I can’t tell you the number of devs I’ve met that know jack shit about infrastructure and networking. Even simple questions like understanding their subnet or how a load balancer works.
That’s because an “IT engineer” is like an “engineer in chemistry”. You have so many sub-disciplines that a given engineer can’t be expected to know everything. Nobody is surprised if an organic chemist doesn’t know much about heat exchanger optimisation, or a chemical process engineer is confused by organic synthesis.
In the same way, I’ve met devops guys that know surprisingly (to me) little about the intricacies of hand writing multi-threaded high performance calculation code. In the other hand, they’re wizards at stuff like load balancing and optimising how we use sub-processes for I/O, CPU-bound or memory bound tasks. You would think the two are very similar, but even then it quickly becomes clear that people really become specialists in their field.
Sure. I consider myself a fairly good dev. Not an amazing genius or anything but pretty competent in my area. Wouldn’t have a clue how a load balancer works beyond what’s in the name, I guess it balances loads. I can assume it does this by distributing requests to different servers because that just makes sense, but I don’t even really know that. Networking might as well be black magic as far as I’m concerned. I’m grateful we have a decent infra team who knows about all that. So I’m not surprised.
I’ve been developing for decades so obviously I know that the subnet is where the Lord Marshall went and returned with powers beyond mortal comprehension.
That’s because they’re back end devs. There is one way to do things, you learn the one way to do the thing, you do it that way, you’re done.
The fun is in front end, there is 15 different ways of doing things, but you don’t like any of them, so you develop a 16th way, eventually you get bored of the 16th way so come up with a 17th way. Eventually you get sick of all of this and go back to SSGs as God intended.
The one correct way to do back-end is obviously to have lots of microservices all talking to a shared database so they can take each other down and also they’re never updated because you’d need about 10x the devs you have to maintain all of them. I know this because that’s everywhere I’ve worked for over a decade.
When you translate it out it bits it’s the number of bits that are active in the net mask
0000 0000 0000 0000 0000 0000 0000 is a /0
1111 1111 1111 1111 1111 1111 0000 0000 is a /24
1111 1111 1111 1111 1111 1111 1100 0000 is a /26
And since it’s used for matching, it follows how many there are.
Edit: typed that quickly, but to expand it further, take a /24 subnet, that is 255.255.255.0, if you look above each group of 8-bits can cover 0-255, so you in essence by having 24 bits turned on that’s what you want for an exact match. You’re basically saying “the first three octets need to exactly match”
Look I’m mostly in infrastructure and I don’t know much beyond core principles behind software development, I can’t really cast stones. But yeah they can be surprisingly ignorant sometimes.
I can’t tell you the number of devs I’ve met that know jack shit about infrastructure and networking. Even simple questions like understanding their subnet or how a load balancer works.
That’s because an “IT engineer” is like an “engineer in chemistry”. You have so many sub-disciplines that a given engineer can’t be expected to know everything. Nobody is surprised if an organic chemist doesn’t know much about heat exchanger optimisation, or a chemical process engineer is confused by organic synthesis.
In the same way, I’ve met devops guys that know surprisingly (to me) little about the intricacies of hand writing multi-threaded high performance calculation code. In the other hand, they’re wizards at stuff like load balancing and optimising how we use sub-processes for I/O, CPU-bound or memory bound tasks. You would think the two are very similar, but even then it quickly becomes clear that people really become specialists in their field.
Those are the ones who can maintain eye contact, aren’t they?
Sure. I consider myself a fairly good dev. Not an amazing genius or anything but pretty competent in my area. Wouldn’t have a clue how a load balancer works beyond what’s in the name, I guess it balances loads. I can assume it does this by distributing requests to different servers because that just makes sense, but I don’t even really know that. Networking might as well be black magic as far as I’m concerned. I’m grateful we have a decent infra team who knows about all that. So I’m not surprised.
You are not a real dev unless you can build your own CPU atom by atom
where is the lemmy community for those who self-host chip fabrication?
everyone else is a poseur.
I’ve been developing for decades so obviously I know that the subnet is where the Lord Marshall went and returned with powers beyond mortal comprehension.
You keep what you kill.
That’s because they’re back end devs. There is one way to do things, you learn the one way to do the thing, you do it that way, you’re done.
The fun is in front end, there is 15 different ways of doing things, but you don’t like any of them, so you develop a 16th way, eventually you get bored of the 16th way so come up with a 17th way. Eventually you get sick of all of this and go back to SSGs as God intended.
The one correct way to do back-end is obviously to have lots of microservices all talking to a shared database so they can take each other down and also they’re never updated because you’d need about 10x the devs you have to maintain all of them. I know this because that’s everywhere I’ve worked for over a decade.
There’s a perfect way to do anything in frontend but only on the dev’s one computer. Everyone else will get garbage.
Yeah, but why is CIDR notation like that? You put a bigger number at the end and your subnet becomes smaller? Why would you want that?
When you translate it out it bits it’s the number of bits that are active in the net mask
0000 0000 0000 0000 0000 0000 0000 is a /0
1111 1111 1111 1111 1111 1111 0000 0000 is a /24
1111 1111 1111 1111 1111 1111 1100 0000 is a /26
And since it’s used for matching, it follows how many there are.
Edit: typed that quickly, but to expand it further, take a /24 subnet, that is 255.255.255.0, if you look above each group of 8-bits can cover 0-255, so you in essence by having 24 bits turned on that’s what you want for an exact match. You’re basically saying “the first three octets need to exactly match”
…well I don’t have to know that because AWS is like, serverless, dude.
It’s in the cloud
It’s a series of clouds.
Not just your industry. Every industry is like this.
Look I’m mostly in infrastructure and I don’t know much beyond core principles behind software development, I can’t really cast stones. But yeah they can be surprisingly ignorant sometimes.
The subwhat?
It’s the water based version of skynet.