The current state of LoRaWAN security

Internet of things (IoT) devices have been adopted across many areas and use cases, from the production
lines of manufacturing industries to the operation of smart cities. IoT is expected to have an even more
significant impact on daily life in the near future; however, barriers such as increased power consumption
prevent many organizations from adopting the technology. A subset of IoT technologies that use Low Power
Wide Area Network (LPWAN)1 was created to resolve the problem of power consumption. Among these
LPWAN technologies, we can find cellular IoT systems known as Narrow-Band IoT (NB-IoT), SigFox, and
LoRa.
The LoRa (long-range) protocol helps companies connect low-powered IoT devices to the internet using a
wireless connection. Long Range Wide Area Network (LoRaWAN) is a radio-based technology that works
in conjunction with LoRa. Since they are affordable and convenient to use, many industries have adopted
LoRaWAN technology. Specifically, they have been used in soil management, animal monitoring, weather
monitoring, water meter reading, environment sensors for dams, asset tracking and fleet management,
failure prediction for buildings, and more. These technologies have been introduced as an alternative to
short-range systems such as Bluetooth, ZigBee, and Z-Wave. However, questions must be raised when
adopting these technologies in sensitive environments: are they secure, and do they present a risk to
organizations using them?
This article will discuss LoRa modulation technology and the LoRaWAN protocol by first summarizing
previous security research material on the topic. The succeeding sections will cover more solutions and
introduce our tools that add to past research. We also provide insight into overcoming the limitations in
defining attacks against these technologies and propose solutions that can reinforce LoRaWAN devices
against radio and hardware attacks.
Introduction
LoRa PHY is the physical layer commonly used by the LoRa Wide Area Network (LoRaWAN) Medium
Access Control (MAC) stack, as shown in the OSI network model:

Figure 1. The LoRaWAN (MAC) protocol stack is implemented on top of LoRa modulation (PHY).2
There are three dominant LPWAn technologies: LoRaWAN, NB-IoT, and Sifox. Compared to NB-IoT,
LoRaWAN is perfect for single-building applications as it can be set up and managed without any telecom
operator dependence on a gateway. It also has a longer battery life but a lower data rate and more
latency than NB-IoT. LoRaWAN has also started to be more widely deployed in Europe compared to
Sigfox, although Sigfox’s radio module costs less.
We also noticed other differences between the three technologies:

Table 1. A Comparison of the different Low-Power Wide Area Network (LPWAN) technologies (fixed version from
National Center for Biotechnology Information3
)
Both SigFox and LoRa use ISM bands4 (i.e., 868 MHz in Europe, 915 MHz in North America, and 433 MHz
in Asia). Like SigFox, LoRa also transmits signals using a proprietary spread spectrum technique
modulation scheme4 based on Chirp Spread Spectrum modulation (CSS) to encode information. This
modulation technique was created by Semtech, which provided specifications on its website.5 Invented
initially for radar applications in the 1940s and then used by military and secure communications
applications, CSS was adopted by the IEEE for long-range and mobility for the Low-Rate Wireless Personal
Area Networks (LR-WPANs) standard, 802.15.4.
LoRa end-devices can also use different modulations as supported by the SX12xx series (like the SX1261):
 FSK (Frequency-Shift Keying)
 GFSK (Gaussian Frequency-Shift Keying)
 MSK (Minimum-Shift Keying)
 GMSK (Gaussian Minimum Shift Keying)
 LoRa
It should also be noted that LoRa modulation is used over Frequency Shift Keying (FSK) to send data over
long distances because of the LoRa Spread Spectrum’s high sensitivity. FSK provides a higher data rate.

Figure 2. Two SX1276 transceivers used in a Dragino Gateway
LoRa Modulation
Symbols are transmitted around a sin wave around a defined carrier frequency, noted as carrier frequency
(fc), with a linear frequency that varies over time:
Figure 3. Linear frequency-modulated up-CHIRP6
This sin wave is called a CHIRP (Compressed High-Intensity Radar Pulse). A chirp is a signal that varies
in frequency and can increase or decrease across the whole bandwidth. The following example shows an
up-chirp containing a message, m(t), with eight symbols encoded in 8 bits using a speed of 1 KBaud:

Figure 4. Byte associated to chirp symbols
In the case of down-chirps, the signal frequency decreases over time.
The CSS technique presents many advantages:
 Transmitting signal across vast distances
 Immunity to multipath and fading
 Resistance to doppler shift (moving devices, high clock tolerance)
 Good sensitivity
 Simple to implement
 Low power and adapted for low data rates
Two parameters define the data bit rate:
 Bandwidth
 Spreading factor
The following expression defines the modulation bit rate Rb as:
(where SF = spreading factor (7...12); BW = modulation bandwidth)
LoRa CSS modulation uses three bandwidths, using the 125 kHz bandwidth the most:
 125 kHz
 250 kHz
 500 kHz
LoRa uses six different spreading factors, from SF7 to SF12, as shown in the following graphic:

Figure 5. Comparison of LoRa Spreading Factors: SF7 to SF127
This gives us an idea of the different configurations we can find in the EU868, EU433, CN780, and AS923
bands. The data rates8 are as follows:
Data Rate Configuration Bits/s Max Payload
DR0 SF12/125kHz 250 59
DR1 SF11/125kHz 440 59
DR2 SF10/125kHz 980 59
DR3 SF9/125kHz 1,760 123
DR4 SF8/125kHz 3,125 230
DR5 SF7/125kHz 5,470 230
DR6 SF7/250kHz 11,000 230
DR7 FSK: 50kpbs 50,000 230
Table 2. Data rates associated to Spreading Factor and bandwidth configuration
This allows us to conclude that a higher spreading factor equates to a higher over-the-air time, which means
the information will take longer to send. On the other hand, a lower spreading factor equates to a higher
data rate, which means the information will be sent quicker.
The LoRa physical layer includes eight up-chirps for preamble symbols (in red), two down-chirps
synchronization message symbols (in green), a physical payload (in pink), plus an optional CRC depending
on message direction:

Figure 6. LoRa modulated up-link frame
The demodulation process of a LoRa modulated message can be done by multiplying the received signal
su(t) with a computed inverse chirp as ref(t), such as the signal s(t) before symbol extraction is s(t) =
su(t)*ref(t), in order to isolate significant components and obtain the symbols to extract after:
Figure 7. Decode LoRa modulated symbols9
To extract these symbols, the recovered s(t) needs are passed to a fast Fourier transform (FFT) analysis
process after being filtered. Symbols can then be extracted properly, as seen in the following flowgraph:

Figure 8. Decoded LoRa modulated IQ symbols10
The following diagram details a complete IQCSS transceiver:
Figure 9. LoRa PHY transceiver block diagram11
To send and receive LoRa-modulated signals, Semtech provides cheap transceivers that can be used for
uplink and downlink messages and are generally used by endpoint-devices as gateways.
But a few software-defined radio implementations, especially two GNU Radio modules, were also released
to work on LoRa:
 gr-lora from rpp0 from Pieter Robyns, Eduard Marin, William Thenaers, and Clayton Smith
 gr-lora from Matt Knight of Bastille Threat Research Team

Before releasing the gr-lora module, Matt Knight made a very instructive presentation on LoRa PHY at the
33rd Chaos Communication Congress (33c3),14 which should be viewed by anyone interested in LoRa PHY
implementation. However, this implementation hasn’t been updated in three years and needs to be
readapted to the current version of GNU Radio. The next part of this blog series will discuss the use of the
actively updated gr-lora module from the rpp0 repository, which we can also directly use to decode uplink
and downlink messages.
IoT devices can use LoRa PHY to transport messages, but since they do not provide native security
mechanisms, developers need to implement their own security to protect messages from interception,
injection, replay attacks, and other malicious activity.
Another layer on top of LoRA PHY, called LoRaWAN, has been created to simplify communications and
address these security problems.
LoRaWAN
LoRaWAN is the cloud-based MAC layer protocol. This layer is used to communicate between LPWAN
gateways. The following diagram shows the classic architecture of a LoRaWAN network:
Figure 10. LoRaWAN architecture12
We reproduced a real-world environment with two development LoRaWAN kits, a LoRaWAN GPS tracking
badge, and a LoRaWAN door sensor connected to a Dragino LG308 Gateway. In the following picture, we
can observe the environment connected to one of The Things Network’s servers:

Figure 11. LoRaWAN real-world testbed with LoRaWAN sensors and a gateway

From LoRaWAN 1.1, a separate Join server can also be used to store and generate root keys as well as
send them to the application server and the network server.
LoRaWAN is not mandatory when using a LoRa device. It is also possible to use raw MAC communication
to directly send commands and messages, like in P2P (Peer-to-Peer), for example. But in that situation,
100% of the security depends on the user, as the raw MAC communication does not provide any encryption
and integrity the same way LoRaWAN does.
Regional parameters of LoRaWAN can be found in documentation from the Lora Alliance.13
Different server solutions are available to connect gateways:
 The Things Network14
 Tencent cloud
 Custom/private solutions
 Network server solution for IoT devices using LoRaWAN and with security in mind, like LORIOT15
Different classes exist for LoRaWAN communications:
1. Class A: The uplink transmission of each end-device is followed by two short downlink receive
windows.
2. Class B: End-devices of Class B allow for more receive slots — the end-device opens its
receiving window at the scheduled time, and it receives a time-synchronized beacon from the
gateway.
3. Class C: The window for receiving is always open; however, this results in increased battery
consumption.
Figure 12. The LoRaWAN stack
Two different versions of LoRaWAN are currently in use: LoRaWan 1.1, and LoRAWAN v1.0.3, which can
be found in many cheap devices sold in the market (such as Dragino gateways), including end-devices and
shields. Solutions based on Mbed OS version 5.8 (the operating system that runs the LoRaWAN stack on
most embedded devices) and more expensive solutions support LoRaWAN version 1.1 and provide even
more security features for version 1.0.3.

We will review the security impact and different modes of operation in the following sections.
LoRaWAN Security
There are two different modes that define/compute keys for MAC frame payload encryption:
 OTAA
 ABP
Regardless of the mode used with LoRaWAN communication, messages are protected by two session keys,
AppSKey and NwSKey, which are used to encrypt messages in Counter with CBC-MAC (CCM) mode, a
variation of Counter (CTR) mode16:
Figure 13. Use of CTR mode encryption within LoRaWAN
CTR mode is used to encrypt the payload using AppSKey and authenticates the message with a Cipher-
based Message Authentication Code (CMAC) based on the NwkSkey.
In the different sections, we will review these different modes to understand their impact on security.
OTAA mode: Over-The-Air Activation
If the end-device supports the Join function and can store dynamically generated keys, a join procedure
can be performed to compute keys for encrypting and protecting packet integrity.
The process for computing new keys:
1. The end-device sends a Join Request
2. The network will generate keys
3. If the device can Join the network, a Join Accept message is sent by the network encrypted with
AppKey, providing an App Nonce
4. With the given parameters sent via the Join Accept message, the end-device can compute the
new encryption and integrity keys

Figure 14. LoRaWAN 1.0 Join procedure in OTAA
Some differences exist between 1.0 and 1.1 in terms of security.
LoRaWAN 1.0
As described in LoRaWAN 1.0.3 specifications,17 the Join Request is sent in clear-text to the
gateway with the following parameters:
 DevEUI: unique end-device identifier in IEEE EUI64 address space
 AppEUI: the application identifier in IEEE EUI64 address space
 A random DevNonce of 2 bytes
From the specifications, DevNonce values are tracked to avoid replay attacks.
The message integrity code (MIC) for this message, which also enables the network to check if the
AppKey is correct, is computed as follows:
cmac = aes128_cmac(AppKey, MHDR | AppEUI | DevEUI | DevNonce)
MIC = cmac[0..3]
If the device is allowed to join the network and the MIC is correct, the network sends an
encrypted Join Accept message with the following fields:
 AppNonce in LoRaWAN 1.0: random value (3 bytes)

 NetID, called Home_NetID in 1.1: network ID (3 bytes)
 DevAddr: Device ID (3 bytes)
 DL Settings: downlink parameters
 RxDelay: delay between TX and RX (1 byte)
 CFList: optional list of channel frequencies (16 bytes)
The Join-accept payload is encrypted as follows:
aes128_decrypt(AppKey, AppNonce | NetID | DevAddr | DLSettings | RxDelay | CFList | MIC)
The sessions keys are computed in the network and end-device sides as follows:
NwkSKey = aes128_encrypt(AppKey, 0x01 | AppNonce | NetID | DevNonce | pad16)
AppSKey = aes128_encrypt(AppKey, 0x02 | AppNonce | NetID | DevNonce | pad16)
All the communications will then be encrypted using AppSKey, and the integrity protected with
NwkSKey.
Nevertheless, only one AppKey key is used to compute the MIC as well as AppSKey and
NwkSKey, which leaves a bigger opening for an attacker to crack either the MIC of the Join
procedure’s message or the Join-accept message to retrieve this key. Retrieving the key could
allow an attacker to eavesdrop on the messages between an end-device and a gateway.
LoRaWAN 1.1
The 1.1 version introduces more key diversification to make key-cracking attacks more complex
on LoRaWAN. Everything is documented in the LoRaWAN 1.1 specifications.18
In this version, LoRaWAN defines two root keys, NwkKey and AppKey, that are AES-128 keys
specific to the end-device and assigned during the fabrication.
For the Join-request message, values transmitted by the end-device are generally the same, but
they are renamed:
According to the specifications, DevNonce is a counter that starts at zero when the device is initially
powered up and is incrementally increased by one with each Join-request. This DevNonce value
will never be reused for a given JoinEUI value, but since some end-devices cannot store this
counter in non-volatile memory, they may discard the Join-request of the device to the server. In
that case, the JoinEU (AppEUI renamed in v1.1) must reset as well.

To protect the message's integrity, the MIC is computed using a dedicated NwkKey:
cmac = aes128_cmac(NwkKey, MHDR | JoinEUI | DevEUI | DevNonce)MIC = cmac[0..3]
If the device is allowed to join the network, a Join-accept message is sent to the device with the
following encrypted parameters:
 JoinNonce in LoRaWAN 1.0: random value (3 bytes)
 NetID, called Home_NetID in 1.1: network ID (3 bytes)
 DevAddr: Device ID (3 bytes)
And this Join-Accept message is encrypted as follows using the NwkKey:
aes128_decrypt(NwkKey, JoinNonce | NetID | DevAddr | DLSettings | RxDelay | CFList | MIC)
 DL Settings: downlink parameters
 RxDelay: delay between TX and RX (1 byte)
 CFList: optional list of channel frequencies (16 bytes)
The key used to encrypt the Join-Accept message depends on the Join or ReJoin Request
message that triggers it.
The Join-Accept message is encrypted as follows:
aes128_decrypt(NwkKey or JSEncKey, JoinNonce | NetID | DevAddr | DLSettings | RxDelay |
CFList | MIC)
Like in the 1.0 version, an exception exists by using the NwkKey only instead of the AppKey to
encrypt this message. Indeed, in the downlink configuration field DLsettings, 7-bit subfields are
indicated if the Network Server implements the LoRaWAN 1.0 (unset) or 1.1 (set) protocol.
Referring to the specifications, when this OptNeg bit is set:
 The protocol version is further (1.1 or later) negotiated between the end-device and the
Network Server through the RekeyInd/RekeyConf MAC command exchange
 The device derives FNwkSIntKey & SNwkSIntKey & NwkSEncKey from the NwkKey
 The device derives AppSKey from the AppKey

ABP: Activation by Personalization
The ABP method is simpler than OTAA as there is no Join Procedure. Nevertheless, it has downsides in
terms of security, as session keys are hardcoded on versions 1.0 and 1.1.
Indeed, session keys stay the same until the user manually changes them or when a firmware
update/upgrade is applied. Because of this, ABP is more vulnerable to a cryptanalysis attack compared to
OTAA.
MAC Frame Payload Encryption (FRMPayload)
With given session keys, we can protect MAC frame payloads on Data Up and Down messages.
LoRaWAN 1.0
First, the key "K" is chosen depending on the FPort of the data message:
For the encryption, the message is split into a sequence of blocks: "Ai" for i = 1..k with k
=ceil(len(pld) / 16) (with pld = FRMPayload).
Note that the Direction (Dir) is defined by "1" for a downlink and "0" for an uplink.
So, the block Ai is encrypted as follows:
Si = aes128_encrypt(K, Ai) for i = 1..k
S = S1 | S2 | .. | Sk
Encryption and decryption of the blocks are then performed as follows:
(pld | pad16) xor S
The MIC is computed as follows:

msg = MHDR | FHDR | FPort | FRMPayload
cmac = aes128_cmac(NwkSKey, B0 | msg)
MIC = cmac[0..3]
Where B0 is defined with the following parameters:
LoRaWAN 1.1
The key “K” used depends on the FPort as for LoraWAN 1.0 version:
For each data message, the algorithm defines a sequence of Blocks Ai for i = 1..k with k
=ceil(len(pld) / 16):
Blocks are computed and encrypted/decrypted the same way as for LoRaWAN 1.0:
Si = aes128_encrypt(K, Ai) for i = 1..k
S = S1 | S2 | .. | Sk
Encrypted/decrypt = (pld | pad16) xor S
For computing the MIC, there are differences between Uplink frames and Downlink Frame.
Downlink Frames
For Downlink frames, the MIC is computed with the SNwkSIntKey as follows:
cmac = aes128_cmac(SNwkSIntKey, B0 | msg)
MIC = cmac[0..3]
And the B0 blocks is defined differently by introducing a ConfFCnt field:

In all cases, this ConfFCnt = 0x0000, but if the device is connected to a LoRaWAN 1.1 network
and the ACK bit of DL is set (acknowledging an uplink "confirmed" frame), then the ConfFCnt is a
frame counter with a value of "confirmed" uplink modulo 2^16.
Uplink frames
If the device is connected to a LoRaWAN 1.0 Network Server, the MIC uses B0 block defined as
follows:
And so the MIC is computed as follows:
cmacF = aes128_cmac(FNwkSIntKey, B0 | msg)
MIC = cmacF[0..3]
If the device is connected to a 1.1 Network Server, the MIC then uses B0, but also a B1 block
defined as follows:
And the MIC is computed as follows:
cmacS = aes128_cmac(SNwkSIntKey, B1 | msg)
MIC = cmacS[0..1] | cmacF[0..1]
Attacks against LoRaWAN
Most of the attacks documented below have been performed on LoRaWAN class A devices. However, a
class B attack is also possible; these are usually done to drain the end-device’s battery.
DoS in ABP mode

Given that the counter in FRMPayload is only 16 bits long In LoRa 1.0, a denial-of-service (DoS) attack is
actually possible. Xueying Yang, Evgenios Karampatzakis, Christian Doerr, and Fernando Kuiper (a team
of security researchers based in the Netherlands)19 demonstrated this. They posited that a malicious actor
could replay a captured packet, wait until the counter overflows, and replay this packet to realize a DoS
attack:
Figure 15. An example of a replay attack for ABP20
Session keys stay the same until manually changed or with a firmware update/upgrade. Because of this,
ABP could be more vulnerable to a cryptanalysis attack than OTAA.
Eavesdropping
In their paper,21 Xueying Yang, Evgenios Karampatzakis, Christian Doerr, and Fernando Kuiper also
highlighted that the key could be retrieved if the counter is reset, as it is not used securely in LoRaWAN
1.0. Indeed, when the counter is reset, the keystream will be reused and could allow an eavesdropping
attack. Another factor that leaves it vulnerable is the use of CTR to encrypt information.
The same researchers also highlighted a bit-flipping issue in v1.0 that is backward compatible with v1.1.

Bit-Flipping
Regardless of the version, the integrity of LoRaWAN messages is protected thanks to the MIC and
encryption. Nevertheless, the messages decrypted by the network server and sent to the application are
no longer protected:
Figure 16. The setup of a bit-flipping attack
To protect messages from the network server to the application server, it would require designing the
network to use an SSL tunnel, preferably with a client SSL certificate in order to protect the
communication or by using another MIC field inside the MAC Layer Payload computed by the AppSKey.
Ack Spoofing
An acknowledgment mechanism was introduced in LoRaWAN to maximize battery life by reducing the time
the radio needs to be powered up. But as highlighted by Xueying Yang, Evgenios Karampatzakis, Christian
Doerr, and Fernando Kuiper once again in their paper,22 the ACK message does not state which message
it is confirming.

Figure 17. ACK messages may be repurposed to acknowledge frames other than the ones the application provider
originally received
To illustrate this issue, researchers proposed to selectively jam the downlink when an end-device sends a
confirmed message to the gateway that will confirm the reception. The confirmation message will never get
to the end-device, and the confirmed message will be retransmitted seven times. Then, the message will
be considered lost or refused, and the status “mac_err” will be raised.
But during the jamming session, the attacker can capture the downlink message for the confirmation and
can play the confirmation for the first message when the end-device sends a second confirmation message.
LoRa Class B Attacks
Most setups use the class A mode, which specifies that downlink traffic must/can follow an uplink one.
Class B tends to reduce the amount of spent energy by telling end-devices to wake up periodically to wait
for any incoming messages during a receiving window. The durations of the receiving windows are specified
by beacon broadcast messages that are comprised of a PHY layer header followed by a beacon payload:
Table 3. Beacon frame content for the EU 863- 870MHz ISM band23, 24

An attacker can easily compute the different publicly known fields with malicious parameters25 to:
 Find the location of a LoRa gateway thanks to a GwSpecific field with an InfoDesc subfield
containing GPS coordinates
 Drain the battery by sending crafted beacons framed with an extreme wakeup time value
According to their paper26 (as well as another LoRaWAN security evaluation paper by Frank Hessel, Lars
Almon, Flor Álvarez27) and LoRaWAN v1.1 specifications,28 the integrity of beacon frames is still an issue.
It would be better to use a MIC instead of a CRC to check the time’s field integrity.
Root Key Management
On top of all the security mechanisms we previously discussed is the management of keys. Indeed, the
backend is exposed to the internet, which leaves it open to attacks (LFI, SQL injection, deserialization
vulnerability, etc.). A malicious actor would be able to get the secret key, read the data, craft downlink
packets, and more. All risks related to key management procedures are enumerated in a complete paper
by F-Secure Labs (ex-MWR Labs).
Here are some security points to check in a LoRaWAN setup:
 Use randomly generated keys
 Avoid the exposition of key management servers and services (exposed key management
service accessible on the internet)
 Preferably use HSM (Hardware Security Module) to keep the keys
 Preferably use OTAA mode and LoRa version 1.1.
Further Reading
An interesting paper on the MobiSPC 201829 highlights security improvements brought by LoRaWAN v1.1.
To understand the differences between LoRaWAN v1.0 and v1.1, we also suggest watching Renaud
Lifchitz’s presentation at The Things Conference 2019.30 During the presentation, Renaud stated that there
will be some vulnerabilities present even if replay attacks are used to transmit an earlier captured packet;
he also outlined how LoRaWAN v1.1 has better authentication, confidentiality, and integrity. Indeed,
message padding enforcement is not present to prevent default ECB mode encryption risks such as
information leaks of the plaintext (length, prefixes, common substrings, etc.). Moreover, backend networks
do not have a defined secure standard to protect messages between the network server and other servers
because it is sent in plaintext in UDP by default, rather than using MQTT within a TLS session.

The State of Security
When looking at LoRaWAN, we can see that very few security tools exist for the technologies released
starting this year. Some researchers from IO Activ released a framework called LAF31 (LoRaWAN Auditing
Framework). They also provided a tool that can parse, send, craft, analyze, audit a setup, and crack some
LoRaWAN packets using weak/default keys. In their paper, the researchers presented LoRaWAN 1.0.3
attacks mainly using their tool.
Nevertheless, this framework still has some limitations we can try to overcome:
 It only works with a Gateway
 It can only listen to uplink packets
 It can only listen to eight out of 64 channels
 Generation and fuzzing depends on LoRaWAN (Go) using an inflexible format such as JSON
Over the same period, another type of tool called “LoRa Craft”32 was released to intercept packets using
Software Defined-Radio and craft packets using dedicated LoRaWAN v1.0 and v1.1 Scapy layers
developed for this tool. But this tool is mainly a do-it-yourself (DIY) tool and needs much more assistance
than those already released, like the crypto-helpers for Join-Accept payloads and MIC to help crack weak
keys.
Later in May, another framework called ChirpOTLE33 was released by the SEEMOO Lab. This
demonstrated two attacks affecting the availability of LoRaWAN networks, like time drifting in LoRa class
B and a novel ADR spoofing attack to manipulate frame metadata. But still, the researchers’ setup was
limited since they only chose a few default channels to demonstrate their attack on each node.
Conclusion: Let’s Go Further!
In this first technical brief, we introduced LoRa and LoRaWAN technologies as well as security
improvements from version 1.0 to 1.1. We also outlined research and practical tools to help against
attacks on LoRaWAN networks and noted the limitations encountered by security researchers.
LoRaWAN devices may be low-powered, but they are used in critical areas and should be secured as
comprehensively as possible. For example, in 2016, a large-scale LoRaWAN network was deployed in
Taiwan,34 and cities in the country have adopted many LoRaWAN devices over the years. They are used
to monitor water levels in rivers, dam management, waste disposal, and more. Attacks against sensors
used in smart cities could potentially affect the safety of a sizeable population. In terms of the private
sector, organizations could also be fed falsified information on monitoring devices, which could impact
their bottom line. These are only some of the issues that need to be resolved.
The next part of the series will explain how we managed to overcome security research limitations by
going in another direction. We look into the use of Software-Defined Radio (SDR) and optimization when
attacking both uplink and downlink on multiple channels, as well as multiple spreading factors using a
very cheap SDR device. We will also introduce tools that would help analyze LoRa PHY as well as
LoRaWAN radio communications. As discussed earlier, the final part of the series will be on hardware
attacks and security mechanisms concerning LoRa end-devices.

References
1
TechTarget. (n.d.). IoT Agenda. “LPWAN (low-power wide area network).” Accessed on Jan. 11, 2020 at
https://internetofthingsagenda.techtarget.com/definition/LPWAN-low-power-wide-area-network.
2
Dong-Hoon Kim,Eun-Kyu Lee, and Jibum Kim. (March 30, 2019). MDPI. “Experiencing LoRa Network Establishment on a Smart
Energy Campus Testbed”. Accessed on Jan. 11, 2020 at https://www.mdpi.com/2071-1050/11/7/1917.
3
Philip J. Basford et al. (Jan. 20, 2020). The National Center for Biotechnology Information. “LoRaWAN for Smart City IoT
Deployments: A Long-Term Evaluation.” Accessed on Jan. 11, 2020 at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7038353/.
4
Peter McNeil. (March 22, 2018). Pasternack. “What are the ISM Bands, and What Are They Used For?” Accessed on Jan. 11,
2020 at https://blog.pasternack.com/uncategorized/what-are-the-ism-bands-and-what-are-they-used-for/.
5
Semtech. (May 2, 2015). Semtech. “AN1200.22 LoRa™ Modulation Basics.” Accessed on Jan. 11, 2020 at
http://wiki.lahoud.fr/lib/exe/fetch.php?media=an1200.22.pdf.
6
IoTOne. (n.d.). IoTOne. “Chirp Spread Spectrum.” Accessed on Jan. 11, 2020 at https://www.iotone.com/term/chirp-spread-
spectrum/t110.
7
Sakshama Ghoslya. (n.d.). Sghoslya. “LoRa: Symbol Generation.” Accessed on Jan. 11, 2020 at https://www.sghoslya.com/p/lora-
is-chirp-spread-spectrum.html.
8
The Things Network. (n.d.). The Things Network. “Modulation and Data Rate.” Accessed on Jan. 11, 2020 at
https://www.thethingsnetwork.org/docs/lorawan/modulation-data-rate.html.
9
Thomas Telkamp. (Nov. 10, 2016). The Things Network. “LoRa Crash Course.” Accessed on Jan. 11, 2020 at
https://www.youtube.com/watch?v=T3dGLqZrjIQ.
10
Thomas Telkamp. (Nov. 10, 2016). The Things Network. “LoRa Crash Course.” Accessed on Jan. 11, 2020 at
https://www.youtube.com/watch?v=T3dGLqZrjIQ.
11
Ivo Bizon Franco de Almeida et al. (Sept. 22, 2020). Arvix. “In-phase and Quadrature Chirp Spread Spectrum for IoT
Communications In-phase and Quadrature Chirp Spread Spectrum for IoT Communications.” Accessed on Jan. 11, 2020 at
https://arxiv.org/pdf/2009.10421.pdf.
12
The Things Network. (n.d.). The Things NetworkI. “LoRaWAN Architecture.” Accessed on Jan. 11, 2020 at
https://www.thethingsnetwork.org/docs/lorawan/architecture.html.
13
LoRa Alliance. (Feb. 20, 2020). LoRa Alliance. “RP002-1.0.1 LoRaWAN® Regional 40 Parameters.” Accessed on Jan. 11, 2020
at https://lora-alliance.org/sites/default/files/2020-06/rp_2-1.0.1.pdf.
14
The Things Network. (n.d.). The Things Network. “The Things Network.” Accessed on Jan. 11, 2020 at
https://www.thethingsnetwork.org/.
15
Loriot. (n.d.). Loriot. “Loriot.” Accessed on Jan. 11, 2020 at https://www.loriot.io/.
16
Matt Knight. (Dec. 29, 2016). RC3 33c3. “Decoding the LoRa PHY.” Accessed on Jan. 11, 2020 at
https://www.youtube.com/watch?v=NoquBA7IMNc.
17
LoRa Alliance. (2018). LoRa Alliance. “LoRaWAN 1.0.3 Specification.” Accessed on Jan. 11, 2020 at https://lora-
alliance.org/sites/default/files/2018-07/lorawan1.0.3.pdf.
18
LoRa Alliance. (Oct. 11, 2017). LoRa Alliance. “LoRaWAN™ 1.1 Specification.” Accessed on Jan. 11, 2020 at
https://lora-alliance.org/sites/default/files/2018-04/lorawantm_specification_-v1.1.pdf.

19
Xueying Yang et al. (April 17, 2020). IEEE Xplore. “Security Vulnerabilities in LoRaWAN.” Accessed on Jan. 11, 2020 at
https://ieeexplore.ieee.org/document/8366983/footnotes#footnotes.
20
21
22
23
LoRa Alliance. (2018). LoRa Alliance. “LoRaWAN 1.0.3 Specification.” Accessed on Jan. 11, 2020 at https://lora-
alliance.org/sites/default/files/2018-07/lorawan1.0.3.pdf.
24
LoRa Alliance. (2018). LoRa Alliance. “LoRaWAN 1.1 Specification.” Accessed on Jan. 11, 2020 at https://lora-
alliance.org/sites/default/files/2018-04/lorawantm_specification_-v1.1.pdf.
25
26
27
Frank Hessel, Lars Almon, and Flor Álvarez. (May 23, 2020). Arvix. “ChirpOTLE: A Framework for Practical LoRaWAN Security
Evaluation.” Accessed on Jan. 11, 2020 at https://arxiv.org/pdf/2005.11555.pdf.
28
LoRa Alliance. (2018). LoRa Alliance. “LoRaWAN 1.1 Specification.” Accessed on Jan. 11, 2020 at https://lora-
alliance.org/sites/default/files/2018-04/lorawantm_specification_-v1.1.pdf.
29
Tahsin C.M. Dönmez and Ethiopia Nigussie. (Jan. 2018). ScienceDirect. “Security of LoRaWAN v1.1 in Backward Compatibility
Scenarios.” Accessed on Jan. 11, 2020 at https://www.sciencedirect.com/science/article/pii/S1877050918311062.
30
Renaud Lifchitz. (Feb. 11, 2019). The Things Network. “From LoRaWAN 1.0 to 1.1: Security Enhancements.” Accessed on Jan.
11, 2020 at https://www.youtube.com/watch?v=FsO5zxYHfKw.
31
Cesar Cerrudo, Esteban Martinez Fayo, and Matias Sequeira. (Jan. 2020). IO Active. “LoRaWAN Networks Susceptible to
Hacking: Common Cyber Security Problems, How to Detect and Prevent Them.” Accessed on Jan. 11, 2020 at https://act-
on.ioactive.com/acton/attachment/34793/f-87b45f5f-f181-44fc-82a8-8e53c501dc4e/1/-/-/-/-
/LoRaWAN%20Networks%20Susceptible%20to%20Hacking.pdf.
32
Cesar Cerrudo, Esteban Martinez Fayo, and Matias Sequeira. (Jan. 2020). IO Active. “LoRaWAN Networks Susceptible to
Hacking: Common Cyber Security Problems, How to Detect and Prevent Them.” Accessed on Jan. 11, 2020 at https://act-
on.ioactive.com/acton/attachment/34793/f-87b45f5f-f181-44fc-82a8-8e53c501dc4e/1/-/-/-/-
/LoRaWAN%20Networks%20Susceptible%20to%20Hacking.pdf.
33
Frank Hessel, Lars Almon, and Flor Álvarez. (May 23, 2020). Arvix. “ChirpOTLE: A Framework for Practical LoRaWAN Security
Evaluation.” Accessed on Jan. 11, 2020 at https://arxiv.org/pdf/2005.11555.pdf.
34
Semtech. (Aug. 16, 2016). Semtech. “Semtech LoRa® RF Technology Adopted by Asia Pacific Telecom for Rollout of
Commercial Low Power Wide Area Internet of Things Network in Taiwan.” Accessed on Jan. 15, 2020 at
https://www.semtech.com/company/press/semtech-lora-rf-technology-adopted-by-asia-pacific-telecom-for-rollout-of-commercial-low-
power-wide-area-internet-of-things-network-in-taiwan.

TREND MICROTM
RESEARCH
Trend Micro, a global leader in cybersecurity, helps to make the world safe for exchanging digital information.
Trend Micro Research is powered by experts who are passionate about discovering new threats, sharing key insights, and
supporting efforts to stop cybercriminals. Our global team helps identify millions of threats daily, leads the industry in
vulnerability disclosures, and publishes innovative research on new threats techniques. We continually work to anticipate new
threats and deliver thought-provoking research.
www.trendmicro.com

The current state of LoRaWAN security

More Related Content

What's hot (18)

Similar to The current state of LoRaWAN security (20)

More from 📡 Sebastien Dudek (17)

Recently uploaded (20)

The current state of LoRaWAN security