Iot sdk thesis

Figure 1.3.1 LWM2M protocol stack 
[4]
 
1
Wide area IoT: wirelessly accessible from more than 10km at least;
2
Short-range IoT: not accessible from more than a few meters (e.g. Bluetooth, ZigBee accessible devices) 
3
Also known as CAGR: “
The growth rate that gets you from the initial investment value to the ending 
investment value if you assume that the investment has been compounding over the time period 
[38]
.”

---

6 
As seen, secure LWM2M device management is possible in two ways: over 
connection-oriented TCP by TLS cryptographic protocol and over connectionless UDP with 
DTLS cryptographic protocol. A number of cellular IoT (CIoT) technologies are supported. 
The use of LoRa stack, which defines its own physical layer (e.g. modulation) and LoRaWAN 
medium access control (MAC) layer is supported. The messaging could be a REST-style CoAP 
messaging, with optional application level security provided by Object Security for Constrained 
RESTful Environments (OSCORE) method [5].
LWM2M is designed for client-server topology and supports basic functionalities of 
enrollment, provisioning, monitoring, updating, bootstrapping of the devices, throughout the 
device lifecycle [4]. Client-Server architecture of LWM2M has been shown in Figure 1.3.2 
(Refer section 2.6 for the explanation in detail) 
Figure 1.3.2 LWM2M Client-Server Architecture 
[6]
 
1.4
 
The Problem 
Eclipse Wakaama is a LWM2M implementation designed with POSIX portability in 
mind [7]. Which means, Wakaama requires the support of Linux libraries on the devices that 
will act as client and server. Wakaama is written in C, which opens the possibility to make it 
compatible even with the microcontroller-based development boards that don’t have Linux 
kernel.
Microcontroller board SARA by SODAQ supports cellular LPWAN connectivity by 
means of an NB-IoT modem, whereas Arduinos can be interfaced with independent modem 
over serial communication link to have cellular LPWAN support. 

---

7 
 
SODAQ or Arduino do not provide any LWM2M libraries that can be used out-of-the-
box to enable device management on their microcontrollers. That is, there is no abstraction from 
low-level details of LWM2M stack to the developers, if they decide to run a LWM2M client 
on these microcontrollers. Thus, in order to build an NB-IoT powered application with 
LWM2M device management on such a development board, the developers must have expert 
knowledge of all the technologies involved in the protocol stack and the development board. 
Thus, being able have a software development kit (SDK) with libraries, to add LWM2M 
support to such microcontroller development boards will make application development less 
tedious, without the developers having to possess expertise in each and every technology 
involved. It will also make many use-cases feasible, especially in bandwidth and power-
constrained geographies.
With this as a motivation, for duration of my thesis, I worked on making Wakaama 
LWM2M client implementation completely compatible on microcontroller boards with 
Arduino form factor, to enable device management over NB-IoT network. The structure, 
objectives and short description of the tasks has been explained in subsequent sub-chapters.
To summarize, during the thesis I worked on studying and implementing technologies 
present at different layers of LWM2M protocol stack.