UNIVERSITY OF EDINBURGH COLLEGE OF SCIENCE AND ENGINEERING SCHOOL OF INFORMATICS
INFR11146 INTERNET OF THINGS SYSTEMS, SECURITY, AND THE CLOUD
May 2020 13:00 to 15:00
INSTRUCTIONS TO CANDIDATES
Answer QUESTION 1 and ONE other question.
Question 1 is COMPULSORY. If both QUESTION 2 and QUESTION 3 are answered, only QUESTION 2 will be marked.
All questions carry equal weight. This is an OPEN BOOK examination.
Year 4 Courses
Convener: T.Komura
External Examiners: S.Rogers, S.Kalvala, H.Vandierendonck
THIS EXAMINATION WILL BE MARKED ANONYMOUSLY
1. THIS QUESTION IS COMPULSORY
(a) Explain the role and characteristics of General Purpose Input/Output in embedded systems. How would you use GPIO to control the speed of a motor or fan?
(b) AWi-Fiaccesspointservestwostations,eachofwhichuploadsvideocontent to a cloud service over UDP. The medium access operation is as follows:
◃ Each station contends with a CWmin = 16 and uses maximum back-off stage K = 1.
◃ The stations transmit packets with size L = 1,500 bytes and send 8 frames upon each channel access.
◃ Payloads are preceded by a 20 μs preamble and a 24-byte header (A- MSDU aggreggation). A 4-byte Frame Check Sequence (FCS) is ap- pended to each aggregate.
◃ The channel must be sensed idle for AIFS = 34 μs to trigger the access procedure and the duration of an empty slot is σ = 9μs.
◃ The RTS/CTS mechanism is employed to improve efficiency, whereby a 18-byte RTS and 14-byte CTS are exchanged; Each of these control frames is preceded by a 20 μs preamble and the frames are separated by a Short Inter-frame Space (SIFS) of 16 μs.
◃ RTS/CTS frames are transmitted at 24 Mbit/s. If the RTS/CTS frame exchange is successful, the A-MSDU is sent within SIFS, at a bit rate of 65 Mb/s.
◃ Successfully received packets are acknowledged within SIFS, and the total duration of an acknowledgment frame is 24 μs.
i. Assuming the probability of two or more consecutive collisions is negli- gible, what is the average transmission attempt rate of a station?
(you may approximate to 2 decimal points)
ii. What is the total throughput that a station can achieve?
iii. Assume RTS/CTS is robust to errors, but a data frame transmission may be corrupted with probability p = 0.02. How long will it take on average for a station to successfully transmit a data frame?
(c) Discuss three advantages of using CoAP instead of HTTP in IoT. What mechanism can be used with CoAP to facilitate firmware updates and what does its underlying operation consist of?
(d) A64-bitplaintextpassword0xC0AB4ACADAB4AAC0isencryptedwithaFeistel cipher that employs the following round function
QUESTION CONTINUES ON NEXT PAGE
Page 1 of 4
[3 marks ]
[5 marks ] [4 marks]
[3 marks ]
[2 marks ]
QUESTION CONTINUED FROM PREVIOUS PAGE
F(Ri,Ki)=((Ri <<8)⊕Ki)>>4,
where << denotes the circular shift left operation, >> circular shift right, and ⊕ is the XOR. If the sub-keys at rounds 1 and 2 are K0 =0x4C004CFF and K1 =0xCB8BCb8B, what is the cipher text obtained after the plaintext passes through these rounds?
[8 marks]
Page 2 of 4
2. ANSWER EITHER THIS QUESTION OR QUESTION 3
(a) Explain the differences between Bluetooth Classic and Bluetooth Low En- ergy, in terms of channel properties and the procedure to establish a channel hopping sequence between two peers.
(b) Consider a system that uses the Qn.m fixed point format to represent num- bers on 12 bits, whereby 8 bits are reserved for the fractional part.
i. What is the range of numbers that can be covered with this format?
ii. Convert the decimal number ‘-7.828125’ into Qn.m format, giving your answer in both binary and hexadecimal.
iii. How should (n) and (m) be modified, to be able to represent the value 21.578125?
(c) Describe the concept of ‘hypervisor’, explaining the roles this entity fulfill, as well as the key differences between different classes of hypervisors.
(d) TheRoutingProtocolforLow-PowerandLossyNetworks(RPL)isemployed in an IoT network, resulting in the following destination-oriented directed acyclic graph (DODAG) topology.
i. If the expected transmission count time on each link is 1.2, what is the rank of node E?
ii. List the path from E to G if the DODAG is non-storing.
iii. List the path from G to H if the DODAG is storing.
iv. Assume, the link E–B drops and node E routes through node F. List the path from E to G in the new topology, if the DODAG is storing.
v. Give at least two reasons for using RPL in IoT settings, instead of other well-established routing protocols, such as RIP or OSPF.
Page 3 of 4
[3 marks ]
[4 marks ] [3 marks ] [3 marks ]
[3 marks ]
[2 marks ] [2 marks] [1 mark]
[2 marks ] [2 marks ]
3. ANSWER EITHER THIS QUESTION OR QUESTION 2
(a) Explain how authorisation is performed using the OAuth standard. Discuss the difficulties facing OAuth with IoT devices.
(b) A LoRaWAN network operates over a 125kHz bandwidth. Chirp spread spectrum modulation with spread factor SF = 7 and forward error correction (FEC) encoding with rate 3/4 is configured on all end devices. A maximum packet payload of 100 bytes is enforced, which is accompanied by a 20-bytes MAC overhead. 1% duty cycle is imposed by regulation.
i. Assuming no packet loss, what is the volume of useful data an end device transmits to a gateway in 30 mins?
ii. In case frame transmissions are lost with probability p = 10% due to channel errors or collisions, how much time will be required to transmit 2kB of data?
(c) Explain the LE Secure Connections pairing mechanism introduced with the Bluetooth 4.2 specification, considering IoT devices using the “Just Works” authentication method. Explain how the new features introduced solve the passive eavesdropping problems of LE Legacy Pairing.
(d) Consider an IoT system that hosts a microcontroller, which operates with three distinct phases: 1) low-power run, during which the microcontroller performs computations at a 50mW power consumption profile; 2) power- save, during which the microcontroller operates at with reduced performance to save energy, thereby consuming only 10mW; and 3) stand-by, during which most functionality is turned off and the microcontroller only listens for critical interrupts, thus consuming 2mW.
If the device is powered by a Li-Polymer battery with 2.5V nominal voltage, it spends 2% of the time in the first phase and 10% of the time in power-save mode, what should the battery capacity (in mAh) be to guarantee a device lifetime of 10 days?
[3 marks ]
[5 marks ]
[5 marks ]
[4 marks ]
Page 4 of 4
[8 marks ]