Re[2]: SYN-scan and TCP-connect scan time difference.

[Anton Konvalyuk <w.o.l.f.paradox () mail ru>]

 Dan,

I didn't find <taskbegin> in the XML output. But I tested the problem one more time (thank you). So:

sudo nmap --open -T4 -F -n -oA report xxx.xxx.xxx.xxx/18
16384 IP addresses (5472 hosts up) scanned in 10995.48 seconds

nmap --open -T4 -F -n -oA report_2 xxx.xxx.xxx.xxx/18
16384 IP addresses (163 hosts up) scanned in 158.18 seconds

Ndiff showed 20 more hosts with open ports (--open option) in the first case. There two hosts with 80 open port. Don't 
know why, because nmap sends packets on this port even in TCP-connect case. Then I captured some packets in both cases. 
Nmap sends packets on 80, 443 ports to identify live hosts in TCP-connect one and packets on 80, 443 ports and ICMP 
Echo request in the SYN-scan case. 

And now I don't know why TCP-connect misses a lot of live hosts. I thought it got any requests from closed 80 and 443 
ports.


Четверг, 23 января 2014, 12:00 -06:00 от Daniel Miller <bonsaiviking () gmail com>:
> On 01/23/2014 10:33 AM, Anton Konvalyuk wrote:
> > Hello.
> >
> > I was scanning a large network (/18). I've noticed strange  occasion. When I use 'sudo nmap --open -T4 -F -oX report 
> > xxx.xxx.xxx.xxx/18', scanning lasts for more than 2 hours. But if I use  'nmap --open -T4 -F -oX report 
> > xxx.xxx.xxx.xxx/18' it takes approximately 2 minutes. And no big difference when using '-n' option.
> >
> > Could you tell what the reason is? The only information I've found is  http://seclists.org/nmap-dev/2006/q1/370 . So 
> > why is TCP-connect faster than SYN-scan? And why is the difference really big?
> >
> > Thanks!
> > Nmap version: 6.00
> > OS: Debian 6.0 x86_64
> Anton,
>
> I am surprised that you are finishing a /18 scan in 2 minutes, no matter 
> what options you choose. Have you compared the results to be sure that 
> you are not losing data? Against a large network without many hosts, the 
> host discovery phase would possibly be the source of delay. With root 
> privilege, Nmap sends 4 probes to determine whether a host is alive, but 
> only 2 probes without privileges. Have a look at the <taskbegin> and 
> <taskend> elements in the XML output to see how long each phase of the 
> scan took.
>
> Dan


-- 
С уважением,
Конвалюк Антон
_______________________________________________
Sent through the dev mailing list
http://nmap.org/mailman/listinfo/dev
Archived at http://seclists.org/nmap-dev/