Weak or nonexistent implementations in computer security
software can leave otherwise-secure computers wide open for attack – so open,
in fact, that in some cases it’s as if there’s no
firewall running at all.
Speaking at the annual HOPE (Hackers on Planet Earth) conference in New
York, security researcher Joe Klein of Command Information said
that the internet is full of computers surreptitiously running IPv6,
unbeknownst to their owners. Compounding the problem is the number of operating
systems shipped with IPv6 enabled by default, which includes Windows Vista,
Linux’s 2.6 kernel, Sun’s Solaris, Mac OS X, and a variety of cell phones
operating systems, including Windows Mobile 5 and 6.
Computers with a lackluster IPv6 setup – even if they have a strong IPv4
firewall or Intrusion Detection System (IDS) in place – are just as naked in
IPv6 space as they would be in IPv4-space without a firewall, with any program
that listens for connections allowed to accept them. Most operating systems, by
default, use a handful of “listeners” used for networking and internal
processes – and it is these listeners that are frequently the first to be
targeted in an attack.
A number of computer worms, including Blaster
and its follow-up Welchia,
worked by exploiting a buffer overflow with Windows’ internal RPC
infrastructure, which listens on port 135 and is ordinarily covered up by a
firewall.
Network administrators who don’t keep tabs of their systems face a huge
risk, said Klein. Operational dangers aside, administrators who work for
organizations that have to comply with regulations like HIPAA or Sarbanes-Oxley
risk non-compliance if they don’t secure their IPv6 implementations – whether
they realize they have one or not.
“Essentially, we have systems that are wide open to a network,” said Klein.
“It's like having wireless on your network without knowing it.”
Security researchers have for some time found hackers exploiting IPv6. A 2002
post from Lance Spitzer of the Honeynet project observed a hacker that
broke in to a Solaris-based honeypot through normal means, enabled IPv6
connectivity in the OS, and then set up a tunnel out of the network that went
into another country. The break-in was only discovered due to network
packet-sniffing, and even then Spitzer says he was unable to decode the data
being sent out.
One of the biggest threats is the variety of backwards-compatibility schemes
designed to tunnel IPv6 traffic through an IPv4 system, like Teredo or the 6to4 system: the very act
of tunneling often circumvents firewalls by nature.
“Teredo/ISATAP is currently and will continue to be a major red flag for
networks that have both IP versions enabled, because tunneling confuses the
heck out of a lot of firewalls and IDS deployments,” said an unnamed DoD
security specialist, in an interview with Wired’s Threat Level.
With internet progressives trying to switch the internet to IPv6 as fast as
it can – a widget on Command Interface’s web site estimates that the internet
will run out of IPv4 addresses in about two and a half years – some fear that
technological progress may be outpacing the security that keeps it safe.