unknown-tcp / udp - please explain

[cryptochrome]

Hi,

I know that these two applications stand for unrecognized traffic. It worries me though that for some of the other applications to work, I have to add unknown-tcp/udp to the firewall rule. Example for this would be Bittorrent traffic. To allow Bittorrent, I also have to allow web-browsing and unknown-tcp and unknown-udp.

Can someone please elaborate on this? If I only want to allow Bittorrent, but also add web-browsing and unknown-tcp, I will open up the firewall for unwanted traffic. I really have a hard time understand this concept.

Thanks

[kfindlen]

Allowing unknown-tcp/udp to allow BitTorrent traffic should not be required.  On my device I have utilized BitTorrent with and without encryption over the last few weeks and the traffic logs show that none of the sessions are being identified as unknown-tcp/udp.  It's possible this issue is specific to the torrent you are accessing. 

The web-browsing component could be required for the tracker communication which can utilize HTTP. 

If you see this issue on the latest content then I would recommend opening a support case for investigation.

[NGS_SOC]

Bittorent doesn't depend on unknown tcp/udp, only web-browsing on tcp/udp dynamic ports. If you have 5.0.x this dependence is already done, otherwise a rule has to be inserted for allowing web-browsing before bittorent.

Verify the app dynamic update (latest 373-1793) and in case of other error/warning  during commit also I suggest opening a support case.

[cryptochrome]

Ok, you are right, there are no warnings anymore about uknown-tcp on commit. However, you are saying that 5.0 automatically resolves those dependencies, does that mean it will actually include the needed services without me specifying them in the rule? That would mean it will still open unknown-tcp/udp.

Bittorrent was just an example. I've seen this dependency with other apps as well. If other apps rely on unknown-tcp/udp, doesn't that make the whole thing completely insecure? I am opening up the firewall for unknown traffic.

[NGS_SOC]

In PANOS 4.x all application dependencies have to be explicit allowed in security rules, otherwise warning during may appear and related application could not work properly. Sometimes in large scale this requirement could be annoying or worse.

Version 5.0.x changes this behavior allowing application dependencies if they are granular web-browsing, ssl, ftp and few more. Never unknown traffic, if needed, this have to be allowed with an explicit rule.

[cryptochrome]

So that means, if there are dependencies:

   1.) It will resolve them automatically and add the needed services, invisible to the user.

   2.) Because of that, I don't what what I actually allow through my firewall.

Excuse my ignorance, but are you guys serious?

[NGS_SOC · Accepted solution]

I quote what is reported in PANOS v.5 release note:

Application Dependency Enhancement – For some protocols, you can allow an application in security policy without explicitly allowing its underlying protocol. This support is available if the application can be identified within a pre-determined point in the session, and has a dependency on any of the following applications: HTTP, SSL, MSRPC, RPC, t.120, RTSP, RTMP, and NETBIOS-SS. Custom applications based on HTTP, SSL, MS-RPC, or RTSP can also be allowed in security policy without explicitly allowing the underlying protocol. For example, if you want to allow Java software updates, which use HTTP (web-browsing), you no longer have to allow web-browsing. This feature will reduce the overall number of rules needed to manage policies.

This means that few applications can use this enchantment and you never allow unwanted applications. Be aware of how PA recognize application: for example application facebook relies on web-browsing because before facebook the frewall recognize in fact web-browsing app. So the programmer ask themself why not having an implicit application allowing?

Please do not mix together app and service (ports) these are different variables in security rules, as an advice try to use always application defaults as policy enforcement.

If you keep log option for security rule you are always able what traversed the firewall. Also always in session browser (cli/gui) you can see which kind of app traffic is flowing even with a permit all policy, this is the strength of these devices.

[K_Celenza]

Is this the same cryptochrome from the infamous why "NSM is a piece of crap" forum? I happened to be the first one to reply to that post.

[cryptochrome]

yep. that's the same Cryptochrome :smileygrin:

[mikand]

Slightly off-topic but I guess this is the thread you both are refering to ? 🙂

Want some examples why NSM is a piece of junk? - J-Net Community