a large network, transactions are broadcast constantly from different nodes, and those transactions will take varying amounts of time to reach different points of the network. Hence, it is difficult to have a definitive order of transactions – especially important if two transactions attempt to credit the same resource, leading to double-spending and two parties disagreeing on who has the right to a given asset.
Blockchain solves this problem by having newly broadcast transactions go, not directly onto
the ledger, but into a holding space. These transactions are periodically bundled together
into a block, which is then officially considered to have posted with a simultaneous timestamp, thereby propagating the transaction to all users. In order to prevent blocks from clashing, and to avoid the need for a central authority to do the block-making process, blockchains use various methods to impede the process of making (‘mining’) new blocks. The process for bitcoin, for example, automatically adjusts the difficulty of the process (which involves complex mathematics), so that on average a new block is formed every 10 minutes. Different nodes compete to solve these mathematical problems, so no central party controls the process. Successful mining is rewarded with new bitcoins and a transaction fee. A system that only included trusted parties of known identity can simplify this by reducing the amount of verification that is needed for each transaction.
verification that is needed for each transaction.
So that’s the ‘block’ part covered – what about the chain? Well, new blocks don’t just
contain the list of approved transactions, they also contain the timestamp of the block, and
the hash – a unique cryptographic signature – of the previous block. Because the block
references the immediately preceding block, its order in the chain is unambiguous. What’s
more, an attempt to change a previous block would be immediately obvious, as the hash
signature would change and no longer match the backward reference in the following block.
Consequently, changing something in a blockchain after the fact is not viable and blockchain
records are permanent.
Some blockchains are set up to contain, not only details of transactions and ownership, but
also executable programming code. Parties can agree to add code to a blockchain in the form of a smart contract, that is, code that will carry out agreed transactions when triggered. This allows for automation of new transactions, and allows some blockchains to
be programmable.
Comentarios