One of the main reasons that ConsensusJ-Namecoin is still in the Beta Downloads section of Namecoin.org is that it carries several patches against upstream ConsensusJ that haven’t yet been upstreamed. This presents two resultant issues:
- There are fewer eyes on the ConsensusJ-Namecoin code.
- ConsensusJ-Namecoin takes some extra time to benefit from new code from upstream, because I need to manually rebase against that new code (and fix whatever conflicts might show up).
Each of these issues reduces the security of ConsensusJ-Namecoin, which means we should fix them by getting our patches upstreamed. We’ve been making some progress on this recently, and last month we saw some additional progress, as I’ve submitted the WalletAppKit support from ConsensusJ-Namecoin to upstream, and upstream has merged it. Mommy, what’s a WalletAppKit? The BitcoinJ library, on which ConsensusJ is based, includes quite a few different modes of operation. The most well-known mode is a mode that downloads blockchain headers and does SPV validation of them. A lesser-known mode simply opens P2P connections to other Bitcoin nodes and lets the user figure out what to do with those P2P connections. The latter mode (known as PeerGroup) is utilized as a component of the former mode (known as WalletAppKit). Prior to my pull request, upstream ConsensusJ’s daemons (
namecoinj-daemon) used PeerGroup mode, and simply assumed (for the purposes of the
bitcoind-style RPC API) that the peers they’ve connected to are telling the truth. ConsensusJ-Namecoin, meanwhile, uses WalletAppKit mode, so that we can be reasonably confident (within SPV’s threat model ) that the name transactions we’re given are actually from the chain with the most work.
As expected, upstream ConsensusJ developer Sean Gilligan noticed some things that could be improved as part of the code review process (which is exactly why we want to get this stuff upstreamed). I made the relevant changes, and the code is merged. Now that WalletAppKit is merged, the most likely next candidate for upstreaming will be the
name_show RPC call, and associated backend code. It’s likely that the
name_show upstreaming will focus on
leveldbtxcache mode, for a few reasons:
leveldbtxcachemode is by far the most secure SPV mode for Namecoin name lookups (substantially more secure than any SPV clients that exist in the Bitcoin world).
leveldbtxcachemode has much faster lookups than the other
libdohj-namecoinname lookup methods, which use an API server.
- The API server that
libdohj-namecoinuses by default when not in
leveldbtxcachemode is WebBTC, which has been down for maintenance for quite a while. Also, WebBTC is problematic due to its lack of consensus safety (it re-implements the consensus rules instead of relying on Namecoin Core’s consensus implementation, and this has caused it to chainfork from the Bitcoin chain before).
- The API-server security model isn’t actually a bad one (it’s a lot easier to Sybil the P2P network than to impersonate an API server), but I tend to think that Electrum is a better (and more standardized) approach to the API-server security model than the
libdohj-namecoinimplementation. It’s conceivable that we could combine the two approaches in the future (maybe make
libdohj-namecoinquery an Electrum instance in addition to the P2P network?), but this is not exactly a high priority for our R&D budget given that
leveldbtxcachemode isn’t in any immediate danger of encountering scaling problems.
Hopefully we’ll see more progress on this front in the near future. In the meantime, I’ve released ConsensusJ-Namecoin v0.3.2.1 (on the Beta Downloads page as usual), which incorporates the latest improvements from upstream ConsensusJ (including Sean’s code review of my WalletAppKit support).
This work was funded by NLnet Foundation’s Internet Hardening Fund.
 The SPV threat model makes a security assumption that the chain with the most work is also a valid chain. This is, of course, not guaranteed to be a correct assumption, although there are some game-theoretic reasons to believe that as long as an economically strong fraction of the network is using full nodes, then SPV is safe for everyone else. You should definitely run a full node for your Namecoin resolution if you have the ability to do so; you’ll be more secure yourself, and you’ll also be making the network more secure for anyone who’s using an SPV node.