Where is pfSense support for HTTP/3 and QUIC protocol support?
-
@mer Yep which is why i think most would agree focusing on the endpoint through some type of software installed is the best way to go. I know my company uses Sophos endpoint agent along with FireEye's on the laptops. The sophos client is what is doing the URL control and application filtering. Even though Palos are used all over the environment and its a NGFW, its not at all feasible to do this type of control on the firewall. Defense in depth.
But to stress, to a certain degree, if we are going to expand the definition of a NGFW than you can argue to some degree, that pfsense is that - it is, after all, a UTM. But everyone has there own definition of NGFW.
-
@michmoor said in Where is pfSense support for HTTP/3 and QUIC protocol support?:
has there own definition of NGFW.
True - but without any packages is just a traditional stateful firewall providing inbound and outbound inspection of the state, etc.
While it can do say ips/ids with the addition of a package.. What those packages might do with quic would be up to those packages, and not really pfsense.
UTM is another term.. again if I throw IPS package on it I can now call it my NGFW UTM ;) which all just words without meaning without understanding context..
Still what exactly does OP feel should happen here.. If your curious what the IPS packages are going to do with quic, that they haven't already done in the last 10 years. Should prob ask @bmeeks
But to me this thread doesn't make much sense - unless your just asking for the web gui to be served up via quic? There is nothing for the pfsense the traditional stateful firewall it is to do to "keep up" or not "fall behind"..
-
@johnpoz said in Where is pfSense support for HTTP/3 and QUIC protocol support?:
But to me this thread doesn't make much sense
yep and this is not to belittle the OP but first and foremost as the original question was confusing.
At the end of the day, the majority of communications on a network are going to rely on ports. Allow or Block ports as required. The payload is irrelevant if it's encrypted and if it's not encrypted then you have IDPS systems that will scan the payload and work on defending your network.I think we can all see how UTM or NGFW or whatever term comes next, its really the marketing teams who have won this and confused us all :)
-
@johnpoz said in Where is pfSense support for HTTP/3 and QUIC protocol support?:
OP is asking for http/3 quic support in pfsense - for what? Serving up the web gui? There is nothing pfsense can do now that gets removed with quic udp over whatever port..
Yes, this. I could imagine it makes Squid more obsolete. That's about all.
-
@stephenw10 said in Where is pfSense support for HTTP/3 and QUIC protocol support?:
it makes Squid more obsolete
heheh - yeah true, but squid has be obsolete for what 10 years anyway..
-
Largely, yes. And encrypted SNI will only make it more so.
It still has a place in some specialist deployments. And in those you can just block udp 443.
Steve
-
@johnpoz I was not expecting much other than to start a dialog as the reality is the f/w world had to adapt (note I didn't say "adopt") to the new reality.
This has been in motion for a decade and just was ratified and an increasing number of content providers are using it.
I didn't know if QUIC was on the roadmap or not and given that Wireshark is STILL evolving to handle it (other than just dump the raw data) to perhaps build a stream/session table to show the "sessions" in progress.
My expectations for pfSense would be to provide a little diagnostic data in terms of percentage of traffic it's using, the destination hosts, and any stream/session tabular data which is in the unencrypted part of the header.
By definition, the majority of the header is encrypted not just the payload so I wasn't expecting much. However, pfSense is also known as an IDP/IPS platform with the addition of several packages.
So if there could at the very minimum get some summary traffic and percentage of traffic, perhaps as another item on the traffic graphs, that would be helpful.
Again, I didn't know if there were any plans, so I simply asked the question to get a baseline of where we are.
-
@lohphat any of that stuff would have to come from packages would be my take.
So again I would tag @bmeeks to throw in his take - he is the ips/ids guy to be sure.
To be honest most everything becoming encrypted, not something new with quic - ips and ids is becoming less and less if you want my take on it.
Not much ips/ids can do from an encryption tunnel - there isn't much to look at to see if something bad is happening inside that tunnel.
Not much the ips/ids can tell from just looking at packets without the payload..
-
@johnpoz said in Where is pfSense support for HTTP/3 and QUIC protocol support?:
Not much the ips/ids can tell from just looking at packets without the payload..
You'd be surprised. Certain applications have a pattern of behavior which can be deduced by the size and frequency of transmissions. It's a bit of alchemy but there are trends which can be triggers on their own.
-
There is essentially zero chance that Snort 2.9.x (currently what is used on pfSense) will get HTTP/3 or QUIC support. The 2.9.x branch is not getting any new features from upstream. Everything is going into Snort3.
Suricata does have an open Feature Request for incorporating QUIC app-layer support, but nothing has been settled there yet. Here is the link to the latest iteration of the long discussion: https://github.com/OISF/suricata/pull/7095. But this really does not mean much as Suricata could only detect that QUIC was passing, it would not be able to see anything in it. So not really very useful IMHO. I am unaware of anything happening in Suricata with regards to HTTP/3.
As @johnpoz alluded to, IDS/IPS is getting increasingly harder as more and more network traffic gets encrypted. Pretty soon nothing will be visible at the network layer except very basic source/destination info. Any IDS/IPS will have to move to the endpoints (servers, workstations, mobile devices, etc.).
-
@bmeeks said in Where is pfSense support for HTTP/3 and QUIC protocol support?:
ny IDS/IPS will have to move to the endpoints (servers, workstations, mobile devices, etc.).
spot on and exactly as i stated in the previous post..
The way I see it, its like adding ClamAV. Belts and Suspenders to your overall security footprint. Doesn't hurt to have it enabled but to be clear it has little to no impact on defending you at the perimiter. Most of the work is/should be taking place on the endpoint.Ah well, so the debate rages on....
-
Hi,
just a quick question. DNS over QUIC support is coming to Unbound.
There's a chance it will be in the upcoming 1.16.3 release. Does Netgate plans to support this via GUI? -
@netblues Well, I'm trying to be pragmatic.
A f/w's job is to monitor and control traffic. Given that QUIC makes inspection a moot point what the f/w CAN do is help control where those packets come from.
e.g. The DoH settings in pfBLocker-devel to permit DNS over HTTP is a good example of how to control a type of traffic. A list of well known sites from which the f/w admin can select to permit/block traffic.
Since QUIC's strength is with streaming/complex websites, I think a page for QUIC which allows to admin to permit QUIC traffic from well known sources would be a reasonable first step. e.g. Google/YouTube, Facebook, Amazon, (and their CDNs).
Let the 80/20 rule stand and let the protocol do its magic where needed but force fallback to TCP for unknown/all other sites.
Yes, this can be done in the f/w ruleset page, but making a dedicated QUIC UI control page would be much friendlier. Perhaps this is a job for a package, but IMHO it's a low-level protocol issue and should be handled by pfSense.
-
@lohphat said in Where is pfSense support for HTTP/3 and QUIC protocol support?:
but IMHO it's a low-level protocol issue and should be handled by pfSense.
Well the same could be said for any cloud provider using 443 currently over tcp. So you want a feature to put in what CDNs are allowed for rules you create.
So in general if i create a rule - you want a drop down list to only allow to specific CDN ASNs that can be picked from a drop down.
-
@johnpoz Perhaps.
I've just started digging into QUIC and logging accesses so I need to be smarter about the scope of the requests. e.g. Is the QUIC request going to the FQDN of the content source (e.g. YouTube) or the CDN? If it's the CDN then was there an initial QUIC request to the FQDN, then a session ID created then subsequent QUIC requests go to the CDN? I don't know yet. I'm acknowledging my ignorance thus the reason I posted my question in the first place.
How would pfSense enumerate the CDN ASNs unless it were running BGP?
-
@lohphat said in Where is pfSense support for HTTP/3 and QUIC protocol support?:
IMHO it's a low-level protocol issue and should be handled by pfSense.
why?
does any other big box vendor do this? -
@michmoor QUIC was just ratified, so I don't know.
The world is not static. Things change.
-
@lohphat well the point of my question wasn't about ratification - quic has been around for quite some time - it's if other firewalls provide the ability to filter from which sources you want quic enabled or not.
Overall, the requst for pfsense and really for any other vendor is moot because there is no need to do this at all. as been suggested, block udp/443. -
@michmoor All or nothing is not acceptable and a lazy solution.
QUIC may be desirable for known sources and not for others. That's what I'm asking: if there's a middle ground where well known, trusted sources could be MORE EASILY accepted since QUIC is much more efficient, while blocking other sources.
BTW QUIC is not the same as gQUIC (Google QUIC) when it was first developed in 2012 -- the ratified RFC version is its own thing.
-
@lohphat said in Where is pfSense support for HTTP/3 and QUIC protocol support?:
trusted sources could be MORE EASILY accepted since QUIC is much more efficient, while blocking other sources.
There is no built in feature like that - you would either have to create your own lists for your rules, or use something like pfblocker to create the aliases for you.
If you are filtering outbound to only trusted destinations, or more trusted destinations then you should already have these rules in place be it they use normal https over tcp, or over udp - not really sure how the trust changes be it tcp or quic to be honest.
Why would I allow a client to destination X IP via tcp 443 but not udp 443?