Re: ICMP Flooding vs SMURF Attack

From: Peter Plak (plukkie@gmail.com)
Date: Tue Aug 22 2006 - 03:56:08 ART


David,

The first 2 icmp lines, only detect an attack which tries you to be the
reflector (type echo).
I think you also need the icmp echo-reply, like those first 2 lines in your
list. Suppose someone uses reflektors, which will in turn reply to you, and
hists all your complete subnet?

I just did some thinking with a drawing which has the Attacker, Reflector
and Victim. If you have the following scenario: internet------
R1-----internal
If you have to log icmp flooding/smurf attack on R1, you will fullfill with
the following list (for icmp that is):

acl 123 permit icmp any 0.0.0.0 255.255.255.0 echo log-input (smurf from
Attacker to Reflector)
acl 123 permit icmp any 0.0.0.255 255.255.255.0 echo log-input (smurf from
Attacker to Reflector)
acl 123 permit icmp any 0.0.0.0 255.255.255.0 echo-reply log-input (from
Reflector to Victim)
acl 123 permit icmp any 0.0.0.255 255.255.255.0 echo-reply log-input (from
Reflector to Victim)
acl 123 permit icmp any any echo log-input (flooding to Victim or Reflector)
acl 123 permit icmp any any echo-reply log-input (flooding to Victim)

For the udp it's still not clear to me. How does a unix host respond, when
it receives a udp echo?
Does it respond with src-port echo or dst-port echo?

gr

On 8/22/06, David Redfern (AU) <David.Redfern@didata.com.au> wrote:
>
> Peter,
>
> I agree that there probably is further clues in the question. Eg, log icmp
> flood/smurf attack. Everything I can find on an 'icmp flood attack'
> indicates it uses unicast icmp echo's rather than echo-replies, which the
> smurf reply uses. If i was trying to log an icmp/smurf attack, i'd maybe
> add that too.
>
>
>
>
> permt icmp any 0.0.0.255 255.255.255.0 echo log
> permit icmp any 0.0.0.0 255.255.255.0 echo log
> permit udp any 0.0.0.255 255.255.255.0 eq echo log
> permit udp any 0.0.0.0 255.555.255.0 eq echo log
> permit icmp any any echo-reply log
> permit upd any eq echo any log
>
> permit icmp any any echo log
>
>
> First four lines are for smurf/fraggle reflector network, lines 5 and 6
> are for smurf/fraggle echo-replies to ANY end target on your network, last
> line is for icmp flood to ANY host in your network.
>
>
> Let me know what you think of the above acl for logging icmp/smurf attacks
> and if you think im missing something?
>
>
>
>
> ------------------------------
> *From:* Peter Plak [mailto:plukkie@gmail.com]
> *Sent:* Monday, 21 August 2006 8:09 PM
> *To:* Aamir Aziz
> *Cc:* David Redfern (AU); ccielab@groupstudy.com
>
> *Subject:* Re: ICMP Flooding vs SMURF Attack
>
>
> not 100% sure anymore about syntax of udp I did:
>
> acl 130 permit icmp any 0.0.0.0 255.255.255.0 echo log-input
> acl 130 permit icmp any 0.0.0.0 255.255.255.0 echo-reply log-input
> acl 130 permit icmp any 0.0.0.255 255.255.255.0 echo log-input
> acl 130 permit icmp any 0.0.0.255 255.255.255.0 echo-reply log-input
> acl 130 permit udp any 0.0.0.0 255.255.255.0 eq echo log-input
> acl 130 permit udp any 0.0.0.255 <http://0.0.0.0/> 255.255.255.0 eq echo
> log-input
> acl 130 permit udp any eq echo 0.0.0.0 <http://0.0.0.255/> 255.255.255.0log-input
> acl 130 permit udp any eq echo 0.0.0.255 255.255.255.0 log-input
> permit ip any any
>
> On 8/21/06, Aamir Aziz <aamiraz77@gmail.com> wrote:
> >
> > Peter,
> >
> > You said you lost points, what was ur ACL in the exam, I mean what did u
> > put?
> >
> > Thanks
> > Aamir
> >
> >
> > On 8/21/06, Peter Plak <plukkie@gmail.com> wrote:
> > >
> > > David,
> > >
> > > You're 100% right about missing the unicast replies in the .255 and .0
> > > subnets. But the correct answer is maybe hide in the question.
> > > My exam question was titled "log icmp attack / smurf attack".
> > >
> > > Smurf is often an attack to those .255 and .0 addresses.
> > > ICMP flooding often just to a unicast.
> > > Fragle with udp.
> > >
> > > So again, it's unsure what they want to see.
> > > In my question, I had to only log. My list was not allowed to block.
> > > I did not log icmp to unicast addresses, and i also logged udp. Maybe
> > > that;s why I didn't get the points.
> > >
> > > gr
> > >
> > >
> > > On 8/21/06, David Redfern (AU) <David.Redfern@didata.com.au > wrote:
> > > >
> > > >
> > > > Only problem I see is with the icmp echo-reply lines.
> > > >
> > > > deny icmp any 0.0.0.255 255.255.255.0 echo-reply
> > > > deny icmp any 0.0.0.0 255.255.255.0 echo-reply
> > > >
> > > > If you are the end target of the attack receiving echo-replies from
> > > > the
> > > > reflector network then this echo reply would/could be destined for
> > > > host
> > > > addresses in your network. A server for example the DOS is
> > > > targeting.
> > > > The acl only block echo-replies to the 0 or 255 address, which is
> > > > not
> > > > where the target will be.
> > > >
> > > > Maybe you have to block icmp any any echo-reply coming in but this
> > > > this
> > > > stops you from being able to ping the backbone. Only way around this
> > > > I
> > > > know is to to a reflective acl.
> > > >
> > > > If you get an answer to this one please let me know
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > -----Original Message-----
> > > > From: nobody@groupstudy.com [mailto: nobody@groupstudy.com] On
> > > > Behalf Of
> > > > Peter Plak
> > > > Sent: Monday, 21 August 2006 7:47 AM
> > > > To: Aamir Aziz
> > > > Cc: ccielab@groupstudy.com
> > > > Subject: Re: ICMP Flooding vs SMURF Attack
> > > >
> > > > Is it possible to have a udp with source echo, sourced from the
> > > > network
> > > > (
> > > > x.x.x.0) or broadcast (x.x.x.255)?
> > > > The source udp echo is probably from the reflector, so it's replied
> > > > to
> > > > the destination network or broadcast I would presume.
> > > >
> > > > Then I would say for the udp streams it's:
> > > >
> > > > deny udp any 0.0.0.255 255.255.255.0 eq echo deny udp any 0.0.0.0
> > > > < http://0.0.0.255/> 255.255.255.0 eq echo deny udp any eq echo
> > > > 0.0.0.255 255.255.255.0 deny udp any eq echo 0.0.0.0 <
> > > > http://0.0.0.255/>
> > > > 255.255.255.0
> > > >
> > > > gr
> > > >
> > > >
> > > >
> > > >
> > > > On 8/20/06, Aamir Aziz < aamiraz77@gmail.com> wrote:
> > > > >
> > > > > Yes i agree with you that the UDP source is missing here, but the
> > > > > question is what is most suitable or lets say what is required in
> > > > the
> > > > > lab, how about if we go for something like this:
> > > > >
> > > > > deny icmp any 0.0.0.255 255.255.255.0 echo deny icmp any 0.0.0.0
> > > > > 255.255.255.0 echo deny icmp any 0.0.0.255 255.255.255.0echo-reply
> > > > > deny icmp any 0.0.0.0 255.255.255.0 echo-reply deny udp any
> > > > 0.0.0.255
> > > >
> > > > > 255.255.255.0 eq echo deny udp 0.0.0.255 255.255.255.0 eq echo any
> > > > > deny udp any 0.0.0.0 255.255.255.0 eq echo deny udp 0.0.0.0
> > > > > 255.255.255.0 eq echo any permit ip any any
> > > > >
> > > > > this one makes any sense?
> > > > >
> > > > > Thanks
> > > > > Aamir
> > > > >
> > > > > > >
> > > > >
> > > > >
> > > > > On 8/20/06, Peter Plak < plukkie@gmail.com> wrote:
> > > > > >
> > > > > > Hi Aziz,
> > > > > >
> > > > > > I have also spent lot of time to this task. I found a link which
> > > >
> > > > > > enters the explanation of smurf / fragle and protection best so
> > > > far.
> > > > > >
> > > > > > http://www.windowsecurity.com/whitepaper/Characterizing_and_Tracing_
> > > >
> > > > > > Packet_Floods_Using_Cisco_Routers.html
> > > > > >
> > > > > > <http://www.windowsecurity.com/whitepaper/Characterizing_and_Tracing
> > > > > > _Packet_Floods_Using_Cisco_Routers.html+>
> > > > > >
> > > > > > If I look at your list, I would say, almost there. What in my
> > > > > > opinion misses is the udp source eq echo.
> > > > > > I would replace the udp lines with any any. Cause udp echo is
> > > > rarely
> > > >
> > > > > > used nowadays, it's likely that you will have many hits compared
> > > > to
> > > > icmp.
> > > > > >
> > > > > > So, I think the list totally will be then:
> > > > > > deny icmp any 0.0.0.255 255.255.255.0 echo deny icmp any
> > > > 0.0.0.0
> > > > > > 255.255.255.0 echo deny icmp any 0.0.0.255 255.255.255.0echo-reply
> > > >
> > > > > > deny icmp any 0.0.0.0 255.255.255.0 echo-reply deny upd any
> > > > any eq
> > > > > > echo deny upd any eq echo any permit ip any any
> > > > > >
> > > > > > What you think?
> > > > > >
> > > > > >
> > > > > > On 8/20/06, Aamir Aziz < aamiraz77@gmail.com > wrote:
> > > > > >
> > > > > > > Hi there ppl
> > > > > >
> > > > > > I just wanted to clear something, if the tast says that certain
> > > > > > router is experiencing attack via ICMP and UDP flooding does it
> > > > mean
> > > >
> > > > > > SMURF ATTACK?
> > > > > >
> > > > > > and would the following ACL work to mitigate this flooding
> > > > issue?
> > > > > >
> > > > > > deny icmp any 0.0.0.255 255.255.255.0 echo deny icmp any 0.0.0.0
> > > > > > 255.255.255.0 echo deny icmp any 0.0.0.255 255.255.255.0echo-reply
> > > > > > deny icmp any 0.0.0.0 255.255.255.0 echo-reply deny upd any
> > > > > > 0.0.0.255 255.255.255.0 echo deny upd any 0.0.0.0 255.255.255.0echo
> > > >
> > > > > > permit ip any any
> > > > > >
> > > > > > Thanks
> > > > > > Aamir
> > > > > >
> > > > > >
> > > > ____________________________________________________________________
> > > > > > ___ Subscription information may be found at:
> > > > > > http://www.groupstudy.com/list/CCIELab.html
> > > >
> > > > _______________________________________________________________________
> > > >
> > > > Subscription information may be found at:
> > > > http://www.groupstudy.com/list/CCIELab.html
> > > >
> > > >
> > > > ******************************************************************************
> > > > - NOTICE FROM DIMENSION DATA AUSTRALIA
> > > > This message is confidential, and may contain proprietary or legally
> > > > privileged information. If you have received this email in error, please
> > > > notify the sender and delete it immediately.
> > > >
> > > > Internet communications are not secure. You should scan this message
> > > > and any attachments for viruses. Under no circumstances do we accept
> > > > liability for any loss or damage which may result from your receipt of this
> > > > message or any attachments.
> > > >
> > > > ******************************************************************************



This archive was generated by hypermail 2.1.4 : Fri Sep 01 2006 - 15:41:58 ART