[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: I-D.ietf-v6ops-cpe-simple-security-10



On Thursday 22 April 2010, Mark Townsley wrote:
> On 4/22/10 6:38 PM, Timothy Baldwin wrote:
> > There are many ssh servers with default or poorly chosen passwords, so
> > that might not be a good idea.
>
> So Bob's bad password means Alice has to figure out how to configure a
> pinhole for ssh because she wants her ssh to be reachable from the
> Internet.

On the other hand: Alices (I would call Alice a geek - she is using ssh 
after all) desire to have a reachable ssh port means that Bob (let's assume 
he knows nothing about computing, but is a very fine chef) is vulnerable to 
attacks. Just because nobody told him to click a specific button that he 
knew nothing about.

I would assume Alice capable of opening a port, but would guess Bob gets 
very apprehensive when he is told to configure his home router.

> Hardly fair, particularly when we don't even know what the 
> intrusion rate would be for a given service/protocol over IPv6.

I think we can assume it to grow proportionately with the growth of IPv6 - 
it is practically nil now because attackers don't consider IPv6 gurus dumb 
enough to have default passwords on open ssh ports.

> Modern IPS firewalls actually have provisions for sniffing and blocking
> login attempts from the outside using poor/default passwords. But this
> isn't foolproof, and requires digging deeply into packets.

I don't think the master chef Bob would have an IPS firewall - all he wanted 
was to connect his kitchen computer to this Internet thing because we heard 
there were good recipes out there.

> It would be really nice if applications would be easily configurable for
> what IPv6 address scope they could use. Then when you turn on SSH, you
> are simply asked whether it should be made available on the Internet or
> not.

Open a feature request on your favorite bugzilla... ;-)

More realistically: there will always be thousands of very popular 
applications out there with insecure defaults. I spent the last ten years 
(that's my entire career) trying to educate my programmer colleagues and 
friends about secure programming. I think I can count the ones I've reached 
on one hand - either I'm a bad teacher or it is not a popular topic.



	Konrad

Attachment: signature.asc
Description: This is a digitally signed message part.