On real security of cryptographic primitives: from theory to practice
Contents TOC o “1-3” h z u Summary PAGEREF _Toc77941807 h 4Introduction PAGEREF _Toc77941808 h 6Data Analaysis AS a Service. PAGEREF _Toc77941809 h 6PROCESS ANALYTICS PAGEREF _Toc77941810 h 7PRIVACY ISSUES├┐├┐IN DATA ANALYTICS AS A SERVICE PAGEREF _Toc77941811 h 7MARKETING ANALYTICS PAGEREF _Toc77941812 h 7PROBLEM STATEMENT PAGEREF _Toc77941813 h 9THESIS CONTRIBUTION PAGEREF _Toc77941814 h 9REFERENCES PAGEREF _Toc77941815 h 9The Real Security of Cryptographic Primitives PAGEREF _Toc77941816 h 131.0Introduction. PAGEREF _Toc77941817 h 131.2 Practical Cryptanalysis. PAGEREF _Toc77941818 h 151.3 practical cryptanalysis of sha-1 on ASIC 9 PAGEREF _Toc77941819 h 161.4 Cost Implications of ASIC hardware crackers: a sha-1. PAGEREF _Toc77941820 h 181.5 Hash Functions and Cryptanalysis PAGEREF _Toc77941821 h 191.6 Hardware Birthday Cluster. PAGEREF _Toc77941822 h 19Hardware for Birthday Slaves – Conceptual Design PAGEREF _Toc77941823 h 191.7 Cluster Nodes PAGEREF _Toc77941824 h 201.8 Hardware Differential Attack Cluster Design PAGEREF _Toc77941825 h 201.9 Neutral Bits PAGEREF _Toc77941826 h 212.0 Storage PAGEREF _Toc77941827 h 212.1 Architecture PAGEREF _Toc77941828 h 222.2 Chip Design PAGEREF _Toc77941829 h 222.2.1 Chip Architecture. PAGEREF _Toc77941830 h 223.1 ASIC Fabrication and Running Cost PAGEREF _Toc77941831 h 23Birthday Attack(2/64) PAGEREF _Toc77941832 h 23Birthday Attack (2/80) PAGEREF _Toc77941833 h 24Prefix for Differential Collision Attack PAGEREF _Toc77941834 h 25Side-Channel Attacks PAGEREF _Toc77941835 h 262.0 Introduction PAGEREF _Toc77941836 h 262.1.1 The source of the leaks PAGEREF _Toc77941837 h 262.1.2 Models of leakage PAGEREF _Toc77941838 h 262.1.4 Leakage models PAGEREF _Toc77941839 h 272.2 ATTACKS ON TIMING PAGEREF _Toc77941840 h 272.3 DESIGN CONSUMPTION COULD BE IMPOSSIBLE PAGEREF _Toc77941841 h 292.4 TIMING ATTACKS PAGEREF _Toc77941842 h 302.4.1 Simple modular exponentiator cryptanalysis PAGEREF _Toc77941843 h 312.5 POWER CONSUMPTION ATTACKS PAGEREF _Toc77941844 h 322.5.1 Analysis of Differential Power (DPA) attacks PAGEREF _Toc77941845 h 332.6 ATTACKS ON THE DIFFERENTIAL FAULT ANALYSIS (DFA) PAGEREF _Toc77941846 h 342.7 ATTACKS ON CHANNELS PREVENTION PAGEREF _Toc77941847 h 342.8 COUNTERMEASURES FOR ATTACKS IN GENERAL PAGEREF _Toc77941848 h 342.9 POWER ANALYSIS ATTACKS PAGEREF _Toc77941849 h 352.9.1 Signal Size Reduction. PAGEREF _Toc77941850 h 362.9.2 Alterations to algorithmic design are being considered. PAGEREF _Toc77941851 h 362.9.3 Double-Encryption PAGEREF _Toc77941852 h 36LIGHTWEIGHT IMPLEMENTATIONS PAGEREF _Toc77941853 h 373.0 Introduction PAGEREF _Toc77941854 h 373.1 Symmetrical Chips PAGEREF _Toc77941855 h 373.2 Implementation of symmetric cyphers in computer software PAGEREF _Toc77941856 h 383.3 Asymmetric Chips PAGEREF _Toc77941857 h 383.4 A lightweight Elliptic Curve Engine PAGEREF _Toc77941858 h 393.5 Inversion PAGEREF _Toc77941859 h 393.6 Multiplication point PAGEREF _Toc77941860 h 393.7 Hardware and software code Design in ECC’s PAGEREF _Toc77941861 h 393.8 Implementation of the ECC software PAGEREF _Toc77941862 h 403.9 The elliptic curve in arithmetic. PAGEREF _Toc77941863 h 414.0 Fault Injection Attacks and Countermeasures PAGEREF _Toc77941864 h 41DISCUSSION PAGEREF _Toc77941865 h 43ACHIEVEMENTS PAGEREF _Toc77941866 h 43REFLECTION PAGEREF _Toc77941867 h 44FUTURE WORK PAGEREF _Toc77941868 h 44
Most of today applications are enhanced by connectivity features enabling unprecedent possibilities to improve our everyday life. New connectivity protocols such as 5G or LoraWan are quickly adopted to interconnect objects, devices and sensors. On one hand, Internet of Everything promises many benefits (eg. optimised supply-chains, optimised energy ressource utilisation, better management of health crisis, simplification of payment process).
On the other hand personal information about users daily routines is collected and is vulnerable to privacy violation issues. Moreover, many applications (military, financial or automotive applications ) are security sensitive.
The information stored on a device is vulnerable to privacy violation by compromising nodes existing in an IoT network. Connected devices can be deployed in an hostile environment, i.e and adversary has physical access to or control over the devices, enabling physical attacks.
Thats? why connected devices implements cryptographic algorithms and protocols to garan- tee the confidentiality and authenticity of information exchanges.
Cryptgraphic algorithms are designed to be robust against attacks such as differential cryptanalysis at least during the lifetime of the device. Cryptographic standards are used to allow the interoperability between devices over the internet. But with the evolution of the cryptanaysis research and the increasing computation power folowing Moor?s Law, a cryptographic standard such as SHA-1 can be cryptographicaly broken but takes times time to deprecate due to interoperability issues.
During this thesis the first chosen-prefix collision SHA-1 ASIC prototype cracker is designed and compared with a GPU-based solution. We show that modern technologies allow to compromise up to a 80-bit security cryptosystem. Even though more modern ciphers such as SHA-2 or AES seems to be resistant against attacks such as linear differential cryptanalysis, it might be possible to attack the implementation.
In Internet of Everything, throughput is usually not a problem but energy, power and area are sparse ressources. Mass-produced microcontrollers can be used to provide the ability of an update, and to reduce the time to market.
This thesis have been carried out during the NIST lightweight competition that aims at creating a new standard of lightweight cryptgraphic primitives for Internet of Everything and connected ressource constrained devices.
I demonstrate that the bloc cipher GIFT that is the basic bloc of several NIST lightweight cryptograhy comptetition candidates provides very good embedded software performances compared to AES and is even close to the NSA lightweight ciphers SIMON ans SPECK both in masked un unmasked setting.
Finaly, different advanced side-chanel metrics and analysis technics for software protected implementations are presented. We show that when an attacker has full knowledge of the underlying implementation, the security of a masked software implementation doesnt neces- sarility grows with the masking order.
The last considered aspect in this thesis is the fault injections attacks. We demonstarte the feasability of Electromagnetic fault injections on microcontroller and a physical sensor based detection technic.
Big data became popular for companies in the last decade as data generated by new computer technologies increased considerably. 2.5 bytes of quintillion were generated everyday in 2016 [1]. This quantity is expected to amount to 146,880 GB per person by 2020[2]. The huge volumes of computer data are not idle. Companies are collecting and analyzing data to improve services and goods, understand the conduct of their customers and reduce the danger of cyber attacks[3]. Companies have taken data analytics instead of a supplemental tool as a business booster. In 2018, 59% of companies employed data analysis technology with certain industries, such as telecoms, reaching 90%[4]. Companies struggle to analyze data. A lack of internal experience is a problem with data analysis. Companies, in particular small and medium-sized ones, do not have professionals in data analytics. Companies can either hire new staff at greater costs or train existing staff at higher costs. Managers hesitate to fund teams for data analysis[5]. Another problem in-house data analysis is the lack of computational resources. Without a solid computer infrastructure, investment in human resources is insufficient. In addition, even if firms have sufficient human and computer resources, it’s tough to keep up with the latest advancements in data analysis.
Data Analaysis AS a Service.
Outsourcing data analytics jobs can help overcome the limitations of in-house data analytics. The Data Analytics-as-a-Service (DAaaS) model gives companies data.
Outsourced analytics services [7]. In this model, a data analytics company provides its expertise and computing resources to other organizations that want data analytics. Companies like data outsourcing because it saves money and time. Delegating analytics to a professional firm leads in better analytics since the expert firm has the latest analytics technologies and is aware of industry advancements [6]. Furthermore, an external team can offer accurate results without prejudice [6]. With all the benefits, organizations are opting for outsourced analytics, which will account for over half of all data analytics work by 2019 [8]. Notably, data science experts do not encourage firms to totally outsource their data analytics duties. On should outsource certain analytics apps and develop an in-house data analytics team for the remainder [6]. So they can react faster in an emergency while keeping up with the latest data analytics technology. Several firms have acquired outsourced analytics applications.
Process analytics helps businesses improve their business processes [9]. Companies may identify bottlenecks in their systems and take action by visualizing business processes. Process analytics is useful for analyzing internal workflow, clarifying job duties, and internal auditing. Many firms prioritize process efficiency [10]. Process mining is a popular method for discovering,…

error: Content is protected !!