closure-library VS tinc

> "I recently learned that some Node.js engineers prefer stream classes when building strings, for performance reasons." Pretty much tells you everything you need to know about node js, I guess.

Google Closure Library includes a StringBuffer class. [1]

I recall it having explanatory notes, but I don't see them in the code now. JavaScript engines can optimize a string concatenating to in-place edit, if there is only one reference to the first string. The StringBuffer class keeps the reference count at one, guaranteeing this optimization is available, even if the StringBuffer itself is ever shared.

[1] https://github.com/google/closure-library/blob/master/closur...

This made me wonder what the oldest go-link (from inside Google) discoverable on the public internet is. So far I've found one going back to 2013 (but there should definitely be some from the mid 2000s): https://github.com/google/closure-library/blame/11ed104958a2...

(Fun fact: go-links are so critical to Google ops, that they're expected to be accessible in a "everything is down" scenario.)

For data type I always use DECIMAL(#, 6). So 6 decimal places. This accommodates all generally agreed upon rounding standards for calculating interest, taxes, discounts, etc. So for me, something like $12.75 USD will look like 12.750000 in the database. I store the 3-letter currency code in a separate column. I know what the base value for each currency is on the code side in our framework, which has a currency library that reads a dataset that has every currency stored like this (the 'rules' key is a bitmask per Google's i18n currency.js):

But both Nebula and tinc max out at around 1 Gbit/s on my Hetzner servers, thus not using most of my 10 Gbit/s connectivity. This is because they cap out at 100% of 1 CPU. The Nebula issue about that was closed due to "inactivity" [2].

I also observed that when Nebula operates at 100% CPU usage, you get lots of package loss. This causes software that expects reasonable timings on ~0.2ms links to fail (e.g. consensus software like Consul, or Ceph). This in turn led to flakiness / intermittent outages.

I had to resolve to move the big data pushing softwares like Ceph outside of the VPN to get 10 Gbit/s speed for those, and to avoid downtimes due to the packet loss.

Such software like Ceph has its own encryption, but I don't trust it, and that mistrust was recently proven right again [3].

So I'm currently looking to move the Ceph into WireGuard.

Summary: For small-data use, tinc and Nebula are fine, but if you start to push real data, they break.

[1]: https://github.com/gsliepen/tinc/issues/218

[2]: https://github.com/slackhq/nebula/issues/637

[3]: https://github.com/google/security-research/security/advisor...

From a purely networking perspective, there are far better solutions than tailscale.

Have a look at full mesh VPNs like:

https://github.com/cjdelisle/cjdns

https://github.com/yggdrasil-network/yggdrasil-go

https://github.com/gsliepen/tinc

https://github.com/costela/wesher

These build actual mesh networks where every node is equal and can serve as a router for other nodes to resolve difficult network topologies (where some nodes might not be connected to the internet, but do have connections to other nodes with an internet connection).

Sending data through multiple routers is also possible. They also deal with nodes disappearing and change routes accordingly.

tailscale (and similar solutions like netbird) still use a bunch of "proxy servers" for that. You can set them up on intermediate nodes, but that have to be dealt with manually (and you get two kinds of nodes).

Two other options are Tinc https://tinc-vpn.org/ or Nebula https://www.defined.net/nebula/

And there is Tinc; the OG overlay network. I don't have experience with this. Seemed a bit of a pain to setup. https://tinc-vpn.org

For what its worth I have used the open source Tinc VPN [1] for mesh multihop routing for ages. It is nowhere near as fast as Wireguard but I could envision Tinc incorporating support for Wireguard if the author were so inclined. Like you mentioned Tinc does not mesh with other VPN's AFAIK.

[1] - https://tinc-vpn.org/

This is actually very simple in concept and is just as simple or even simpler to do with tinc (https://tinc-vpn.org).

Since I can use tinc in bridge mode, I can run tinc on the upstream server and on a local machine which then provides access to several physical machines without running extra software on each of those machines, which is particularly useful for machines that are resource limited, like my Macintosh LC II and LC III+:

http://elsie.zia.io/

It'd be nice if it weren't so difficult to get public addresses.

I clicked expected some broken analogy between https://tinc-vpn.org/ and the Catan board game, but instead it is a Catan implementation. Fair enough.

will generate a real-time network graph using the Graphviz DOT language. It's a cool feature that I find quite useful.

[0] https://tinc-vpn.org/