TCP/IP daddy Cerf: 'Don't rewrite the internet for security' There is no need to rewrite the basic internet protocols to beef up security, Vint Cerf has said. He also warned that governments are making increasingly heavy-handed attempts to take control of the interwebs. Cerf, co-creator of TCP/IP and currently chief internet evangelist at Google, told delegates at the Atmosphere …
This post has been deleted by its author
Well that would depend on how you define "secure Internet" then wouldn't it.
To you "secure" probably means "all my stuff is secured with encryption" whereas your common or garden government type would define secure as "the proles are not able to hide behind encryption and make snooping on them difficult"
"Cerf, co-creator of TCP/IP . . . told delegates at the Atmosphere conference[. . ] that it was perfectly possible to add security features to the basic internet protocols without requiring a ground-up rewrite, simply by using currently available technology [.. . .] 'The technology is available to do the job, it can be adapted.' he explained. 'Don’t listen to those who tell you it can’t be adapted.'”
Sorry, but the source of your paycheck makes every word out of your mouth suspect.
If you had any knowledge of his background and what he has accomplished over the years you would realize how stupid your remarks are.
And even if you disagree with the company he works for (maybe you would be happy if worked for for Microsoft) how does that in anyway take away from the concepts he is advocating. Do you think the internet should be government controlled and censored?
I hear China is great this time of year.
Hmm this is an issue with the El Reg comments system as I replied to the 1st reply which had "your an idiot" in the subject - I wasn't referring to the OP at all (as, you so/ nicely/ pointed out, he'd done it correctly)
Obviously you couldn't be arsed looking beyond the OP (correct use of your) and looking at the 1st reply (incorrect use of your) but I forgive you as it's El Reg fault!
FAIL: On El Reg comments system, Runcible being a lazy git and me for not explicitly explaining which post I was replying to :)
It does explain the downvotes though!! Maybe I'll get some upvotes instead!
"If you had any knowledge of his background and what he has accomplished over the years you would realize how stupid your remarks are."
I do know about Cerf's background, and I don't care, because it doesn't matter. *Anything* coming out of Google regarding anything that even remotely impinges on public policy and/or Google's business, is immediately suspect.
And if you think otherwise, you are nothing but a gullible fool.
Vint's claim to fame is inventing TCP. Some inventors see their baby as one product to be improved upon, a snapshot in time. Vint sees his as a perpetuity, perfect and never to be replaced. Of course it has many, many flaws, of which security is merely the most egregious.
The ARPANET was the military's, but it was a research net, for a closed community, with host-based security assumed to suffice. There were no networked PCs in 1975! TCP/IP wasn't designed for mission-critical secure use. It is long, long past its sell-by date. Piling hack upon hack onto it is just a stopgap.
(And yes, I suggest looking at RINA at http://www.pouzinsociety.org/ as a substitute.)
Any new Internet protocol or fiddling with existing protocols invites infection on that protocol with content management.
Look at IPv.6, in IP6 land they will know who you are and what you are doing with your packets.
It started off as a innocent solution to the running out of addressing thing and mutated into a IP lawyers wet dream.
The current security challenges are almost exclusively due to buggy implementations of internet endpoints or internet services.
The latest flaw in Windows is a good example: They had a buggy font format parser in the kernel, which could be used to inject malicious code.
If people stopped to used plain C pointers, plain (unchecked) C arrays and plain C++ STL (which is also unchecked), many of these endpoint security risks would simply disappear. Protocol-type risks such as the SSL weakness are not a sign of a general weakness of Internet protocols and can be fixed quickly.
Buggy implementations just cause more problems. The basic design of TCP/IP is that every node can reach every other node, and scan its ports for that matter. It was not designed to be open to the public. Think of a giant motel where every single room faces a street in a very bad neighborhood. Not nearly as secure as a building with a lobby, halls, etc., but IP is a motel.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
