Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Enabling Reliable, Efficient, and Secure Computing for Energy Harvesting Powered IoT Devices

Xie, Mimi (2019) Enabling Reliable, Efficient, and Secure Computing for Energy Harvesting Powered IoT Devices. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

[img]
Preview
PDF
Download (7MB) | Preview

Abstract

Energy harvesting is one of the most promising techniques to power devices for future generation IoT. While energy harvesting does not have longevity, safety, and recharging concerns like traditional batteries, its instability brings a new challenge to the embedded systems: the energy harvested from environment is usually weak and intermittent. With traditional CMOS based technology, whenever the power is off, the computation has to start from the very beginning. Compared with existing CMOS based memory devices, emerging non-volatile memory devices such as PCM and STT-RAM, have the benefits of sustaining the data even when there is no power. By checkpointing the processor's volatile state to non-volatile memory, a program can resume its execution immediately after power comes back on again instead of restarting from the very beginning with checkpointing techniques.

However, checkpointing is not sufficient for energy harvesting systems. First, the program execution resumed from the last checkpoint might not execute correctly and causes inconsistency problem to the system. This problem is due to the inconsistency between volatile system state and non-volatile system state during checkpointing. Second, the process of checkpointing consumes a considerable amount of energy and time due to the slow and energy-consuming write operation of non-volatile memory. Finally, connecting to the internet poses many security issues to energy harvesting IoT devices. Traditional data encryption methods are both energy and time consuming which do not fit the resource constrained IoT devices. Therefore, a light-weight encryption method is in urgent need for securing IoT devices.

Targeting those three challenges, this dissertation proposes three techniques to enable reliable, efficient, and secure computing in energy harvesting IoT devices. First, a consistency-aware checkpointing technique is proposed to avoid inconsistency errors generated from the inconsistency between volatile state and non-volatile state. Second, checkpoint aware hybrid cache architecture is proposed to guarantee reliable checkpointing while maintaining a low checkpointing overhead from cache. Finally, to ensure the security of energy harvesting IoT devices, an energy-efficient in-memory encryption implementation for protecting the IoT device is proposed which can quickly encrypts the data in non-volatile memory and protect the embedded system physical and on-line attacks.


Share

Citation/Export:
Social Networking:
Share |

Details

Item Type: University of Pittsburgh ETD
Status: Unpublished
Creators/Authors:
CreatorsEmailPitt UsernameORCID
Xie, Mimimm.xie@pitt.edu
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairHu, Jingtongjthu@pitt.edu
Committee MemberDickerson, Samueldickerson@pitt.edu
Committee MemberXiong, Fengf.xiong@pitt.edu
Committee MemberYang, Junjuy9@pitt.edu
Committee MemberZhang, Youtaozhangyt@cs.pitt.edu
Date: 11 September 2019
Date Type: Publication
Defense Date: 30 May 2019
Approval Date: 11 September 2019
Submission Date: 8 July 2019
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 125
Institution: University of Pittsburgh
Schools and Programs: Swanson School of Engineering > Electrical and Computer Engineering
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Energy harvesting, IoT devices, Reliability, Efficiency, Security
Date Deposited: 11 Sep 2019 14:13
Last Modified: 11 Sep 2019 14:13
URI: http://d-scholarship.pitt.edu/id/eprint/37060

Metrics

Monthly Views for the past 3 years

Plum Analytics


Actions (login required)

View Item View Item