VLAN configuration on a "smart" switch for a guest WiFi network


Recommended Posts

So I finally got round to buying a proper access point (Unifi nanoHD), which is capable of having multiple SSIDs, mapped to VLANs.  I have a "smart" switch (TP-Link TL-SG2008), and a PFSense based firewall (SG2220).

 

Assuming that:

1) The AP is connected to the switch on port 1

2) The firewall is connected to port 8

3) There a bunch of other "trusted" devices on the other 6 ports.

4) I want my "trusted" network to run on VLAN 11.

5) I want my "untrusted" network to run on VLAN 99.

 

Is it right that I?

Set up a "trusted" SSID on the AP, which is on VLAN 11.

Set up an "untrusted" SSID on the AP, which is on VLAN 99.

Set up port 1 on my switch to preserve VLANs

Set up port 8 on my switch to preserve VLANs

Set up ports 2 through 6 to tag packets on entry with VLAN 11 and strip VLANs on the way out

 

Does anyone know how to achieve the configuration on my particular switch?

Also what changes do I need to make to pfSense to treat the VLANs as logical interfaces, each with their own subnet, DHCP ranges, internet access and prevent any routing between them?

 

Is there a specific sequence I need to do this all in so I don't lose connectivity to the various components whilst I make the changes?

In your pfsense config, you will want to create your vlans.  What you will do is create a rule to block traffic coming from your guest vlan to your private vlan, it will be able to communicate with all other networks.  

 

You will trunk your vlans on a port (usually you just have to enable trunking, but you can tell it to include those vlans on that trunk port) to the switch.  You will probably have to configure the switch port that is coming from pfsense to trunk.  You will have to configure the switch with the vlan id's that are coming from the pfsense router.  Then you can assign those vlans to ports.  

 

If you have a AP that is vlan capable (**cough** ubiquiti **cough**), you would trunk all of the vlans to that and make your AP default to the private vlan or a management vlan of some sort (so add another vlan for management of devices that can only have source traffic come from your private vlan, another pfsense access control list) to manage networking devices.  Then you can have your AP host both the Guest VLan via a Guest SSID and your Private vlan via Private SSID.

 

 

Draw it out on paper first on how you want things to work.  It will then become clear to you what you have to do and if you will or will not experience an outage.  Understand that you will probably have 1 lan  for internet traffic to go across, 1 lan for house, 1 lan for guest, if any guest or house will need to share devices like printers, another for printers/shared devices, and maybe for the hell of it one for IoT devices.  (fyi, I don't see a way to not have an outage of some sort...you can create the networks and test, but when you move devices over to the new LANs they may have to reboot).

 

On 5/24/2019 at 2:33 PM, sc302 said:

Draw it out on paper first on how you want things to work.

This is very good advice... This will allow you to understand exactly what has to be done, and where.

1 minute ago, sc302 said:

And vlan 2 (or whatever vlan....could be vlan1 but that is just lazy) going between modem and router. Internet traffic isolated on its own vlan to not mix secured from unsecured devices

Can you explain that?  Given that the modem is ISP provided, with no control whatsoever in terms of VLANs, how would I achieve this?

The isp goes into the router. Whatever that is it is a Vlan or untrusted network.  This would be an untagged port. If we are treating this as an outside/untrusted network on a firewall/router nothing else is needed to be done.  

31 minutes ago, BudMan said:

So both of those vlans are tagged or is one native (untagged) and the other tagged? 

Is one of these choices better than the other? If I choose to leave one untagged, would it be the trusted or guest one?

 

32 minutes ago, BudMan said:

At the switch and router? 

Don't understand this question.

 

The picture is meant to be of what I want to end up with, not what I have right now (single SSID, no tagging at all anywhere)

16 minutes ago, Fahim S. said:

The picture is meant to be of what I want to end up with, not what I have right now (single SSID, no tagging at all anywhere)

tagging/untagging is how switches work.  an untagged port is an access port.  An access port is an endpoint port where a device on the other end is essentially dumb and doesn't know the difference between tagged and untagged traffic.

 

Tagged is how a trunk works, this allows the port to encapsulate all of the vlans you choose on a single port.  You can have a native vlan (untagged) and several encapsulated vlans (tagged) on a single port that connects to a switch that can decipher this type of traffic (known also as 802.1Q).  

 

For your ap to work, it will have one port.  That one port can support all of the vlans you want to send over to the AP.  You can have the AP on both a tagged and untagged port...the tagged vlan will be the guest vlan and the tagged/untagged vlan will be your secured vlan.  just like in your picture.  

 

You need to understand the terminology, that is all.  If you don't know ask, don't assume that you aren't tagging anywhere, you have to tag for vlans to function across a single port.

 

edit: so you don't get confused, and being that budman has more time with helping, I will let him work with you.  If you get stuck or need simpler explanation please ask.

As to tagged or untagged doesn't matter which... Its just how you set it up.. Its normally more intuitive on say your router where the actual physical interface network is left untagged.  Vs not putting any network on the physical interface, and only enabling vlans that run on that phy interface.

 

As to switch and router, this is where native or untagged vlan will come in to play. For example out of the box on a switch the vlan 1 is untagged. 

 

On your AP if you do not set a vlan for an SSID, then it would be native untagged... And that would be need to be set on the switch port the AP is connected too.  If you tag both SSIDs on AP with vlan IDs then you would have to set them as tagged on the switch port the AP is connected too.

 

On the interface to the router same thing - if you set both as tagged vlans on your router, then they would both have to be tagged on the switch port that connected to your router.

 

On a port that carries more than 1 vlan, only 1 could be untagged (native) all other vlans would have to be TAGGED... Or all of them could be TAGGED... All depends on the device your connecting to that switch port and how its configured for native or all tagged, etc.

 

To be honest I think tag and untagged is what confuses the most new users to vlans.

11 minutes ago, sc302 said:

edit: so you don't get confused, and being that budman has more time with helping, I will let him work with you.  If you get stuck or need simpler explanation please ask.

Thanks for the offer (I genuinely am grateful), but with the very greatest respect I never find your explanations very "simple".

 

4 minutes ago, Fahim S. said:

Thanks for the offer (I genuinely am grateful), but with the very greatest respect I never find your explanations very "simple".

 

Interesting, but ok.  I do take great pride to simplify things, but completely understandable.  Everyone has different understanding levels, some people require many different approaches until they finally understand (or think they do). 

21 minutes ago, BudMan said:

As to tagged or untagged doesn't matter which... Its just how you set it up.. Its normally more intuitive on say your router where the actual physical interface network is left untagged.  Vs not putting any network on the physical interface, and only enabling vlans that run on that phy interface.

 

As to switch and router, this is where native or untagged vlan will come in to play. For example out of the box on a switch the vlan 1 is untagged. 

 

On your AP if you do not set a vlan for an SSID, then it would be native untagged... And that would be need to be set on the switch port the AP is connected too.  If you tag both SSIDs on AP with vlan IDs then you would have to set them as tagged on the switch port the AP is connected too.

 

On the interface to the router same thing - if you set both as tagged vlans on your router, then they would both have to be tagged on the switch port that connected to your router.

 

On a port that carries more than 1 vlan, only 1 could be untagged (native) all other vlans would have to be TAGGED... Or all of them could be TAGGED... All depends on the device your connecting to that switch port and how its configured for native or all tagged, etc.

 

To be honest I think tag and untagged is what confuses the most new users to vlans.

OK... but in my switch I can set a port (on a per VLAN basis) as Untagged, Tagged, or Not Member. I can also give a port a PVID.  The switch doesn't have an option to set a port as an access port or trunk as such. 

 

I am pretty sure that for VLAN 99 I want to set port 1 and 8 as tagged and the others as Not Member. 

 

But what do I do for VLAN 11? Set them all to Tagged? What PVID should they have?

 

 

So if you put a port in vlan 11, and your going to connect a computer to it then that would be untagged 11 with pvid set to 11... This tells the switch when it sees untagged traffic coming into that port that its vlan 11.

 

When you connect say your router that is using untagged (native interface on the router) and you want that as 11, then same thing untagged 11, pvid 11

 

For the vlan 99 which you run on top of that physical interface, on the switch port it would add tagged 99.

 

For your access point same sort of thing.. if you do not put a vlan ID on one of your SSID that would be the untagged and pvid setting, with the other vlan set to tagged.

 

Your running pfsense as your router?  I can show you some screenshots of what I mean by native and vlan on pfsense.

got it.. thank you!

I decided to keep my trusted network untagged and decided that VLAN 100 would be a better choice for guest.

 

OK.. now the pfSense set up...

I set up a VLAN for 100, and then a (sub)-interface for this VLAN

I then set the interface with a static IP (I used 192.168.100.1 /32).  Kept everything else as default.

When I go to add a DHCP server, I don't even see the tab for my Guest network. 

 

Have I done something wrong?

 

2 minutes ago, Fahim S. said:

/32).

that is wrong!  You prob want /24 which would be 192.168.100.1-254 would be valid IPs on that network.

 

/32 is all 32 bits.. so 192.168.100.1 is the ONLY address.  So can not run a dhcp server on that ;)

Just now, BudMan said:

that is wrong!  You prob want /24 which would be 192.168.100.1-254 would be valid IPs on that network.

this is because /32 is a single IP Address and /24 would be a block of IP addresses? I don't understand how that works..

 

OK.. so now I have a DHCP server, giving out addresses 192.168.100.10 through 192.168.100.100.

Now no matter which WiFi network I get on, I can get out to the internet, which is good, but both networks can see devices in the trusted network.

 

How do I stop this?

Ooops...Completely forgot the firewall rules.

 

I added 2 rules.. an allow all, and a deny access to the trusted network, both to the Guest interface.

Do I need to add a similar deny rule to stop the trusted network being able to access the Guest network?

I can reach the pfSense administrative interface through both networks.  192.168.0.1 on the trusted network and 192.168.100.1 on the guest network.

Is there a way to stop access to this UI from the guest network?

Yeah put in a firewall rule to block it ;)

 

Normally on a guest network it would be pretty locked down..

 

Rules are evaluated top down, first rule to trigger wins, no other rules are evaluated... Post up our rules on our guest vlan interface and we can discuss

 

So what do you want to allow and what do you want to block?  If you just don't want clients to access gui.. Then put a rule above the any rule that says block dest lan address port XYZ, where xyz is the ports (or ports) via an alias that your gui is listening on.

 

example, if your gui just running on 80 (http)

example.thumb.png.7803110dc3e73282ac1f4b02f3583b23.png

 

Keep in mind that such rules would allow guest to actually hit your gui via your wan IP..

 

You could do something like this

otherblocks.thumb.png.51b83a10b4f6cb4e0fc82f1f32c02b86.png

 

So you allow guest to "ping" pfsense guest address. So client can validate they have connectivity to the gateway.

 

But then any other access to firewall is blocked - all IPs, lan, guest, optX, wan, etc.. "this firewall" is a drop down option for dest.

 

This would require clients to be using some outside dns - which is what you normally hand "guest" clients anyway - say 8.8.8.8 for example.

 

Or you could allow clients to use pfsense guest IP for dns and ping - but block all other access

icmpdnswlabels.thumb.png.1ae6bbf33a892c38c2084204340339a9.png

 

Given that when I show "test" on my screenshots you would use your "guest" ;)

 

Since this is local network and not public internet you might want to use "reject" vs just block.. This will tell the client F Off!! Vs letting the client keep trying with retrans, waiting and retrans again.. Client will get told instantly sorry blocked!

reject.thumb.png.37dfce04439b8e59d7d29d33953c019a.png

 

While reject is normally good for your local networks.. You would normally not want to reject any blocks you do from the internet.. Just block (drop) them.. Vs sending any sort of response.

 

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Posts

    • OpenAI's new GPT-5.5-Cyber tops Claude Mythos 5 in vulnerability benchmark by Pradeep Viswanathan OpenAI today announced a major expansion of Daybreak, a cybersecurity initiative designed to help defenders find, validate, and fix software vulnerabilities earlier in the development process. The availability of powerful AI models has definitely changed the cybersecurity landscape by making vulnerability discovery much faster. However, the bigger bottleneck for the industry is now patching those vulnerabilities. Impacted software teams need to validate the discovered issues, understand their impact, develop fixes, test them, and deploy patches. Back in March, OpenAI launched a preview of Codex Security, which uses agentic reasoning with automated validation to discover high-impact issues and actionable fixes specific to the codebase. Since then, it has scanned more than 30 million commits across over 30,000 codebases; more than 70,000 findings were marked as fixed by human reviewers, while over 500,000 findings were automatically determined to be fixed. Now, OpenAI is releasing an updated Codex Security plugin that can run deep scans, review recent code changes, generate security reports, trace attack paths, validate findings, and create codebase-specific patches for human review. It can also triage findings from existing scanners, advisories, bug bounty reports, and ticketing systems. OpenAI says the plugin can export results to vulnerability management systems and integrate with workflows using SARIF files, CodeQL queries, the Codex CLI, and the Codex app. Back in May, OpenAI announced the preview of GPT-5.5-Cyber, a new model built on top of the recently released GPT-5.5, designed for specialized cybersecurity work. Today, OpenAI launched the full version of GPT-5.5-Cyber through a limited release for verified defenders. On CyberGym, GPT-5.5-Cyber scored 85.6%, compared with 81.8% for GPT-5.5 and 83.8% for Claude Mythos 5. It also scored 39.5% on ExploitGym, compared with 25.95% for GPT-5.5, and 69.8% on SEC-bench Pro, compared with 63.1%. OpenAI also announced the new Daybreak Cyber Partner Program, which will allow security vendors and service providers to use GPT-5.5 with Trusted Access for Cyber in their products and services. Accenture, Akamai, Cisco, Cloudflare, CrowdStrike, IBM, Palo Alto Networks, Proofpoint, SentinelOne, Wiz, Zscaler, and others were listed as initial partners for this program. OpenAI is also launching Patch the Planet with Trail of Bits, HackerOne, Calif, researchers, and maintainers. More than 30 open-source projects have committed to participate, including cURL, Go, Python, Sigstore, and pyca/cryptography.
    • AMD confirms 26.6.2 FSR driver breaks on many Windows PCs by Sayan Sen Earlier today AMD released a major graphics driver update as it brings support for FSR 4.1 to Radeon RX 7000 series GPUs. The new update, version 26.6.2, also brings support for Assassin's Creed Black Flag Resynced and more. And while the driver technically supports Windows 10 version 21H2 and newer, the tech giant has confirmed that there is a major issue with the new driver on non-Windows 11 PCs as it fails to launch properly on such systems. The error message says, "The version of AMD Software that you have launched is not compatible with your currently installed AMD graphics driver." Therefore on the surface it looks like a compatibility problem. AMD has also confirmed that the device manager will display the yellow bang or yellow exclamation sign alongside your GPU under the Display adapters dropdown. Here is what the Radeon team's official advisory recommends to affected users: "Users Running Windows 10 and AMD Software: Adrenalin Edition 26.6.2 May Encounter Yellow Bang in Device Manager Affecting AMD Radeon RX Series Graphics ... Our Engineers are currently investigating this issue and will provide a fix once it is available. Affected users may revert to AMD Software: Adrenalin Edition 26.6.1 as a temporary workaround." As such you should revert back to the previous 26.6.1 driver which was released earlier this month. In case you were looking to play Assassin's Creed Black Flag Resynced and DOOM: The Dark Ages | Revelations you will probably have to wait a while if you want the driver to support those games officially. You can find the support article here on Microsoft's website.
    • https://uupdump.net/selectlang...7829-4524-978d-7b5fe79263e3
    • A McDonald's restaurant uses about 1.5 to 2 million gallons of water per year for operations like food preparation, cleaning, and restrooms. That is a lot less than the 2,083 gallons of water per megawatt hour mentioned above.
    • Turbo Pascal Original authorAnders Hejlsberg (at Borland) DeveloperBorland Release20 November 1983; 42 years ago[1][2] Operating systemCP/M, CP/M-86, MS-DOS, Windows 3.x, Classic Mac OS PlatformZ80, x86, 68000, PC-98 https://en.wikipedia.org/wiki/Turbo_Pascal It was the one language I actually learned to program in.   I wasn't very good at it and never used it at work.    If anyone has any personal Turbo Pascal stories or personal accomplishments using it, please take a moment to share.   Thanks. Peace
  • Recent Achievements

    • Week One Done
      Almohandis earned a badge
      Week One Done
    • Rookie
      dorf went up a rank
      Rookie
    • First Post
      mike_rumble earned a badge
      First Post
    • Dedicated
      tuben earned a badge
      Dedicated
    • Week One Done
      mnsgroup earned a badge
      Week One Done
  • Popular Contributors

    1. 1
      +primortal
      504
    2. 2
      +Edouard
      208
    3. 3
      PsYcHoKiLLa
      100
    4. 4
      Michael Scrip
      88
    5. 5
      neufuse
      71
  • Tell a friend

    Love Neowin? Tell a friend!