Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
matt-p
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
9 ms
·
181.
▲
by
matt-p
11mo ago
Yeah that obviously only works on /56 and above because networks should be a minimum of /64. I use k3s and each host has a /64; cilium just gives each pod a /80 and the host does NDP and stuff. Works fine, no need to req
182.
▲
by
matt-p
11mo ago
I'm talking about an internal network, not the public connection.
183.
▲
by
matt-p
11mo ago
On your *internal* network e.g the thing between your postgres VM and your webserver (or whatever). Not arguing against it on the public/wan connection.
184.
▲
by
matt-p
11mo ago
It doesn't impeed observability for goodness sakes. It does however impeed accidentally opening up your internal network because you don't really understand your firewall/virtual router/whatever.
185.
▲
by
matt-p
11mo ago
A container should absolutely not even need a /72. The traditional reason for /64 is for slaac but you most certainly don't need that for one container (if at all honestly).
186.
▲
by
matt-p
11mo ago
Yes, not addressable is even safer. Especially so for someone not specialized in networking.
187.
▲
by
matt-p
11mo ago
Why do you need ipv6 on your internal network? Is 10/8 really not enough/overlap? For 99.99% of people it's fine for the internal interfaces and if anything actually simplifies configuration.
188.
▲
by
matt-p
11mo ago
You outsource it because clouflare have more locations than you so offer lower latency and can offer it at a cost that's cheaper or the same price as doing it yourself.
189.
▲
by
matt-p
11mo ago
I've got to say, I feel like bright white full spectrum is the main problem, I would care much less about them if they were red (though yes, we would need a new colour for rear/brake lights).
190.
▲
by
matt-p
11mo ago
DDOS, but I don't really understand why in particular.
191.
▲
by
matt-p
11mo ago
If you are DO you could, you just decided not to bother. They control the origins it's spaces (s3), so they could absolutely spin up further gateways or a cache layer and then turn the CDN off.
192.
▲
by
matt-p
11mo ago
Just spaces CDN, not spaces - you'd think they'd just turn the CDN off for a bit.
193.
▲
by
matt-p
11mo ago
I already know how to do it, I just don't see the value in it.
194.
▲
by
matt-p
11mo ago
There's no public evidence. Lets wait for the blog post.
195.
▲
by
matt-p
11mo ago
Cloudflare down because of a DDOS is extremely funny.
196.
▲
by
matt-p
11mo ago
The economic costs of that fall on the (residential) ISPs and they aren't really incurring very much cost in additional bandwidth from the outgoing attacks. In most cases it will be 0. It's not 'good', as it could affect
197.
▲
by
matt-p
11mo ago
It's so funny every time we build a nuclear plant we say 'ooooh expensive' then by the time it's built it turns out it's ~ at the cost of gas.
198.
▲
by
matt-p
11mo ago
We could probably do with a small amount of storage as we do have days where we pay for turbines to /not/ generate.
199.
▲
by
matt-p
11mo ago
I adore openbsd and have been using it since 4.x however it is still slow, not slow to boot or anything like that but if you run it as a web server it manages about half the req/s of Debian. Network performance is also slower than Debi
200.
▲
by
matt-p
11mo ago
1/16 of a CPU is admittedly more terrifying, I remember wayyy back in the days of shared hosting we didn't give less than 1/5th a CPU, we had all sorts of issues at absolutely anything higher than that.
201.
▲
by
matt-p
11mo ago
I think in fairness it's an apples to oranges comparison. How long will it take you to setup postgres on a VM and sort out and pay for somewhere to put backups? Let's say only a hour, and your time is worth say $30/hour you&#
202.
▲
by
matt-p
11mo ago
That means 25% more datacentre/grid capacity. Genuinely I think most companies are not happy to fund that, in order to save marginally in other areas.
203.
▲
Splice a Fibre
(react-networks-lib.rackout.net)
4 points
by
matt-p
11mo ago
|
0 comments
204.
▲
Datacentre React Components
(react-networks-lib.rackout.net)
1 points
by
matt-p
11mo ago
|
1 comments
205.
▲
by
matt-p
11mo ago
It's interesting; the current state of the railways is clearly a complete joke but 10-15 years ago, my perception at least, was that it was basically a OK and working system. The stats are only a bit worse now than they were then, I
206.
▲
Show HN: Fibre Splice in React
(react-networks-lib.rackout.net)
1 points
by
matt-p
11mo ago
|
0 comments
207.
▲
by
matt-p
11mo ago
I built this while working on a data center design and management tool. Interactive react components for rack and network design. https://react-networks-lib.rackout.net/fibre
208.
▲
by
matt-p
11mo ago
I was not at any point talking about fixed wireless.
209.
▲
by
matt-p
11mo ago
I'm aware, but, um, without sounding insensitive, not for $50 per sub in hardware cost. Tarana is about 10X that. If you have that kind of money labour is cheap enough to run fibre for the same amount.
210.
▲
by
matt-p
11mo ago
I certainly agree about regional licensing. I think the best scenario would actually be to allow some for WiFi and some for carriers, especially since selling licenses is a two way door in a way that ISM isn't.
More ›