Ripple’s distributed consensus mechanism does not serve a clear purpose. This is because rippled nodes are set by default to hand full control over updating ledger to the website server. This is a conclusion by Bitmex Research in their current analysis of the Ripple network.
BitMEX to explore the technology behind Ripple
The team reached this conclusion by installing a copy of the Rippled network last month, for the purpose of their investigations. The node operated by downloading five public keys from the server. Their analysis finds that four out of five keys are needed to support a proposal for it to go through.
“Since the keys were all downloaded from the Ripple.com server, Ripple is essentially in complete control of moving the ledger forward, so one could say that the system is centralised.”
The Ripple story
We look over the history of Ripple and examine various disputes between the founders. We then explore the technology behind Ripple & conclude that by default, https://t.co/7ez1pRYva2's server has full control over the ledgerhttps://t.co/uU0cak1Y7t pic.twitter.com/uyGJ8hjYsw
— BitMEX Research (@BitMEXResearch) February 6, 2018
It is misleading for Ripple to argue that its system was distributed
Their node showed that the keys would expire a few days later. It means the node would have needed to go back to the server to download a new set. Although there is nothing wrong with centralization, the team says it is wrong and misleading for Ripple to argue that its system was distributed.
“In addition to the potentially misleading marketing, the construction involving the quorum process. And 80% threshold is not necessary and merely adds to the obfuscation, in our view.”
However, the team could argue that the keys are editable and customizable, which is true. There is no evidence that many of its users to change the configuration file. Even if users would edit the files, there is no reason to assume that the “system would converge on one ledger”. For example, one user could connect to five validators and another user to five different validators. The two would meet the 80 percent thresholds but for two conflicting ledgers.
“The 80% quorum threshold from a group of servers has no convergent or consensus properties at all, as far as we can tell. Therefore, we consider this consensus process as unnecessary and essentially pointless.”
The authors note that Ripple is missing 32,000 to 33,000 blocks from the start of the ledger and nodes cannot obtain this data. The whole chain can therefore not be audited.
Ripple network has been popular with facilitating cross-border transactions among banks and other financial institutions. Bitmex says the network works like Lightning Network, except for the fact that it has counterparty risks avoided in the Lightning Network.
Bitmex argues that the model of the network is “likely to be unstable”. The trust network “unlikely to be regarded as reliable.”
The research says,
“Either the system would centralize towards a few large banks and fail to be sufficiently different to the existing financial system. Or it would be liable to regular defaults. However, the current Ripple system is very different to this original idea.”
It continues to explore Ripple history noting that disagreement between Chris Larsen and Jed McCaleb was the reason McCaleb left the company.
Here is a video, showing Ripple architecture after McCaleb had joined
Ripple past issues and complications
One Ripple investor wrote about the situation sometimes in 2014,
“Since Jed’s departure, the management of the company has taken a different direction. Sadly, the vision Jed and I had for the project in the early days has been lost. I’m no longer confident in the management nor the company’s ability to recover from the founders’ perplexing allocation to themselves of 20% of the XRP, which I had hoped until recently would be returned. Prior to Jed’s departure from Ripple, I had asked the founders to return their XRP to the company. Jed agreed but Chris [Larsen] declined — leaving a stalemate. This afternoon, I revisited the allocation discussion with the pair and again, where Jed was open, Chris was hostile.”
Other past problems with other companies include R3 suing Ripple in September 2017. R3 argued that Ripple agreed to give it the option of buying 5 billion XRP at an exercise price of $0.0085 before September 2019. R3 said Ripple terminated this deal despite having no right to do so.
Ripple filed a counter case arguing that R3 did not honor part of its agreement. The agreement was, according to Ripple, R3 introducing Ripple to a large number of banking clients or promoting XRP for use in these banking systems. This case still remained unsolved.