How does my ISP (introducing vectoring) know which DSL modem I am using?

I work for an ISP rolling out G.Vectoring so I know a little about it. While the info above is correct, its missing the point.

The crucial thing is that vectoring requires 100% compliance within the group. The group can be a line card (~30 ports), a rack (192 ports with Huawei gear) or a full DSLAM (n*192). For all cases the lines are all doing computations to reduce crosstalk but they all need to inform the port in the DSLAM what noise they're seeing so that it can "co-ordinate" the negation of said noise. If 1 modem in a group doesn't send back this info the entire group (possible a full housing estate) get reduced performance as that line is an unknown and its noise cant be negated.

You're lucky, some ISPs will kill your port when vectoring is enabled if you're non compliant meaning your get no VDSL signal whatsoever and your connection goes dead. Its then up to you to get a compatible modem for it to be turned back on.

While your modem may say it supports it, there have been issues with certain modems not implementing it correctly. One good example of that is fritzboxes which are performing very poorly with vectoring.

Also, they can tell roughly what you're using. The DSLAM and modem communicated during the sync broker process and the DSLAM is informed of chipset brand ID, MAC, firmware version and serial number. There are possibly a few other bits of info too on the backend but they are the main ones. This is all before you get a PPPoE connection back to the BRAS.

Check for firmware updates from Sagem first of all. G.Vectoring is "bleeding edge" and not all the kinks have been worked out. If that doesn't resolve it then you need to replace it. There's a good change Huawei made the DSLAMs (it's them or Ericsson) so going with their gear isn't a bad bet. AFAIK the 658C only goes to ISPs though. You might find a Zyxel VMG8324-B that'd do it though. If you want more advanced features the Draytek Vigor 28XX's are the ones to go for, but they're expensive.

After you become compliant you may need to call them and ask for the port to be reset and re profiled to 50 (or maybe more with vectoring benefits).

Every network device has a unique identifier (ie: a MAC address) that's required for communication. You are connected to their network, so they can see that unique ID on your modem.

The first chunk of a MAC address is assigned to a NIC manufacturing company to use, and they make up the second part as they see fit, so they can create many MAC addresses that they know are unique.

You can determine which company made the NIC on the device based on the MAC address.

If all the ISP's modems are made by the same company, then the modems should all have the same beginning number-set in their MAC addresses. The ISP can then easily identify when a device is not one of theirs.

From reviewing the Wikipedia article on PPPoE, I don't believe your DSL equipment transmits the model number, but the whole exchange starts by the DSL modem sending out a PADI broadcast packet, which includes the MAC address of the DSL modem:

If a user wants to "dial up" to the Internet using DSL, then his computer first must find the DSL access concentrator (DSL-AC) at the user's Internet service provider's point of presence (POP). Communication over Ethernet is only possible via MAC addresses. As the computer does not know the MAC address of the DSL-AC, it sends out a PADI packet via an Ethernet broadcast (MAC: ff:ff:ff:ff:ff:ff). This PADI packet contains the MAC address of the computer sending it.

If you can put your current modem into bridge mode, you might be able to have a computer behind it run the PPPoE stuff and then spoof a Sagem MAC address.

Linux supports PPPoE via rp-pppoe the last time I checked, and you'd have to also configure this Linux system to be a router (will take some digging into iptables) - and also get your current VPI/VCI and login information from the modem before you do that, to configure into rp-pppoe.

You might be able to then do something silly like change the NIC attached to the DSL modem to a Sagem MAC address and see what happens. However, I would imagine they are tracking which MAC addresses initiate requests on their network (and you need to provide your DSL username and password as well) - your ISP will probably know about it, and maybe not be very happy with you. So I wouldn't do it.

However, I believe they are identifying your equipment via the MAC FWIW.

I used to do tech support and train new hires for an ISP that provided DSL and Fiber connections, so this may vary a bit, but the principle should be the same.

All Tech Support agents had a tool that can see the entire network from the Central Office all the way into your home. They can also see if you have a proprietary modem/router with reasonable accuracy(90%+).

If you do have a modem/router provided by them, they can take it a step further. They can see how many other host devices that have been connected, what their computer or device names are, which ones are currently connected, as well as the local IP of said devices.

Now keep in mind this is accessible to just the Front-Line call center agents. People who are in the Central Office almost certainly have additional tools or privileges where they could grab more information.

I am almost certain there is something in the firmware authenticates with the Central Office that verifies if you are using the correct modem. I highly doubt you will be able to fool them and you may just have to suck it up and buy a supported modem.

Having worked for a previous ISP, I could give you a bit more information.

More than likely, your ISP has migrated to a system where the modem's MAC address is tied to the user account. This MAC address is unique to every modem, and is sent whenever the modem communicates with the ISP. Most major DSL providers in the USA and virtually all cable providers have already switched to this type of system.

Let's look at two scenarios:

• User "Bob" purchased a modem himself. He connects to the ISP and establishes his account via PPPoE. The ISP looks up the account information and shows a valid login attempt. The ISP shows that there either is no MAC address on file for this customer, or the MAC address doesn't match. The ISP sets the limit of 7mbps speed.
• User "Alice" purchased a modeem directly from the ISP. Upon purchase, the ISP updates Alice's account with the new MAC address. She connects to the ISP and establishes her account via PPPoE. The ISP looks up the account information and shows a valid login attempt. The ISP shows that the MAC address matches the MAC address on file. Therefore the ISP sets the speed to Alice's current 50mbps plan.

At the company I used to work for, a customer who purchased a supporting modem could update their account with the MAC address. That way they weren't forced to purchase one directly from us, but they still needed to ensure that we had the information.

It really is easy to lock down a monopoly this way, as there's really no way around that requirement. For example, spoofing a MAC address is pointless. At the point that you can spoof the address, you will have already purchased one or in some other way had your modem added to their system.