Re: Snort inline latency

["Abdelmonaim Mokadem" <abdelmonaim.mokadem () abovesecurity com>]

My first objective is to calculate my snort inline latency.

Right now, I'm  trying to determine what is responsible for this high
latency. Since even when the snort inline is down (with a bridge created
with brctl), results remain unsatisfactory, meaning that there is
another problem in my network setup.  The setup you suggested could be a
way to figure out where is the problem,  or a good alternative to
calculate my snort inline latency since it's realized approximately the
same way (just need to encapsulate the whole thing in a VLAN).

 

The pcap injected integrate a high number of machines with different mac
addresses that do not exist on the network. This is why I approached the
problem with a bridge and a snort inline setup to compare the results.
Adding a router in between doesn't seem like a good approach in my
setup. What I'm trying to understand for the moment is what actually is
done by the bridge. The bridge itself seems to be responsible for a part
of the delay, which makes me think that it's partly introduced by the
kernel...  

 

 

Regards,

Abdelmonaim Mokadem.

 

From: Ian Bowers [mailto:iggdawg () gmail com] 
Sent: Monday, May 07, 2012 10:48 AM
To: Abdelmonaim Mokadem
Cc: snort-users () lists sourceforge net
Subject: Re: [Snort-users] Snort inline latency

 

Apologies if I misunderstood your purpose.  The setup I suggested was
only to calculate the difference in latency for when snort is placed
inline vs snort removed from the path.  This would be useful for
determining if your inline snort setup is in fact a source of latency.  

 

Are you trying to determine if snort itself is a hinderance, if the
physical device you've placed inline is a hinderance, or if it is
something else?

 

Regards

-Ian

On Mon, May 7, 2012 at 10:10 AM, Abdelmonaim Mokadem
<abdelmonaim.mokadem () abovesecurity com> wrote:

Hi,

Thanks for the reply.

I'm not sure that putting my snort into a VLAN is gonna bring me
different results from what I actually have...

Can you explain what would putting my IPS into a VLAN change ?

As I understand it, it would only isolate the tests, but I think it
won't change the latency nor help calculate it..

My big problem here is to understand why my latency is so high (nearly
200 milliseconds with clean traffic, even without snort inline activated
results remain unsatisfactory) and to know if my way to calculate this
latency is the right one.

Regards,

Abdelmonaim Mokadem.

 

From: Ian Bowers [mailto:iggdawg () gmail com] 
Sent: Friday, May 04, 2012 4:19 PM
To: Abdelmonaim Mokadem
Cc: snort-users () lists sourceforge net
Subject: Re: [Snort-users] Snort inline latency

 

If you have a VLAN capable switch, use VLANs to place the snort sensor
inline.  run your tests, calculate latency.  Then use VLANs to take the
sensor out of line.  repeat tests.  Do this a couple times to make sure
you have consistent results.   I use this technique all the time in my
lab to toss test appliances or VMs in the path of traffic without having
to mess around with cables like.  IE:

 

Snort inline:

Server --- VLAN 102 ---- snort --- VLAN 103 --- router --- client

 

Snort removed:

Server --- VLAN 102 --- router --- client

   VLAN 104 --- snort --- VLAN 104

 

inline, both sides of snort are on different VLANs, and act as if
they're on their own little switches.  The switchports are access ports
so the end devices don't see the VLAN tags.  In the second scenario,
snort sits in its own little penalty box on VLAN 104 where the two
sensor ports don't talk to anyone.  the router and the server are both
on VLAN 102 at this point and can communicate directly.  I'm assuming a
router is involved but it doesn't have to be.   just eliminate the
router and put the client in its place, works the same.  I pick random
VLANs in the 100s for most examples to discourage people from using
native VLANs by mistake since that may break the scanario.

 

 

Example switch config if you're using Cisco gear:

 

INLINE:

vlan 102

 name servernet_inside

vlan 103

 name servernet_outside

 

interface FastEthernet0/4

 description Server access port

 switchport access vlan 102

 switchport mode access

!

interface FastEthernet0/5

 description Snort HOME_NET port

 switchport access vlan 102

 switchport mode access

!

interface FastEthernet0/6

 description Snort EXTERNAL_NET port

 switchport access vlan 103

 switchport mode access

!

interface FastEthernet0/7

 description Router access port

 switchport access vlan 103

 switchport mode access

 

 

REMOVED:

 

vlan 102

 name servernet_inside

vlan 104

 name on_deck

 

interface FastEthernet0/4

 description Server access port

 switchport access vlan 102

 switchport mode access

!

interface FastEthernet0/5

 description Snort HOME_NET port

 switchport access vlan 104

 switchport mode access

!

interface FastEthernet0/6

 description Snort EXTERNAL_NET port

 switchport access vlan 104

 switchport mode access

!

interface FastEthernet0/7

 description Router access port

 switchport access vlan 102

 switchport mode access

 

 

Regards,

-Ian

 

On Fri, May 4, 2012 at 3:35 PM, Abdelmonaim Mokadem
<abdelmonaim.mokadem () abovesecurity com> wrote:

Hi all,

I am currently trying to calculate my snort inline latency.

I have 2 hosts : one running snort inline and the other one which inject
pcaps (a bridge is set to send back packets from the snort host machine
to the injection one).

I use tcpdump to sniff packets at the emission and at the reception from
my injection host and (I match only ACK packets with a little program)
then calculate the difference between the timestamps (emission and
reception), but unfortunately my results are really bad...

Can anyone tell me if I'm doing right? Or does anyone know what is the
best way to calculate snort inline latency?

Thanks,

AM.

 

 


------------------------------------------------------------------------
------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond.
Discussions
will include endpoint security, mobile security and the latest in
malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Snort-users mailing list
Snort-users () lists sourceforge net
Go to this URL to change user options or unsubscribe:
https://lists.sourceforge.net/lists/listinfo/snort-users
Snort-users list archive:
http://www.geocrawler.com/redir-sf.php3?list=snort-users

Please visit http://blog.snort.org to stay current on all the latest
Snort news!

 

 

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

_______________________________________________
Snort-users mailing list
Snort-users () lists sourceforge net
Go to this URL to change user options or unsubscribe:
https://lists.sourceforge.net/lists/listinfo/snort-users
Snort-users list archive:
http://www.geocrawler.com/redir-sf.php3?list=snort-users

Please visit http://blog.snort.org to stay current on all the latest Snort news!