Bitcoin still relies on the internet as rails to reach consensus and is thus susceptible to many traditional chokepoints. Reachable peers that are able to both send and receive connections from the Bitcoin blockchain number only slightly over 10,000. These nodes are capable of fully validating the Bitcoin Blockchain and are a vital check on miners.Mining Centralization
While miners' interests are mostly aligned with users, their move towards industrial scale in areas of cheap electricity makes anonymity difficult with most relying on electricity off government grids. The inherent risk of censored transactions is only ensured against by having a robust and decentralized network of full nodes.
It is a misconception that full nodes are the only non-mining nodes. The vast majority of nodes are unreachable nodes, including mobile nodes, those behind firewalls like Tor or even those shielded by Network Address Translation. While these nodes are less discussed, they represent the majority of the network.
A 2018 study, Towards Better Understanding of Bitcoin Unreachable Peers, finds Bitcoin unreachable nodes are highly centralized. If users elect to forego a full node, they are sacrificing security not just for themselves, but also for the entire network.
The data demonstrate the overwhelming majority of Bitcoin is used through mobile wallets with the majority running on software from only two wallet providers, BitcoinWallet & BreadWallet, providing for 59.9% and 20% of clients respectively. This problem likely is, in part, a result of the prevalence of cryptocurrency purchases over mobile, making a mobile wallet the most obvious destination following a purchase.
To novice users, there are no apparent direct economic benefits to running a full node. Instead, they elect to use wallets that ping between three and six other random nodes to get up to date with the network. Full nodes meanwhile keep a full history of the network dating back to the genesis block.
In controlling nearly 80% of the share of unreachable nodes on the network collectively, these two wallet providers are in charge of a daunting amount of transaction propagation to the point they can be considered critical to Bitcoin's infrastructure. Institutions within Bitcoin should under no circumstances be entrusted to provide the infrastructure for eighty percent of non-listening nodes. When weighing the share of Bitcoiners using mobile, the benefits of decreasing block size in order to implement mobile full nodes becomes even more paramount.Mobile Routing
Mobile wallets connect through telecom companies, meaning the routing falls on a limited number of service providers - namely T-Mobile, Verizon, Comcast, ATT, and Rogers. These traditional communication rails are designed to ease onboarding at the cost of security and user autonomy. They make up 0.07% of the observable Autonomous Systems, and the five major telecom hosts make up 16% of the networks unreachable peers, demonstrating a clear lack of fault tolerance to routing attacks as well as in the software managing transactions. Additionally, at the time of this research, the top 100 Autonomous System routed 61% of transactions for peers across the network. Throughout history, internet service providers have shown a tendency to consolidate, with further consolidation representing greater Bitcoin centralization.
IP centralization exposes Bitcoin to all sorts of malicious attacks. IPs can wreak havoc on the network by delaying the propagation of blocks, or by hijacking all IPs prefixes, then subsequently dropping all interconnected routes to partition the network in half. Two Bitcoins unable to communicate with one another create parallel blockchains and could create havoc on commerce; unaware transactions are only being published to a partitioned network.
It is unlikely that there will suddenly be a shakeup in the telecommunications industry, and if Bitcoin users continue to rely heavily on traditional lines for internet money, they will be exposed to traditional bottlenecks associated with them.
While we may see some elect to provide internet service for themselves, it is unreasonable for the majority of the non-technically community, and thus routing will continue to be an issue going down the road. While software-defined radio represents an interesting way to communicate among nodes without relying on traditional communication rails, users are few and far between, and in the foreseeable future, it is not a reputable option for most users. However, the Bitcoin holders' strong community and deep pockets are an advantage that earlier bootstrapping of amateur radio did not have.
While many may argue that the satellite infrastructure provided by Blockstream can give users an alternative route to connect to the network and a constant peer to help avoid partitions, the reality is users are pointing their dish into the sky toward only one relaying peer who is controlled by a regional telecommunications company. While satellites do avoid IPs, only one off-grid accessible peer exposes users to similar risks.
100 IPs are associated with 89% of all the propagations, and 50 unreachable IPs are involved in 43% of all the propagations.
Network centralization can easily create a "bottleneck" threat where the likes of BreadWallet & BitcoinWallet or a major telecommunication company create chokepoints in the network. The Bitcoin ecosystem would benefit tremendously from more competition in the mobile wallet space. Going forward, it appears it is becoming more important to find alternative ways to broadcast and receive Bitcoin transactions as a dependance on centralized telecommunications companies creates inherent risk for Bitcoin. Protocols like SABRE have been put forth to protect Bitcoin block promulgation by providing a stable infrastructure capable of dealing with complex denial of service attacks. It will take a comprehensive combination of new protocols and routing mechanisms in order to protect Bitcoin from its own users if they continue to pile in behind the same gates.