Is Cardano thinking about using RINA for DApps or communication among the network? I remember a discussion a while back on how Cardano might be implementing RINA and Charles discussing the use of RINA back in late 2017, but since than, its been crickets.
If anyone is a RINA expert let me know, its an amazing weird way of thinking of how the internet is flat, outdated, and communication is based on distributed communication instead of TCP/UDP.
It is planned to be majorly worked on in “Basho” (~2019). It should help solve the network communication speed itself and also the problem of small private MPC networks (for off-chain smart-contracts and transactioning)
Thank you. And this is a huge leap of creating a truly distributed system. I don’t think most people understand when RINA is implemented, it changes the landscape of the internet. What are your thoughts @vantuz-subhuman?
I have been talking to a network specialist. He says RINA is indeed an amazing tech, but it may require the whole architecture of internet to be replaced: routers, etc.
And you and I know, companies don’t want that.
So my guess is that Cardano will implement RINA in top internet’s current hardware. Unless they have a way to integrate special hardware on stakepool levels. That would be something marvelous to watch.
I am not sure if it’s at all possible so I will let others chime in…
Unfortunately, I know absolute jacksh*t about specifics of RINA, and I still have tons and tons of documentation to RTFM about it, so I just wait for 2019 to lean back and enjoy the ride, learning as we go
But knowing IOHK I expect my reaction to actual implementation details being somewhat like:
(and all the subsequent thoughts, concerns and jokes)
While I have been working with TCP/IP, BGP, AS, DNS, DDOS and so on for almost two decades, this RINA thing mentioned here really surprised and inspired me from the very first moment.
And yes also for me a big question mark is HOW they simply want to implement such a “layer” on all this highly optimized devices with often put on silicon highspeed packet-forwarding logic.
At one side it offers much more anonymousity and DDOS immunity for the SL/CL-nodes, on the other side it looks like falling back to “old-school” dedicated p2p connections defined and established between single network providers. One provider only knows his direct neighbours and so the first step and direction to an (anonymous) host name somewhere behind of it. But you can’t traceroute it or whois/lookup the destination IP’s network owner. At max you can guess by latency and lightspeed calculations if its relative close or on the dark side of the moon.
I’m not sure yet if this would require a separate network protocol layer beyond or beside IP. And so Cisco/Juniper backbone-class devices must support it. Or if it’s “only” some new kind of routing protocol, who looks more like a hosts file, RIP or IGRP instead of todays dynamic BGP-tables.
If only implemented in a virtual way on top of TCP/IP it would be not more then a “CryptoVPN” (somewhere between OpenVPN and Onion-stuff).