October 2021

December 13th, 2021

Publications

  • Mohammad Sadegh Aslanpour, Adel Toosi, Raj Gaire and Aamir Cheema, “WattEdge: A Holistic Approach for Empirical Energy Measurements in Edge Computing”, ICSOC 2021 (accepted)

An efficient management of energy consumption of highly resource-limited edge computing nodes is a challenge. Past studies have limited or often simplistic focus on energy consumption factors considering computation or communication-only solutions, questioned by either costly hardware instrumentation or inaccurate software-specific limitations. This paper proposes a holistic and accurate energy measurement approach in RaspberryPi based edge computing infrastructure, thereby proposing an all-inclusive energy model. Our study identifies previously underestimated yet important factors and achieves a 95% accuracy in energy usage prediction.

  • Anton V. Uzunov, Matthew Brennan, Mohan Baruwal Chhetri, Quoc Bao Vo, Ryszard Kowalczyk, and John Wondoh: AWaRE2-MM: A Meta-Model for Goal-Driven, Contract-Mediated, Team-Centric Autonomous Middleware Frameworks for Antifragility, 28th Asia-Pacific Software Engineering Conference (APSEC 2021)
  • Zhi Zhang; Wei He; Yueqiang Cheng; Wenhao Wang; Yansong Gao; Minghua Wang; Kang Li; Surya Nepal and Yang Xiang. “BitMine: An End-to-End Tool for Detecting Rowhammer Vulnerability”. Accepted by IEEE Transactions on Information Forensics & Security (Core Rank: A, IF: 7.178). To the best of our knowledge, this paper provides the first end-to-end tool, coined BitMine, that systematically assesses a DRAM chip’s susceptibility to rowhammer bit flips. (Background: rowhammer is a destructive software-induced DRAM fault, which an attacker can leverage to break system security. Both individual customers and enterprise users (e.g., cloud providers) might refrain from using a computing system if it is vulnerable to rowhammer.)
  • Anahita Namvar, Chandra Thapa, Salil S. Kanhere, Seyit Camtepe, “Evaluating the security of machine learning-based IoT device identification systems against adversarial examples,” Accepted ICSOC 2021 (core rank A conference). This paper presented the adversarial attacks for IoT device classification under both the white-box and black-box attack scenarios. Furthermore, it studied the attack stealthiness by statistical test analysis and IDS system evaluation and found that these attacks are highly imperceptible in the IoT network.

Projects

 

  • CSCRC CCOP project has been delivered

.

Cyber attacks are becoming more frequent, sophisticated and targeted. They are often widespread and can go undetected, with criminals ‘preparing a network’ for extended periods in the lead up to an attack. This means that cyber security decision makers need to make faster, critical decisions to contain and mitigate cyber attacks. To keep their companies safe, boards and executives need to understand complex data about cyber security. This presents a challenge. Unfortunately, leaders often struggle to make swift, effective decisions because they are unable to easily assess risks and view the security status of the entire organisation.

Our researchers are developing ways to simplify this complex data so business leaders can easily comprehend, assess and respond to cyber threats. The CCOP platform will help improve executives and managers understanding of vulnerabilities and potential attacks, based on the level of security in their respective organisations. This tool is expected to allow leaders to interpret information more easily with less technical metrics, which will help them allocate resources, and map out an operations plan in response.

Our team is developing a prototype dashboard that captures and presents the current security status of an organisation’s systems, services and networks for its security staff and decision makers, including senior management, through a set of metrics and related visualisations. The security status presented by CCOP emerges from multiple cyber security information feeds, such as IDS (Intrusion Detection System) and SIEM (Security Information and Event Management) alerts, patching coverage, new critical vulnerabilities, and external sources of threat intelligence. The focus is on making sure executive decision makers have the correct information readily available, in a format that empowers them to make informed decisions.

The final deliverable includes the design and implementation of the Cyber security Insights from Real-time Twitter Analytics (CIRTA) tool, which provides real-time insights on cyber security threats, vulnerabilities, and attacks by monitoring and analysing a live Twitter stream. The tool collects cyber-security related tweets in real-time, applies supervised learning to classify them into different cyber security categories of relevance, extracts important keywords and topics from them, and makes the results available via a REST API. These results can be fed into a Cyber Common Operational Picture (CCOP) dashboard to support executive decision makers in making more informed decisions to manage their organisation’s cyber security.

With Marthie Grobler, Mohan Baruwal Chhetri and Seung Jang. https://research.csiro.au/distributed-systems-security/cyber-common-operating-picture-ccop/

  • CSCRC South Australia pilot project: Cyber security uplift has been delivered.

This report presents the main findings of a pilot cyber security maturity assessment done amongst selected South Australian SME participants. The assessments were done through targeted consultations with the SMEs with the specific aim to provide a high-level assessment of the cyber security posture of SMEs and to provide recommendations to improve overall cyber security maturity. This project is done in collaboration with the CSCRC, the Premier of South Australia, and 6 SMEs from South Australia. It is expected that the project report will publicly be launched at the end of November.

With Marthie Grobler and Regine Richelle.

  • US Army (CRYPTO FUNCTION DETECTOR (CFD) – ML/NLP BASED STANDARDS COMPLIANCE) project has been delivered.

This project developed machine learning (ML) and natural language processing (NLP) methods to detect software vulnerabilities and cryptographic primitives. The extension project aims to develop further methods to assess if cryptographic functions comply with standards and are used in authorized ways without introducing new vulnerabilities in applications. Standards compliant and authorized use of cryptographic methods is a crucial component of application security. Cryptographic functions form the foundation of security and trust. Strong cryptographic primitives are used to build robust protocols to establish the required level of security, privacy and trust that an application requires. However, very similar cryptographic primitives have also been commonly abused by adversaries to hide their malicious behaviours or to disguise destructive payloads such as a backdoor. For example, ROS-2 provides security services (authentication, access control and confidentiality) to applications through Data Distribution Service middleware. Hence, additional cryptographic primitives, which do not comply with the functional specification of the application, can be a sign of malicious intent.

Hence, this project focuses on the following objectives:

  1. Crypto Function Detection: This project developed methods that can detect cryptographic functions in application source codes (e.g., ROS applications) by utilizing language models such as Megatron-LM, which enhances BERT and GPT like multi-billion parameter model using model parallelism.
  2. Crypto Function Compliance: The extension project will use the evaluated language models further to develop methods that can assess if cryptographic functions comply with standards so that they won’t be introducing new vulnerabilities

The project showed that large NLP models (up to 1.5 Billion parameters), which are trained within natural languages domain (e.g., text data from the Internet), can effectively be tuned to process programming language to detect different security features. The project first tested these models for vulnerability detection with established datasets compared to the existing methods. Then transformed positive results to crypto function detection with high accuracy.

With Seung Jang, Ejaz Ahmed, Chanda Thapa, Seyit Camtepe, Surya Nepal, Josef Pieprzyk,

Join us

  • Postdoctoral Fellow opportunities

https://jobs.csiro.au/job/Sydney%2C-NSW-CSIRO-Postdoctoral-Fellowship-in-Collaborative-Intelligence-Cybersecurity/802708400/?locale=en_GB

https://jobs.csiro.au/job/Various-Data61-PhD-Scholarships/796808000/?locale=en_GB

Students

  • let’s meet one of our latest students: Ivo Murdoch

Ivo is a PhD student at the University of Queensland in UQ Cyber Security. His research centres on improving neural network interpretability for use in human centric cyber security applications and defence against adversarial attacks. Ivo applied for a Data61 scholarship after finding a project with M.A.P. Chamikara and Marthie Grobler which aligned heavily with his research interests. Outside of study, Ivo enjoys gardening, reading, playing guitar, and training judo.

Achievements/ Good news

  • Chamikara won “2021 RMIT Prize for Research Excellence – Higher Degree by Research (Technology) ”RMIT Prize for Research Excellence – Higher Degree by Research (Technology) recognizes an outstanding HDR candidate who has achieved the highest levels of excellence in research during the completion of their degree in the category of Technology. The selection of the prize recipient is by the Deputy Vice-Chancellor Research and Innovation (R&I) based on recommendations of the Awards Panel, RMIT University. The award comprises a cash award, a certificate signed by the Deputy Vice-Chancellor (R&I), a permanent record of the award on an appropriate webpage (Honour Board) of the University, and a commemorative trophy.
  • Sunil Singh Samant has been awarded his PhD from Swinburne University of Technology and is now working as a Technology Analyst with Accenture Australia. Congratulations Sunil for this milestone.

Events

  • Data61 and DSTG Cyber Security Summer School

Bringing together the Australian cyber security ecosystem

Due to the recent announcements and future dates set for the relaxing on domestic and international border restrictions, the organising committee has made the decision to move our forthcoming event to 21 to 23 February 2022. This event will still be held in a hybrid online/in-person mode, but with the added benefit of an increased proportion of in-person participation, enabled by the new travel rules.  We hope that you are as excited as us at the possibility to meet in person, even if it means a change in date.

The CSIRO’s Data61 & DSTG Cyber Security Summer School (CSSS) (https://research.csiro.au/csss/), in collaboration with the University of Queensland, is an annual summer school focusing on a range of cyber security topics, bringing the Australian cyber ecosystem together.

The CDNG Technology and Science Conference (https://wp.csiro.au/cdng/), in collaboration with Macquarie University, is a first of its kind scientific and technology focused Cyber Defence Conference, with the aim to have the community of Cyber-Security under the “Next Generation Defence Technologies | Cyber Research” program in collaboration with DST-Data61 and partner universities meet together, expose relevant research activities and discuss future collaborations.

This year, the two events will jointly be hosted at Customs House Brisbane, 21 to 23 February 2022.

Attendance is free, subject to invitation.

https://research.csiro.au/csss/

  • SME and Cyber workshop

Raising the bar of cyber security awareness among Australian SMEs

This workshop is crafted with the needs identified by SMEs in this domain in mind and is aimed at bringing practical and relevant research and developments directly to the SMEs.

Strengthen your current understanding of cyber security, engage with cyber security specialists, and learn through practical exercises and demos how to best tackle cyber security challenges within your organisation. Through attending this workshop, you will start your journey to build a stronger cyber security posture.

You will be equipped to better:

  • Recognise cyber threats and emerging cyber trends;
  • Understand the impact of cyber security on your organisation;
  • Build a network of cyber security connections;
  • Focus on cyber security upliftment.

This 3-hour workshop is free to attend, but will require registration.

https://www.csiro.au/en/work-with-us/funding-programs/programs/Innovate-to-Grow/Cyber-security-awareness for registration

 

  • The Human Centric AI Seminars Series

The Human Centric Security team are running a new monthly series “The Human Centric AI Seminars” that will focus on various research topics in human centered AI.

For more info contact: Kristen Moore and Tina Wu, https://research.csiro.au/distributed-systems-security/the-human-centric-ai-seminars-series/

Free access to anyone interested in Humans and AI

  • Join us for our monthly SAO seminars in collaboration with the Cyber Security CRC.

 

 

Talks/conferences

  • Zhi Zhang has been Invited to be a PC member in Australasian Information Security Conference (AISC) 2022.
  • Marthie Grobler presented on Trends and Innovation in Cyber at the Innovate to Grow: Cyber masterclass, the opening of a 10-week program for Small to Medium Enterprises (SMEs) to investigate new R&D opportunities related to Cyber Security solutions. The course is open to established SMEs who are either interested in investigating new R&D opportunities or are in the early decision-making stages about engaging with a research organisation on a new project. Muhammed Esgin, Sharif Abuadbba and Chandra Thapa serves as SME mentors for this program.

HS-Me Day

HS-Me Day 2021 was on Tuesday 19 October. It was a day for all staff to reconnect in their own way… with their health, personal safety and environment.

CSIRO people were given the opportunity to plan their day with a variety of activities to rejuvenate mind and body, and enjoy the local environment.