Can I register an ABP device that has fixed DevAddr, NwkSKey and AppSKey?

Thx @arjanvanb . I connected this node : http://webshop.ideetron.nl/ID140015-01 to TTN using your instructions:

ttnctl devices register personalized your DevAddr 2B7E151628AED2A6ABF7158809CF4F3C 2B7E151628AED2A6ABF7158809CF4F3C --relax-fcnt

Since two weeks ttnctl.exe does not seem to work anymore. I can use the web dashboard to modify existing ABP nodes but need the ttncl.exe to generate ABP nodes with fixed DevEUI etc.

D:\ttn>ttnctl user login mw12554@hotmail.com Password: INFO Visit https://account.thethingsnetwork.org to register or to retrieve your account credentials. FATAL Failed to login error=Grant type is not available for the client My password which always worked before is not accepted any longer.

We recently launched version 2.0.0 of our backend, which is no longer compatible with v1-staging. You can configure your v1-staging ttnctl with ``–ttn-account-server https://v1.account.thethingsnetwork.org` to delete your old devices from v1-staging and move them to our v2 backend.

See Launching Production Environment for more information.

1 Like

Pfoe that did the trick. Thanks!.

So we should start migrating. How long will the “old” staging environment be alive?

Maarten

The Roadmap says Q1 2017.

As an aside: network providers (such as TTN) are required to actively block devices that always send on SF11 or SF12, to keep their LoRa Alliance NetID: Devices with fixed hardcoded data rates of SF12 or SF11 must not be allowed to join the network.

1 Like

It’s not possible to register just any DevAdr, as it includes a network identifier:

The most significant 7 bits are used as network identifier (NwkID) to separate addresses of territorially overlapping networks of different network operators and to remedy roaming issues. The least significant 25 bits, the network address (NwkAddr) of the end-device, can be arbitrarily assigned by the network manager.

[…]

The seven LSB of the NetID are called NwkID and match the seven MSB of the short address of an end-device as described before.

The documentation on Addressing and Activation explains some more:

The Things Network Foundation has received a 7-bit device address prefix from the LoRa Alliance. This means that all TTN device addresses will start with 0x26 or 0x27 (although addresses that start with these might also belong to other networks with the same prefix).

(TTN currently uses a NetID of 0x000013, hence its NwkID is the 7 bits binary %0010011. As hexadecimal 0x26 is binary %00100110, and 0x27 is %00100111, both indeed start with the 7 bits NwkID.)

Also, the TTN-assigned DevAddr uses some more encoding for internal routing. From the same documentation:

Within TTN, we assign device address prefixes to “regions” (for example, device addresses in the eu region start with 0x2601). Within a region, the NetworkServer is responsible for assigning device addresses. We are using prefixes here too for different device classes (for example, ABP devices in the eu region start with 0x26011) or to shard devices over different servers.

(This indeed yields less than 12 bits within a region and class to generate a DevAddr, yielding only 4,096 unique values within such region and class. This is not a problem as a DevAddr is not unique.)

@htdvisser wrote in Slack, about just copying some existing DevAdr into TTN Console:

No, that won’t be possible in the public network

We route data based on the NetID that was assigned to us by the LoRa Alliance

Devices with a DevAddr that does not match our NetID, are (according to the LoRaWAN spec) part of another network

If a manufacturer puts fixed DevAddr/NwkSKey/AppSKey in a device without giving the customer the ability to change those, they either don’t know what they’re doing or they don’t really care about the customer or they try to create some kind of lock-in to their own network.

And if if the NwkSKey/AppSKey is 2B7E151628AED2A6ABF7158809CF4F3C I would think twice before buying the device.

And about the test NetID 01/02:

We don’t issue DevAddresses for 00 / 01 but it is possible to “force-register” a device with such address (not in the console, only with ttnctl)

For such devices, we can still route data

But uplink for anything else than 00 / 01 / 26 / 27 will just be dropped before it’s even considered by the backend

Or routed to a roaming partner when roaming is added to LoRaWAN and supported by us

1 Like

Thanks for this info.
Is there somewhere a list of assignments NetID <-> LoRaWAN Network Provider?

@arjanvanb
Thank you for the helpful information. I have devices whose DevAddresses begin with 00. How can I force-register these devices over the ttnctl?

Thank you very much for your help!

ttnctl devices set my-device-id --override --dev-addr 00000000 --nwk-s-key 00000000000000000000000000000000 --app-s-key 00000000000000000000000000000000

4 Likes

As for currently enforcing this, @htdvisser wrote on Slack July 3rd:

But no, we’re indeed not enforcing it at the moment
It’s pretty difficult to detect (with certainty) that a device is hardcoded to SF11/SF12
Sure, we can “assume” it if a device with a good signal is transmitting on SF12
But in my opinion, we should first notify the owner before we start blocking a device
And we’re still working on this notification system

The TouchTag has a fixed DevAddr that starts with 16. When I register it and personalize it via ttnctl according to the console it is given a DevAddr that starts with 26. Whether or not I change that via ttnctl, the gateway traffic only shows transmissions from the DevAddr 16xxxxxx and in neither case does traffic show at the application level.

Does it seem possible that TTN would in the future accept traffic for a device with a fixed DevAddr that starts with 16? Would that require some kind of peering agreement? My guess would be that sensors with a fixed DevAddr would be cheaper to manufacture than ones where the user can alter the DevAddr. And for devices with a fixed DevAddr to need to be specifically manufactured for use only in a particular network would seem regrettably inefficient. But I may have a fundamental misunderstanding of the issue.

Even it would do peering/roaming, you’d then need to register it with EveryNet in Russia.

1 Like

If the devaddr is fixed the node uses ABP (not the preferred activation mode), with OTAA the customer would be able to create an application with the (unique) id, register the nodes EUI and key and be in business.

1 Like

@Verkehrsrot

https://www.thethingsnetwork.org/docs/lorawan/address-space.html

Thanks! This is useful when migrating a device from TTN to a TTI instance or back.

Hi… I have similar problem with my sensor. The sensor that I have already with a prefix DeviceAddress (Address: B00XXXXXX), Application Session Key and Network Session Key. According to the manufacturer, the sensor is configured with ABP.

When I added the sensor to TTN using TTNCLI and force change the Device Address to the sensor Device Address, I couldn’t get the publish in TTN. The instruction that I used to change the Device Address

ttnctl devices set [device_id] --dev-addr [sensor_device_address] --handler-id ttn-handler-asia-se

As i’m using Asia server, I need to put the --handler-id, else there will be error in the process.

Is there any solution for this?

A DevAdrr that starts with B00 will never work with TTN, see the explanation above. You’ll need to reprogram the device to use the TTN-assigned DevAddr for ABP. Or make it use OTAA and configure TTN and the device to use the same values for DevEUI, AppEUI and AppKey.

1 Like

Hi… Thank you for the feedback. Although the data is available in the gateway traffic, application level, the data can’t be shown in the application due to the network identifier?

gateway

Sorry, I don’t know how to explain better than already quoted from the documentation above:

…and as already quoted from Slack:

If you don’t believe it, then click on an item in the gateway traffic, to see that (in your case) TTN will not recognize this as traffic it needs to handle. For example, for a DevAddr starting with 14:

Non-TTN gateway traffic

…and the “drop” along with the reason “no brokers” in:

Dropped gateway traffic

1 Like