The proposed new relay protocol can reduce the “transaction bandwidth” used by bitcoin nodes of up to 75%.
Called Erlay, the proposed protocol changes the way transactions are delivered so that they use significantly less bandwidth, an important resource for nodes that make up the network. The authors include the University of British Columbia researcher Gleb Naumenko as well as two heavy development bitcoin weights: Greg Maxwell and Peter Will.
The bitcoin way works in this nodes around the world and tie together to form a network. Under the hood, after a transaction is broadcast, it rolls through this vast network of hardware.
Erlay changes how to declare these transactions. As Naumenko described in the email dev bitcoin announcing the new proposal:
“The main idea is that instead of announcing each transaction to each colleague, the messages are sent only over a small number of connections (only 8.) Another relay is achieved by periodically running a reconciliation protocol defined for each connection between groups of experienced messages in both directions.”
The results, according to Naumenko: “We are saving half of the node bandwidth consumes, enable almost free connectivity, and side effect, better withstand timing attacks.If the count of outgoing colleagues has risen to 32, Erlay saves a total bandwidth of 75% compared to the current protocol “He said.
One important result of this new protocol, researchers argue, is that by reducing bandwidth it takes a process, nodes can increase the number of connections they hold with other nodes.
As low and technical as it sounds, this is an important study, especially when it comes to the security of the network itself.
Bitcoin security depends at least partly on connections between the nodes. This new protocol can make room for more connections, and the more connected the node is, the more “tough” it is against network attacks.
Naumenko described one such attack to CoinDesk: “The simplest example is an Eclipse attack when the destination node gets isolated from the longest network because all its connections are established with the attacker.” In this case, an attacker, for example, can cause a target node to believe that they have paid The target node (show a short chain with it [deal] b), without actually submitting transactions to the longest network.
How this attack can affect bitcoin is described in more detail in the research study 2015.
So, if the protocol is so important to bitcoin security, what next? Will it be added to Bitcoin Core, the most popular bitcoin software application?
“A few weeks ago I talked with some contributors of Bitcoin Core and feedback was generally positive, although they asked for more experiments.” Now, as these experiments are added, I would give more time for everyone to get to know the new technical bits, “Naumenko told CoinDesk.
In general, new technology is not added to bitcoin unless the most active contributors to the software, as well as the broader ecosystem that operates the nodes (and unlike the reactors, do not accept any kind of built-in subsidy or compensation) agree.
“We have received positive signals from the community, which encourages us to continue working on implementation,” he added. If the community continues to love it, then: “The protocol should be part of one of the main liberators in the future (I hope, the next).”