dgrnbrg

What about the 3rd founder? He got written out of the narrative early on, but he built the original platform and was bought out in the first few years.

At least one commercial company I've heard of has a sophisticated permissioning model, so that the owner of the chain can publish it and have its causal history and/or integrity be validated by all participants, but with specific permission scopes so that subscribes to the blockchain can only see inside of transactions they have permissions for.

It's not just the root's event log--it's all the event logs. The deferrals allow us to batch writes in an optimized way, and do so with a fully incremental algorithm. In the IO cost model, each "block" read or costs 1 IOP to perform, so we can reduce the IOP cost of writes by a factor of 100-1000x, but reads will still have the same # of nodes to fetch. The event logs augmenting the tree are an effective IO optimization on a B+ tree.

Creator here. As other commenters noted, this data structure only requires keys to be sortable, not hashable, and the best sorted performance we can achieve is O(log(n)).

That said, when a Cuckoo hash gets very full and bounces entries around a lot, there might be an advantage to buffering operations and choosing insertion patterns that reduce the batch's insertion time. Then again, Cuckoo hashes already perform so well for situations they're designed for, so it's hard to improve them with an event log overlay.

The code actually includes a benchmarking tool that's meant to help in figuring out this decision. Once you've selected your key size (or distribution), value size (or distribution), and backend, you can play with these factors. Intuitively, targeting "block" size should improve perf. So, 4k-1MB if you're in memory or local, 1.5k-9k if you're over the network (and depending on configuration).

I did some work to split nodes based on size rather than number of children, but that requires accurate & fast size estimation of the decompressed objects, which isn't possible in general.

Indeed that would.

Creator here. You could use this from Java using the Clojure API for Java; however, many of the extension hooks use Clojure protocols. I would recommend writing a shim to Java for your particular application, so that your data types are represented in the way you want.

That is very cool! When you end up doing reads, do you compute all the pending writes along the path from root to leaf, and then run queries on the projection?

Creator here. There's no visualizations of this yet, but I'll be speaking about it this year at Strange Loop. For that talk, I'll be creating visualizations, so stay tuned!

Creator here. I would absolutely agree with this, but let me shed a bit more light to clarify: the underlying idea of B+ trees with inline caches is not patented (or maybe it was, but a long time ago?). Tokutek's patents are all focused on the methods they use to achieve concurrent operations on a single tree--they have a very interesting & novel locking protocol.

The Hitchhiker tree completely avoids their patents, since it makes very different decisions in order to become purely functional.

Hey HNers! One of the authors here, happy to answer any questions.

I think they've improved a lot WRT fixing problems--we had a chat with them after some issues with Datomic in production, and since then (6 months ago) we've had every problem we've discovered get fixed very promptly, and Datomic's continued to scale for us.