api.distributed.press VS node-replication

I think the obvious match here is how IPFS can use a dnslink TXT file to point to the content-addressable top of an IPFS-distributed static web https://docs.ipfs.tech/concepts/dnslink/ . This is how a lot of dweb/web3 websites maintain and advertise a non-http mirror. It works out of the box for sites served using https://distributed.press/ (non-commercial, open source) or https://fleek.co/ (hosted).

If you use Brave, or have the IPFS browser extension, you can access sites like https://ffdweb.org/ as an option, or preferentially.

This should be a link to https://distributed.press/

There are several new and exciting decentralized web protocols. I am working on a tool to publish to all of them. Right now it publishes to IPFS and Hypercore, and will eventually serve signed plaintext so content can be shared to Scuttlebutt and Aether networks.

For example, this site is published using https://github.com/hyphacoop/api.distributed.press and it can be accessed over HTTP, IPFS, and Hypercore.

https://staging.compost.digital

As pointed out at https://nrkernel.systems/book/, we developed some black-box techniques(https://github.com/vmware/node-replication) to convert data structures into their concurrent and NUMA-aware versions. These techniques are mostly based on state-machine replication using operation logs and replicate the state on each NUMA node.

> IPFS, Hypercore or some other protocol available to the mainstream, whether that be in Firefox or by some other means

TIL about hypercore! Looks like an interesting concept based on append-only logs. It reminds me of this work being done at VMware: https://github.com/vmware/node-replication