Plenty of teams still treat a forwarding firewall as a finished control. If the tunnels come up and the policy still matches, plenty of cybersecurity programs still call the perimeter “fine” until next year’s hardware cycle. That assumption just ran into a vendor advisory that tells you to pull the box off the internet now.
On Wednesday, SonicWall told customers still running Gen 6 and Gen 6.5 appliances on SonicOS 6.x to replace them immediately and to stop using them at the internet edge. Those platforms have reached end of support. There will be no more firmware and no more security patches, “regardless of severity.” If that appliance is still doing your threat-protection at the WAN, you are running an unmaintained service with a public IP.

The Vendor Already Walked Away From That Box
Vendors have been polite about sunsets for years. You got a PDF, a date, and a quote for the next chassis. This note is blunter. SonicWall calls the remaining Gen 6 fleet a high-risk exposure and wants internet-facing units gone. Read it the way you’d read a threat bulletin. The company that ships the firmware is telling you it will not fix the next critical bug.
You already know why these boxes linger. Site-to-site tunnels. SSL VPN. NAT for a plant that still has a printer on a public map. Sometimes the only MFA integration you’ve got. Sometimes a HA pair in a colo that needs a drive and a crash cart to touch. They still pass traffic. Traffic is a terrible health metric for cyber security. An abandoned OS can route packets all day while the VPN parser, the SSL stack, and the admin UI age in public.
Attackers do not need a branded campaign with your vendor’s name on the slides. Unsupported edge devices collect brute-force against admin paths, credential stuffing on VPN realms, and exploit attempts the hour a new bug is public. You will not get a patch. Security hardening on the device itself is finished, because firmware is the control you no longer have. What you still control is placement. Get it off the edge.
Defense in depth is the phrase that usually shows up after this kind of advisory, as if extra internal rules make an unpatchable perimeter honest. Keep the internal controls. Also admit what broke: the outer control is now a standing exception. Every WAN-facing management port, every SSL VPN portal, every “temporary” port forward on that generation is a bet you cannot update.
Threat-intelligence reporting this week put Africa back on top for regional attack volume. Nigeria more than doubled year over year. Angola and Kenya landed in the global top five. You do not need a Lagos office for that to matter. Commodity scanners follow firmware families and open VPN ports, not borders. If your WAN edge is a retired OS, you share a blast radius with every other shop that deferred the same refresh.
Your Cybersecurity Console Patched 140 Flaws This Week
While you’re arguing about chassis lead times, the platform you use for threat detection just absorbed a patch bomb. Splunk published fixes for nearly 140 flaws across three core products. The pile includes critical and high-severity issues that add up to authentication bypass, remote code execution, and privilege escalation. That’s the console your incident response team lives in. That’s also the place an attacker goes when they want to edit the story of the firewall they just walked through.

A lot of programs will park a dump that size in the next “hardening window.” Authentication bypass and RCE on a log aggregator do not honor freeze calendars. If search heads, deployment servers, or forwarder management are reachable from anywhere but a tight admin network, treat the release as a security event. Privilege escalation on the box that holds privileged logs is how an intrusion becomes a reconstruction problem.
You wanted the SIEM watching the firewall cutover. Patch the SIEM on the same sprint, or the cutover is theater. Operations that can see a VPN brute-force only after someone else can rewrite the index are collecting telemetry. They are not doing threat detection.
This is the ugly pairing in one week: an edge generation that will never get another fix, and a monitoring stack that just admitted it had a lot of ways to lose the plot. Your job is not to rank which vendor embarrassed you more. Your job is to stop treating either layer as furniture.
Take the Old Edge Off the Internet Before the New One Arrives
Replacements take weeks. Some take a quarter, once you count change boards, circuits, and the one person who remembers the VPN realm mapping. Exposure reduction does not have to wait for the truck. Get the unsupported OS off the public internet even if the new appliance is still a purchase order.
- Inventory every internet-facing and DMZ firewall, VPN concentrator, mail gateway, and “temporary” NAT box. Record vendor, hardware generation, OS train, last patch date, and whether management or VPN is exposed. Include HA standbys and lab units that grew a public IP. Gen 6 and Gen 6.5 on SonicOS 6.x are the urgent set this week. Run the same pass for every other end-of-support edge device you have been politely ignoring.
- Remove end-of-support devices from the internet edge. Put a supported firewall in front, migrate VPN to a current platform, or take the circuit down if the site can bear it. A retired box behind a current one is a different risk than a retired box with a public address. The advisory is about the internet edge. Start there.
- Shrink the management plane the same day. WAN management off. Default HTTPS admin off. No general desktop VLAN path to the GUI. Jump hosts and allow-lists only. For anything that was public, rotate admin credentials, shared secrets, certificates, and VPN user credentials. Disable unused authentication realms so stale accounts stop being a brute-force gift.
- Patch the telemetry path on a short fuse. Apply the Splunk fixes in a compressed window, not the next quarterly bundle. Confirm authentication still fails closed. Look for unexpected admin users, new apps, and saved searches you did not install. Restrict who can deploy apps to the search tier.
- Put end-of-support dates on the same calendar as incident response retainers. Ninety days before EOS, require a written edge plan with an owner. After EOS, internet-facing use needs an exception with an expiry, not a comment in a CMDB. Until the old generation is gone, review VPN and admin failure logs daily the way you’d watch a known-bad port.
Assume the old box was reachable
If a Gen 6 unit spent months on a public IP, walk a short incident response checklist even if nobody has paged you. VPN logs. Admin logins from odd networks. New local users. Unexpected NAT rules or port forwards. Config backups that left the device. You may not get a tidy indicator. You’re looking for a foothold that arrived through a parser nobody will patch.
Keep the internal work honest after the edge moves. Least privilege on jump hosts. Outbound allow-lists so a compromised VPN user does not stroll to domain admin. Segmentation that still holds if the next perimeter bug is yours. That’s the defense in depth you still own when the vendor has left the firmware chat. Security hardening from here is architecture and access, not another checkbox on a dead OS train.
Frequently Asked Questions
- Can a Gen 6 firewall stay in production if it only does internal NAT?
- Internal-only placement is a smaller problem than a public IP, and it is still an unpatchable OS sitting on your network. Put a supported device in front of anything that still faces the internet, then schedule the internal unit for replacement with an owner and a date. Do not leave WAN management or VPN enabled “just in case.”
- Should Splunk wait for the next change freeze if the edge is the emergency?
- Authentication bypass and remote code execution on a log platform are an emergency of their own, especially if those systems are reachable from broad admin networks. Compress the change window, patch, and verify failed logins still fail. You need trustworthy telemetry while you move the firewalls.
- How is end of support different from missing a Patch Tuesday?
- A missed patch is a delay you can still close. End of support means the next critical bug has no vendor fix, “regardless of severity.” That turns every new public exploit against that firmware train into a permanent exposure until you change the architecture.
Sources
- SonicWall urges customers to replace Gen 6 firewalls now
- Splunk Patches Nearly 140 Flaws Across Three Core Products
- Africa Retakes Top Spot as Most Attacked Region, Led by Nigeria Surge
Take Control of Your Server Security
Don't let brute-force attacks slow you down. Try IPBan Pro risk-free for 30 days.
Secure. Automated. Lightweight.
Take Control of Your Server Security
Don't let brute-force attacks slow you down. Try IPBan Pro risk-free for 30 days.
Secure. Automated. Lightweight.
