From: mr at german-secure.de (Marko Rogge | German-Secure)
Subject: AW: Vulnerability ZoneAlarm Pro 4.5.532.000

Hi Mr. La Cour & Readers!

You are missing one fact, which we didn't explicitly state, which is that
the target system was connected through a normal DSL home user's connection
with a bandwidth of
768 kb/s. So, yes, it is clear that the available bandwidth was exceeded in
any case, with or without Zone Alarm. However, as our detailed test report
shows, with Zone Alarm handling that bandwidth of a 768 kb/s stream of
incoming UDP DoS packets, the Zone Alarm was already using close to 100% CPU
on our high-end system, and became unresponsive, and this especially with
tiny UDP packets to random ports apart from other attacks, so we clearly
_have_ a performance flaw of Zone Alarm here.

Marko Rogge / german-secure.de
Mixter / Securityspecialist 

>>|-----Urspr?ngliche Nachricht-----
>>|Von: John LaCour [mailto:jlacour@...elabs.com] 
>>|Gesendet: Donnerstag, 29. Januar 2004 22:07
>>|An: mr@...man-secure.de; full-disclosure@...ts.netsys.com
>>|Betreff: RE: [Full-Disclosure] Vulnerability ZoneAlarm Pro 
>>|4.5.532.000
>>|Vertraulichkeit: Pers?nlich
>>|
>>| 
>>|-----BEGIN PGP SIGNED MESSAGE-----
>>|Hash: SHA1
>>|
>>|Zone Labs response concerning a reported Denial of Service 
>>|vulnerability in ZoneAlarm Pro v4.5.532.
>>|
>>|Zone Labs is aware of a reported Denial of Service 
>>|vulnerability in ZoneAlarm Pro v4.5.532 as reported by Marko 
>>|Rogge of German-Secure on the Full-Disclosure mailing list 
>>|on January 28th.  We first received this report on Tuesday 
>>|January 27th.  
>>|
>>|Zone Labs has reviewed the test results presented by Mr. 
>>|Rogge and used a similar methodology to try and reproduce 
>>|his findings.  We were unable to do so and, as a result, we 
>>|do not believe that Mr. Rogge's tests indicate that there 
>>|are any vulnerabilities in ZoneAlarm Pro or other Zone Labs 
>>|products.  
>>|
>>|In our own testing, using similarly configured systems, we 
>>|do see an increase in CPU utilization at higher packet rates 
>>|- up to approximately 20%.  However, in no cases does the 
>>|system become unresponsive.  Additionally, the firewall 
>>|continues to perform its job of allowing or denying traffic 
>>|based on the configured policy.
>>|
>>|Zone Labs would also like to point out the connection speed of
>>|55 Mbps in the test case reported is 50 to 500 times the 
>>|bandwidth available to a typical broadband user.  In 
>>|real-world scenarios, a user's bandwidth would be exhausted 
>>|prior to the network traffic having a significant impact to 
>>|ZoneAlarm Pro.
>>|
>>|Additionally, Mr. Rogge and Mixter did not report the 
>>|results of the system when the ZoneAlarm firewall was not 
>>|present.  At extreme data rates any system's performance 
>>|will be impaired by a denial of service attack regardless of 
>>|the presence of ZoneAlarm Pro.
>>|
>>|In summary, ZoneAlarm Pro users are not vulnerable to a 
>>|denial of service attack as a result of using ZoneAlarm Pro, 
>>|nor can a denial of service attack be used to circumvent 
>>|ZoneAlarm Pro's protection.
>>|
>>|Zone Labs takes security vulnerability issues very seriously 
>>|and welcomes the opportunity to work with the security community.  
>>|While we appreciate Mr. Rogge bringing the matter to our 
>>|attention, we ask that all security researchers contact us 
>>|on security@...elabs.com (as mentioned in all of our 
>>|security advisories), and that in accordance with industry 
>>|practice, we be given up to 7 days to respond before any 
>>|issues are disclosed publicly.  In all cases, Zone Labs will 
>>|make every attempt possible to acknowledge the report within
>>|48 hours.
>>|
>>|John LaCour
>>|Zone Labs
>>|Security Response Team Manager
>>|security@...elabs.com
>>|
>>|-----BEGIN PGP SIGNATURE-----
>>|Version: PGP 8.0.2
>>|
>>|iQA/AwUBQBl2DqeZbSyAsADEEQImwACg/UWJ64y+IAgs1Nr5I8hTgHcAnzgAoLwu
>>|/axIMKc6zI27IdW4DwrJXCQd
>>|=IXFN
>>|-----END PGP SIGNATURE-----
>>|

Powered by blists - more mailing lists