The past few days I've been reading a lot about NNCP and similar distributed/decentralized/delay-tolerant technologies and, while I haven't felt the need to go all in, it has been mentally...
The past few days I've been reading a lot about NNCP and similar distributed/decentralized/delay-tolerant technologies and, while I haven't felt the need to go all in, it has been mentally stimulating. It's fascinating to read about some technology, feel excited and wonder why everything isn't more like this, and then realize the larger problems which the technology doesn't really address. It's a recurring theme of humanity that we have to choose inferior tools just because they break in slightly different ways.
I don't personally have a use for NNCP but it is still a very interesting toolkit to study. There are a lot of great ideas here.
Anyway. I found this article while reading through John Goerzen's site and wanted to share it. Here are a few more:
I got into the IT field in the early 2000s, and the author touches on a few things that I didn't understand at all back then that I finally understand now. As a fresh new face, I didn't understand...
I got into the IT field in the early 2000s, and the author touches on a few things that I didn't understand at all back then that I finally understand now. As a fresh new face, I didn't understand why you'd use anything but x86 and TCP/IP. Those were the technologies I was familiar with as a home user, and I was suddenly thrown into a world of fiber channel, MPLS, DS3 circuits, as/400s, and so forth. They were inflexible, not interoperable, and offered extremely limited performance benefits if they offered any at all. When I took my CCNA, I learned about serial networking interfaces, and my first thought was "why would ANYONE ever choose to use an interface that's ridiculously slow, you have to worry about timing for, and which requires special cables that you need to special order?"
I had many, many conversation with greybeards who tried to explain all the great benefits all this stuff has which I didn't understand because I never worked in enterprises before.
I realize looking back that the reason all that stuff was used was because they worked well enough and they were what everyone was used to. I just wasn't used to them so they seemed weird and archaic to me, but these guys had been working with this stuff for decades. To them TCP/IP was the weird thing, associated with home internet users, not fit for use in a data center.
It's weird now that I'm in their position, having to let go of old technology that "just works", but is not as capable or financially prudent as other solutions. I remember a lot of what the author was talking about, and I remember it fondly with a sense of nostalgia despite what a massive pain in the ass it was.
I've definitely suggested POTS modems for out of band management before and been told "why the hell would we do that? Just throw a 5g access point in there and be done with it" and then had to concede that I'm just being old.
Hah, yeah, similar experiences. Funnily enough, I deployed a new hub at a major Internet exchange and deployed a 5g modem as out of band emergency access. Only the 5g coverage in this major global...
Hah, yeah, similar experiences. Funnily enough, I deployed a new hub at a major Internet exchange and deployed a 5g modem as out of band emergency access. Only the 5g coverage in this major global metro, inside an Internet exchange, was too spotty. So we paid for a slow DIA circuit in addition to our commercial transport links. The only reason we had tried the 5g modem is because we already used a service that gave us a single access portal to all our sites out of band interfaces, some of which were fairly rural and it just worked. But not in the heart of the Internet!
The past few days I've been reading a lot about NNCP and similar distributed/decentralized/delay-tolerant technologies and, while I haven't felt the need to go all in, it has been mentally stimulating. It's fascinating to read about some technology, feel excited and wonder why everything isn't more like this, and then realize the larger problems which the technology doesn't really address. It's a recurring theme of humanity that we have to choose inferior tools just because they break in slightly different ways.
I don't personally have a use for NNCP but it is still a very interesting toolkit to study. There are a lot of great ideas here.
Anyway. I found this article while reading through John Goerzen's site and wanted to share it. Here are a few more:
I got into the IT field in the early 2000s, and the author touches on a few things that I didn't understand at all back then that I finally understand now. As a fresh new face, I didn't understand why you'd use anything but x86 and TCP/IP. Those were the technologies I was familiar with as a home user, and I was suddenly thrown into a world of fiber channel, MPLS, DS3 circuits, as/400s, and so forth. They were inflexible, not interoperable, and offered extremely limited performance benefits if they offered any at all. When I took my CCNA, I learned about serial networking interfaces, and my first thought was "why would ANYONE ever choose to use an interface that's ridiculously slow, you have to worry about timing for, and which requires special cables that you need to special order?"
I had many, many conversation with greybeards who tried to explain all the great benefits all this stuff has which I didn't understand because I never worked in enterprises before.
I realize looking back that the reason all that stuff was used was because they worked well enough and they were what everyone was used to. I just wasn't used to them so they seemed weird and archaic to me, but these guys had been working with this stuff for decades. To them TCP/IP was the weird thing, associated with home internet users, not fit for use in a data center.
It's weird now that I'm in their position, having to let go of old technology that "just works", but is not as capable or financially prudent as other solutions. I remember a lot of what the author was talking about, and I remember it fondly with a sense of nostalgia despite what a massive pain in the ass it was.
I've definitely suggested POTS modems for out of band management before and been told "why the hell would we do that? Just throw a 5g access point in there and be done with it" and then had to concede that I'm just being old.
Hah, yeah, similar experiences. Funnily enough, I deployed a new hub at a major Internet exchange and deployed a 5g modem as out of band emergency access. Only the 5g coverage in this major global metro, inside an Internet exchange, was too spotty. So we paid for a slow DIA circuit in addition to our commercial transport links. The only reason we had tried the 5g modem is because we already used a service that gave us a single access portal to all our sites out of band interfaces, some of which were fairly rural and it just worked. But not in the heart of the Internet!