Re: Proxy 2.0 secure? (AG vs. SPF)

["Ryan Russell" <ryanr () sybase com>]

> Even in pure SPF mode, VPNs, remote management consoles,
> etc. use the host IP stack of the gateway, which is supposedly
> vulnerable to the same types of attack that the AG's stack is.  You
> can't not have a stack on the SPF if you want to do some things.  In
> those cases, it's not really an AG, it's just a server.

There are VPN implementations that don't rely on the host IP
stack.  Management functions could be done via SPF without
using the host IP stack, too.  It wouldn't be tha hard to use
the SPF VPN code to verify the connection, and then just
extract the data portion of the packets, UDP sytle.

> Yes, it can help. a bug running in a non-privileged context is much less
> serious than one running in such a context.  FWIW, it doesn't mean the
> vendor has to write their own stack,

If the vendor wants to be able to verify all of the security aspects
of their gateway, they have to write the stack themselves, or
get the source and verify it themselves, making any needed changes,
or trusting it for that reason.  I said it for that reason, not for any
kind
of B2 reason, which I am, again, ignorant about.

> Most SPFs have only a single context, so your definition doesn't seem to
> bear out what the market has engineered.

I don't know of any implementation of the kind of SPF I'd like to see.

> > I get a lot of people claiming that tables that maintain transport info
> > are too hard.  I haven't written an IP stack, but isn't that what they
do?

> For some transports they do, for others they don't.  An SPF has to
> maintain it for all transports, even those which don't have state
> information, which means guessing when things are over.

FW-1 uses timers.  Obviously, you'd want to know the protocol
so know when the conversation was over, if it's possible
with that protocol.

> Also, when you
> throw in the application layer code, your state tables will become a
> real mess.  I think it's easy to infer how difficult it is to test
> things by the state of SPFs on the market.

Only if you handle application logic as a state machine.  This would
be pretty easy with some protocols, and obviously not for
others.  Application logic doesn't have to be written like a
state machine to sit on an SPF.

> And I'm saying that the complexity of doing so without using a stack is
> enourmous, and very difficult to verify, especially in the interface to
> the application layer, sequencing, fragmentation handling, and
> extensions to new protocols.  At this point, building packet filtering
> in front of an IP stack, to me, makes much more sense and is
> significantly easier to verify, and verification is very important in
> security.

I assume that IP stacks are hard to verify (I haven't done it
before.)  I assume this because of all the IP stack problems
that crop up.

> Well, if you're going to magically re-create the Berkely sockets
> interface, and full functionality in your SPF, you're going to end up
> with an IP stack, at which point, you'll have gained not a lot, except a
> great understanding of IP stacks IMO.

If you limit yourself to the sockets interface, you lose the ability
to get the lower-layer information you might want to know about
in a security context.  It would be reasonable to have more that
one API into the lower part of the SPF.  One that lets you write more
informed application protocol handling code, and one that
lets you re-compile sockets proxies, so you can make it
act like an IP stack.

> Cisco routers have an IP stack.  That's why you can telnet to them.
> For the purposes of that, SNMP mangagement, and VPNs, they are a server,

It has an IP stack, but does it use it for the VPN?  (I don't know,
I'll check next time I set up an IOS VPN.)

> I've yet to see an SPF with accept(), listen(), connect(), etc.  Once
> again, I'd really like to see articulated what the gains of such would
> be over adding filtering in front of an IP stack.  I don't see any, and
> indeed, I think it's a detriment to come from the other end of the field.

As mentioned above, this lets you have one "IP stack" instead
of two... i.e. just SPF IP stack, instead of SPF and real IP stack.

> Every implementation that I've seen has been either part of the kernel, or
> tied to drivers running in the kernel space.  Even if you were to open a
> raw socket, that lives there.  In some OS', you stand to take a context
> switch for every packet that goes into user space, which isn't going to
> do wonders for your performance.

They have to be root in order to "own" the NIC.  I imagine this could
be fixed.

> I'm still not convinced that recreating the complete behaviour of an IP
> stack just to get the code into user space is worth doing (especially
> if said behaviour includes VPNs or remote access).  I don't
> see the risks of the IP stack in being kernel mode as that
> significant.  The only attack vector that seeks to stop is an attack on
> the gateway itself.  Since IP stack attacks have all been, and seem to
> be limited to DoS attacks, they can be perpatrated just as well for code
> running in a user context.

As long as it's just DoS attacks, and you don't mind going down when
new attacks come out, that isn't a problem.

> Well, I'll assert that IP itself, and indeed any packet based data
> protocol is subject to some flood attacks as a matter of course.  SYN
> flooding certainly lowered the bar on a flood attack, but it didn't stop
> the fact that panix.com could still be flooded off the Net.  I certainly
> have enough bandwidth available to flood them off the Net.  Repeatedly.

Bandwidth attacks can't be defended against on the far end-point.
SYN attacks are a result of bad IP stack design decisions, and can
be defended against at the far end-point.  They're not really the
same class of attack.

> Take a look at the Linux SYN and RST cookies for a solution.  FWIW, most
> of my external systems were already tuned enough to not be threatened by
> SYN flooding due to their normal traffic loads.  SYN flooding only
> lowered the bar to floods, which are a fact of life in a packet based
> network where one doesn't control all the machines.

As I said, most OSes didn't really fix the problem.  I assume that they
didn't because they're either lazy, or its too hard.  So, those vendors
aren't really suitable to have as your IP stack vendors.  As usual,
Linux is the exception to the sloppy unix vendor.  (Everybody relax..
I'm taking jabs at Sun, HP, etc..  not everyone's favorite flavor of
*bsd, etc... :)

> This issue is that fragmentation reassembly is done differently in
> different stacks, so your SPF would have to know how a particular host
> reassembles fragments before it could protect that host if it isn't
> doing frag reassembly itself (see my earlier comment on this).  Of
> course, if it is reassembling fragments, it has to do it one way, so
> non-malicious fragments could break for some instances.

AGs always rebuild frags in one way, correct?  They don't have
to know about each inside stack, correct?  Then SPFs just
need to have the same behaviour.  If rebuilding frags in a particular
way can take out some inside host, the AGs suffer the same, correct?

> At which point, you have gained what exactly?  Are you saying that the
> OS isn't going to handle packets at all?  Why not just make the network
> devs user R/W and run your stack in user space as a user mode driver
> then?  Same gain, nowhere near as much work.

Yes.. then the OS isn't allocating, and "using" the IP stack for itself,
so the gateway isn't vulnerable, and the inside machines have
the same amount of protection.  See my earlier definition of
an SPF.

> Since there isn't yet an SPF solution which truly applies its state
> protection mechanisms to everything it passes, the discussion is rather
> academic,

Yes, I'm trying to have a theoretical discussion.  I use commercial
products as examples to help clarify things (hopefully.)

> but perhaps you'd like to enumerate "handles all the protocols
> desired"?  For me, if a protocol isn't given some sort of transport,
> internetwork, and application layer protections, it's not "handled", and
> I'm not gaining significant protection over a non-stateful packet
> filter.

Yup.  An excellent HTTP proxy doesn't help all that much if
you need to pass 100 other protocols, and you end up having
to packet filter on the AG gateway.  At that point you'd probably
want an SPF for that.

> As for the state tables, the complexity is fairly easy to figure out,
> and you can do this on paper by flowcharting.  Start with the first
> fragment of a fragmented packet for a TCP connection for HTTP.  Note the
> things you can keep information on, such as frag number, size, frag
offset,
> sequence number, source port, dest port, source IP, dest IP, as well as
> some application transport protocol information, and application layer
> data information.  Make a decision box for each of them, and hook it
> into your flow of inspection.  Then do the next few packets, keeping in
> mind packets too large, fragment overlaps, and every problem you can have
at
> the IP layer, and the TCP layer.  Now add in problems at the HTTP layer,
then
> at the data layer.  Look at how much you have to investigate to see if
> the next packet or fragment should get through.  Now, take into account
> that a mistake or something overlooked in any of those areas means that
> an internal host is potentially compromised.  Blocking Javascript on a
> per-packet basis, with out-of-order packets and fragments is left as an
> extra credit item.

While we know from Computer Science classes it's certainly
possible to do the whole thing in a state machine, we could
probably also guess that it wouldn't be much fun.  In practical terms,
you'd want the app logic from a traditional HTTP proxy, with hooks added in
to take advantage of the SPF information underneath.

> Why down versus compromise?  Because, from what I can see in the code of
the
> stacks I've got, the potential problems are with overwriting internal data
> structures, not getting onto the stack.  In the internal machine case, it
may
> be a DoS, or it may be an intrusion, depending on what slips by, and what
it
> effects.  So, yes, it's my assertion that AG's are inherently more secure
> than SPFs, as a design.

I know of at least one example of a really problem with an IP stack.
It's not pushing code unto the remote machine, but it's an interesting
example anyway.  Up until the most recent versions, Sun's IP stack
from MS-DOS (called something like PC-NFS?) had a really interesting
side-effect.  If you sent it a really large ping, like in the 64K
neighborhood,
the last bit of the pick replys wouldn't match what was originally sent.
The
guy who noticed this broke out a sniffer, and loooked at the last packet
to see what was different.  He found the username and password
(in cleartext) of the user who was logged-in on the remote machine.
This was an all-in-one IP stack, so ping was handled by the stack itself.
Don't run your AG on this stack :)

> Implementations are another matter of course, but you seem to assert for a
> perfect SPF versus an imperfect host for the AG, if you level the playing
> field, an imperfect SPF risks much more, and a perfect SPF either starts
> writing applicaiton layer code, or drowns in a sea of state.

I can't really speak to implementation difficulty until/unless I write one
someday.  You don't have to use a state machine to write an SPF.
You seem to imply that ripping IP out of the kernel into userland
wouldn't be too hard, and I claim that's a type of SPF.  The advantage
is that you protect the gateway AND the inside machines.

> For all their blusering of the benifits of SPFs, the current set of
vendors
> has moved rather quickly to writing application layer code to sit on their
> engines.  I'll also assert that it's easier to protect a stack than to
> recreate that stack.

Yes, they suck.  Well, to be fair, I only have personal knowledge of
FW-1 sucking.  I don't know enough about PIX except aparantly
their VPN sucks.  I don't know if the Linux SPF-like packages
suck, but if it's discovered that they do, I assume it will be fixed within
a day or two.

> Whilst keeping state does add some measure of protection, it really only
> brings value to non-TCP protocols, since TCP has a built in state
> mechanism, and that same state can be done in the AG's code for the
> application, or even a generic stateless protocol handler on a raw socket
> without reinventing the wheel for the case of non-TCP protocols.

If all your protocols ride on TCP that might be ok.

                    Ryan