UTM (Home) EOL and moving over to Sophos Firewall... options?

Notes on installing Firewall... thinking I had not done it right etc x 5 - (rufus usb iso)

SW-19.5.1 ISO boots to display the following after an install
"   Booting '19_5_1_278'
 _
"
Yep thats it....

So in actual fact you need to connect to an AP (so wireless router with DHCP disabled,  plugged into one of the LAN ports) to access the box.
Figuring out what ports on the firewall are LAN/WAN is you next step...

The firewall interface is a "shotgun splatter" of activity and config.

It seems to be doing a great job as a "firewall" but setting up user profiles and devices? WTF?

vicegod said:

The firewall interface is a "shotgun splatter" of activity and config.

So true!

Any updates? How is the migration going for you? I'm still leaning towards migrating to one of the *sense's. I've been "playing" with several FW's and XG is just not a top pick for me. I'm hoping ease of use and feature parity with the UTM gets better by 2026. Fingers crossed.

Well, watched some Vids on Youtube... don't think thats what  i want.

With *sense, untangle, etc. You provide your own hardware, upgrade as neeed.  Untangle does have an annual fee associated.  *sense does not (for home user).

I've decided to go with pfsense+. It may not have all the features of web scanning that xg does but offers far more config granularity. There are some things about it that irk me, such as the lack of integration between modules.  IE, in utm you define an object once and can use it anywhere appropriate. Such is not the case with *sense.

Re pf vs opnsense, feature wise are very similar, but ui is quite different. Can't wrap my head around opnsense's ui layout. Reminds me of the old tomato firmware for wrt* series.

As for ubiquity, if one already has hardware, why reinvent the wheel by buying more hardware.

Jay Jay said:

Can't wrap my head around opnsense's ui layout.

I tried OPNsense awhile ago. I was not fond of the UI either, and the default allow firewall rules that allows everything outbound by default is really dumb. I never tried pfSense yet, but since Zenarmor removed support for it, I can't see it making a good next-gen firewall unless you maybe installed Zenarmor on a Linux distro in Layer 2 bridge mode "with filtering" in a separate VM or on a separate device with 2 or more NICs, which sounds interesting.

EDIT: so it seems that free Zenarmor can still be installed in pfSense manually, you just need to create the Zenconsole Cloud Portal. Is it worth it?

The question is, is this protection really needed for a home user.

According to docs.opnsense.org/.../zenarmor.html,  the features it provides include;

At present, I don't have https inspection enabled in utm as that tend to break a number of sites. Depending on one's household, #1 might be useful, not here. A lot of things can be blocked by using pfblockerNG. Of course, all devices also have some form of security locally (AV/firewall, etc).

In this link, https://www.zenarmor.com/docs/installing/installation

They talk about zenarmor working with the CE (community edition) of pfsense but not pfsense+ (their commercial product).

I host a lot of stuff going back to the days when "next gen" firewall weren't the buzz words that they are today. I've always used the UTM's WAF. It successfully filters a lot of bad traffic and it has its share of false positives so the rules need to be monitored and appropriate exceptions need to be added.

One thing that I've never gotten a clear answer on is... Since the UTM's WAF terminates SSL, doesn't  that mean that the UTM is scanning SSL ingress traffic? Does anyone know the definitive answer to this? This question is what's holding me back from giving *sense a deep dive and waiting to see what XG looks like just before the UTM EOL hits. I'm not very concerned about scanning SSL egress traffic (should I be?) because I trust the local users but I am concerned about what's trying to get into my open ports.

I think most security experts argue that a "next gen" firewall with SSL inspection is a must if you host websites and web apps but is it really?

Free is best but I would pay $120 a year for Zenarmor if it were not for the 3 filtering policy limit and the 100 device limit. I could live with the 100 device limit but having only three filtering policies sucks if you host more than three services.

So for us hobbyists that host email, websites and web apps, who can tell us if *sense with HAProxy provides enough protection without Zenarmor?

It's a real pitty.

I have found nothing so far, that could replace the UTM 9 and the clock is ticking.

Jeff, my understanding is WAF does scan ingress traffic based on the firewall policy defined under webserver/firewall profiles.

Good question re haproxy. From what I was able to briefly search, it's primarily a reverse proxy/load balancer. It may be possible to integrate it with snort/suricatta to scan the traffic as well. This needs more research for sure.

WolfgangS said:

I have found nothing so far, that could replace the UTM 9 and the clock is ticking.

Just wondering, is there a specific feature you need besides an easy-to-use UI that the UTM has? Does the XG simply not have the feature you need.... serious question.

 How difficult Sophos makes it to do simple things can make a person livid. The new version 20 coming out barely address the quality of life issues that are turning people away from the product.

 Things like

difficulty of creating static IP and MAC hosts,

lack of a searchable list of  host names  and IP addresses available to use in the web filtering exception Source list, 

truncating words with ellipsis "..."  in the UI so the user can't even fully read it. 

Version 20 is about cramming in more features without fixing the UI flaws first.

So i was looking at your list: 

difficulty of creating static IP and MAC hosts --> Something which is on the backlog to improve but bigger customers do not use it for two reasons: Either they have a DHCP Server and use DHCP Relay or they use only the User Mapping for Firewall rules (So you use SFOS + User Auth and your firewall rules are LAN - GroupA to WAN and not the DHCP Static User like in UTM). But this request is interesting for smaller customers, who do not have authentication enabled or do not use an external DHCP server. 

lack of a searchable list of  host names  and IP addresses available to use in the web filtering exception Source list --> You are talking about the Web Filtering Exceptions? So just curious, you are looking at the Web Filter Exception - Why not using DPI Engine instead? Which makes most of the Web Filter Exceptions obsolete in itself. 

truncating words with ellipsis "..."  in the UI so the user can't even fully read it. --> Looking at the V20.0 changes with the FullHD resolution, you should have more space and most tables in UI shows the full context. There are still some things, which are being looked at to get more context in it, but most tables show the full values. 
Showing the full context of some objects is still to do, but there were progress on that front. 

So TLDR: Are those the points, UTM to SFOS is not "suitable" for Home Users?  

Jay Jay said:

Jeff, my understanding is WAF does scan ingress traffic based on the firewall policy defined under webserver/firewall profiles.

Good question re haproxy. From what I was able to briefly search, it's primarily a reverse proxy/load balancer. It may be possible to integrate it with snort/suricatta to scan the traffic as well. This needs more research for sure.

Yes, this is all true. The question is does either or both the UTM and *sense perform the scan before or after the traffic is decrypted. I'm assuming after but I've not been able to confirm.

Jay Jay said:

Jeff, my understanding is WAF does scan ingress traffic based on the firewall policy defined under webserver/firewall profiles.

Good question re haproxy. From what I was able to briefly search, it's primarily a reverse proxy/load balancer. It may be possible to integrate it with snort/suricatta to scan the traffic as well. This needs more research for sure.

Yes, this is all true. The question is does either or both the UTM and *sense perform the scan before or after the traffic is decrypted. I'm assuming after but I've not been able to confirm.