> Protocol multiplexing exists. But you will have to agree on a single protocol
I may be misunderstanding your message here, but the requirement to agree on a single protocol isn't true when you're using multiplexing. I think you're confusing tunneling with multiplexing.
With multiplexing, you have multiple protocols listening on a single port. The multiplexer server sniffs the first few bytes of what the client sends to determine what protocol is being used, then decides which back-end to forward the connection to.
Neither the client nor the final back-end need to be aware that multiplexing is happening, and likely aren't.
Through this, you can use both HTTPS and Telnet on port 443 without the Telnet client needing to have any changes done.
There’s an expected amount of defects per wafer. If a chip has a defect, then it is lost (simplification). A wafer with 100 chips may lose 10 to defects, giving a yield of 90%. The same wafer but with 1000 smaller chips would still have lost only 10 of them, giving 99% yield.
As another comment referenced in this thread states, Cerebras seems to have solved by making their big chip a lot of much smaller cores that can be disposed of if they have errors.
Indeed, the original comment you replied to actually made no sense in this case. But there seemed to be some confusion in the thread, so I tried to clear that up. I hope I’ll get to talk with one of the cerebras engineers one day, that chip is really one of a kind.
Defects are best measured on a per-wafer basis, not per-chip. So if if your chips are huge and you can only put 4 chips on a wafer, 1 defect can cut your yield by 25%. If they're smaller and you fit 100 chips on a wafer, then 1 defect on the wafer is only cutting yield by 1%. Of course, there's more to this when you start reading about "binning", fusing off cores, etc.
There's plenty of information out there about how CPU manufacturing works, why defects happen, and how they're handled. Suffice to say, the comment makes perfect sense.
That's why you typically fuse off defective sub-units and just have a slightly slower chip. GPU and CPU manufacturers have done this for at least 15 years now, that I'm aware of.
You've got it wrong. It doesn't have to be HTTP[S] traffic.
Reverse proxies can disambiguate based on the SNI. I could run telnetd on port 23, but have port 23 firewalled off, and have my reverse proxy listening on port 443 with TLS forward anything going to telnet.mydomain.com to telnetd. Obviously, my client would need to support that, but a client-side proxy could easily handle that just as well.
Yes you can run any service on any port. But tunneling telnet over another protocol seems like you would just move the problem. I don't know too much about SNI but if "Reverse proxies can disambiguate based on the SNI" wouldn't your network service provider also be able to filter based on SNI?
You would need to agree on a protocol and you would gain all the advantages but also the disadvantages of the tunneling protocol.
> wouldn't your network service provider also be able to filter based on SNI?
Two things:
1. Only if they knew that the hostname in question is indeed being used for telnet tunneling. You can set that host name to whatever you want.
2. Encrypted SNI is a thing.
> You would need to agree on a protocol and you would gain all the advantages but also the disadvantages of the tunneling protocol.
Yeah, admittedly the entire thing is a bit contrived. If your client is capable of speaking the tunneling protocol, then likely you'd just use the tunneling protocol itself, rather than using it to tunnel telnet.
I don't draw the same conclusion. I think they've made teen pregnancy a horror to avoid, which is totally fair.
reply