Help with Threat log SCAN: Host Sweep

[hattracker]

I am looking for assistance interpreting a report that shows “SCAN Host sweep traffic” in my threat log. There are multiple internal sources scanning multiple destination IP addresses that I do not own. The daily number of scans detected from each source is between 2 and 10. The source machine rarely scans the same destination. Is this a low level attack trying to stay under the radar or is there an explanation that does not indicate a problem on my network? I appreciate any feedback.

Type: Scan
Name: SCAN: Host Sweep
From Zone: Inside
To Zone: Outside
Source address: Internal IP address owned by me
Destination Address: various external addresses not owned by me.
Port; 99% of the time 443. Occasionally port 80 or 22222

This screen shot show traffic during one time frame. I see various Source and Destination addresses.

This screen shot is filtered by destination address. This destination address is scanned various times from different Source addresses over the course of several days.

[hattracker]

I see 100 views but no comments.  Is there additional information I can add to help refine my question?

[OtakarKlier]

Hello,

Could be something attempting to beacon out. I would check the hosts for compromise and keep making sure the PAN blocks the traffic for now.

Also check the destination IP's, if they are something that your systems are supposed to be reaching out to.

Regards,

[hattracker]

Thank you for replying.  I am concerned because those addresses are not relevant to our daily business, but each machine is performing only a few sweeps a day.   I am concerned at this traffic, but wondering if I am just being paranoid?

[OtakarKlier]

Hello,

A little paranoia is OK especially if you dont have a baseline. I would investigate the hosts and see what is actually causing the traffic.

Regards,

[hisingh]

Hello Hattracker,

1. Please check your zone protection setting, what is the threshold is set for it?

2. Have you applied for the zone protection on the internal/ trust interface? 

3, Please scan one of the clients with host-based IPS and AV, see if any ports are open. 

Thanks 
Himani 

[prashant.sharma]

Hello,

I have been seeing the same and it is awful lot of traffic. I checked the destination IPs and most of them are Microsoft IPs in my case. However, I did notice some other IPs too which are flagged by some of the threat detection engines. 

The traffic has application 'not-applicable' which tells me there is no real data over the network but it more sort of probing with a packet or two.

I noticed most traffic going out over ports 443/80.

Have you done any packet captures?

Thanks,

[hattracker]

I am sorry, no I do not have any packet captures from those events.  I did some digging at checked out a few machines. At that time I did not find any compromised machines.  It may be a good time to look at those threats again.

[prashant.sharma]

Yes, see if you can get some captures. I would do that myself too but I do not have the approval to do the captures on the target machines. I am running some on FW though as the time of this writing. Will let you now if find something interesting.

[prashant.sharma]

Found a lot of out-of-order and re-transmitted packets.

For me, some of the traffic is going to Microsoft servers with unrecognized text in XML and some other traffic going to some hosted servers which are apparently similar to the line of business. Did not find anything too concerning. 

[mivaldi]

I'd recommend opening a Support ticket for this.

[DelvinC]

I have seen these logs too.

Not been able to figure this, but, I wonder if the alert logs are generated due to source IP meeting the interval and event threshold instead of tracking it per unique source-destinat... pair when configured to alert on the Reconnaissance Protection tab.

Ideally one gets to pick and choose the tracking mechanism when you block but not for alert.

Did anyone make progress with support on this? What did they have to say?

[mivaldi]

Host sweep will detect whenever a source attempts to hit different IP addresses on the same destination port, which if you think of it is by definition internet activity (multiple IP's hit on port 443 and 80). This means that if you enable this protection on an internal Zone with internet access, it is highly likely to trigger FP's continuously for public IP's on the internet on regular internet ports (most frequently 443 and 80).

[trevor_debeer]

I had to reduce the interval and increase the threshold (2 secs and 1000 IPs worked for me and should be good for most scenarios). This way you won't drop normal traffic, but anything firing up more than 500 connections per second to different servers using the same port will get dropped.