Projects

Research Projects

• Human-Centric Software Engineering (Active)

• Multilingual COVID-19 Fake News Detection and Intervention  (Active)

• Edge Workflow Systems  (Active)

• Edge based Smart UAV Delivery System  (Active)

• Fog/Edge Computing  (Active)

• Cloud Workflow System Design and Development

• Workflow Temporal Verification

• Cloud Workflow Data Management

• Cloud Workflow Scheduling

• Social Network/Context based App/Service Recommendation

PhD Scholarships and Positions for Research Associates are available. Please contact me for more details. 

If you are a researcher interested in collaboration on any of the research topics, or if you are a student interested in doing a research degree such as Honours, Master by Research or PhD, or just want to have a discussion on any of the projects, please feel free to contact me.

Please refer here for all funded projects

Human-Centric Software Engineering

Visit here for more details: https://sites.google.com/site/drxiaoliu/research/projects/human-centric-software-engineering 

• Humans are a key part of software development, including customers, designers, coders, testers, and end-users. While most current software engineering research and practices are function, data, or process-oriented, human-centric software engineering focuses on the human factors in engineering software systems. At present, there are major issues with misaligned software applications related to human factors, such as accessibility, usability, emotions, personality, age, gender, and culture. We aim to investigate enhanced theory, models, tools, and capability for next-generation human-centric software engineering aiming to achieve significant benefits of greatly improved software quality and user experience, developer productivity, and cost savings. Major research interests include but not limited to:

  • Human-centric requirements engineering

  • Incorporating human factors into requirements and design e.g., emotions, bias, personality, and culture

  • Context-awareness in human-centric software (and systems) engineering

  • Impact of human factors on development processes and software teams

  • Tools and models for capturing and interpreting user behaviours

  • Software applications that demonstrate the practice of human-centric software engineering

Fog/Edge Computing 

- Fog Computing (also known as Edge Computing ) is an architecture that uses one or more collaborative end-user clients or near-user edge devices to carry out a substantial amount of storage (rather than stored primarily in cloud data centers), communication (rather than routed over the internet backbone), control, configuration, measurement and management (rather than controlled primarily by network gateways such as those in the LTE core network). Due to the greate success and big market for Internet of Things (IoT), Fog Computing is becoming the next generation computing paradigm after the success of Cloud Computing. 

• Following the OpenFog Reference Architecture, the research in this area will cover everything we do in cloud computing with a shift of focus to the edge networks, including: architecture, resources management, intelligence, security, and software development. Especially, we are currently working on the following topics:

  • Setting up a fog computing testbed 

  • Data placement in fog computing to enable collaboration among edge nodes

  • Software development and testing platform for fog computing systems

  • Fog Workflow System

Figure: Fog Computing Grid Computing Architecture Example (Ref)

Ref: OpenFog Consortium, OpenFog Reference Architecture, https://www.openfogconsortium.org/ wp-content/uploads/OpenFog_Reference_ Architecture_2_09_17-FINAL.pdf 

Supported Projects:

• X. Liu and D. Yuan, Smart Data Placement for Enabling Collaboration Among Fog Computing Nodes, Faculty Research Grants Scheme, Deakin University, 2018. 

• X. Liu and Y. Xiang, Investigating How to Build Secure and Energy-Efficient Mobile Cloud Apps, 2017-2018.

Cloud Workflow System Design and Development

- Workflow system is the major software platform for running automated or semi-automated scientific process and business process. 

• Following the workflow reference model, the research in this area focuses on the design of system architecture, the communication between system components and services, the provisioning of resources, and management of service quality. 

• A family of prototype workflow systems have been designed and developed over the recent decades, including SwinDeW-G for grid, SwinDeW-C for cloud and SwinFlow-Cloud which is the latest prototype for running parallel business process in the cloud.

Figure: SwinFlow-Cloud System Architecture

Ref: D. Cao, X. Liu and Y. Yang, Novel Client-Cloud Architecture for Scalable Instance-Intensive Workflow Systems, Proc. of 14th International Conference on Web Information System Engineering (WISE2013), pages 270-284, Nanjing, China, Oct. 2013.

Supported Projects:

Lead by Prof. Yun Yang at Swinburne, Development of SwinDeW family - partly funded by ARC LP0990393 and DP0663841.

Workflow Temporal Verification

- Most real-world workflows are time-constrained, failure of on-time completion will lead to invalid execution results or significantly decrease customer satisfaction. Workflow temporal verification is the major approach to strictly verify and monitor the temporal consistency states of workflow applications and ensure their on-time completion.

• At workflow build time, based on statistics, the probability of on-time completion for the current workflow application against a given deadline can be estimated. If the probability is below a confidence threshold, additional resources or the relaxation of deadlines will be applied.

• At workflow runtime, the real-time temporal consistency states will be monitored and the workflow execution engine will be alerted if the probability temporal consistency state is below a confidence threshold. In such a case, workflow temporal violation handling will be applied to compensate for the time delays.

Figure: Probability based Temporal Consistency Model

Figure: Response-Time vs. Throughput based Workflow Temporal Checkpoint Selection

Ref: X. Liu, Y. Yang, Y. Jiang and J. Chen, Preventing Temporal Violations in Scientific Workflows: Where and How. IEEE Transactions on Software Engineering, 37(6):805-825, Nov./Dec. 2011.

Ref: X. Liu, D. Wang, D. Yuan, F. Wang and Y. Yang, Throughput Based Temporal Verification for Monitoring Large Batch of Parallel Processes, Proc. of International Conference on Software and Systems Process (ICSSP14), pages 124-133, Nanjing, China, May 2014.

Supported Projects:

Partly funded by ARC LP0990393 (Lead by Prof. Yun Yang) and NSFC 61672034 (AHU, China).

Cloud Workflow Data Management

- Generally speaking,workflow data includes the raw data required by the workflow application (which must be stored) and intermediate data generated during workflow execution. We are particularly interested in how to manage the workflow data effectively to improve the workflow execution efficiency and save the data storage cost in a cloud environment. 

• Cloud workflow data placement: given different physical locations of workflow data (some of them are movable and some of them are not), how to reduce the data transfer cost during workflow execution is a challenging issue. This issue can be addressed by workflow data placement strategy where workflow data are placed in advance to reduce data transfer overhead. 

• Computation and storage trade-off: some intermediate data especially scientific workflow data needs to be reused for validation or refinement purpose. For those intermediate data, it can be stored in the cloud, or it can be deleted and regenerated using data provenance information. In the later case, computation cost is required instead of storage cost. Given the data sizes, the provenance information (dataflow), and the usage frequency, how to achieve the minimum data storage cost is a challenging issue. This issue can be addressed by applying heuristics to search the whole space for possible delete-store options of all workflow data.

Figure: A matrix based k-means clustering strategy for Workflow Data placement

Figure: Data Dependency Graph to Cost Transitive Tournament 

Ref: D. Yuan, Y. Yang, X. Liu and J. Chen, A Data Placement Strategy in Cloud Scientific Workflows. Future Generation Computer Systems, Elsevier, 26(6):1200-1214, Oct. 2010.

Ref: D. Yuan, Y. Yang, X. Liu, W. Li, L. Cui, M. Xu and J. Chen, A Highly Practical Approach towards Achieving Minimum Datasets Storage Cost in the Cloud. IEEE Transactions on Parallel and Distributed Systems, 24(6):1234-1244, June 2013.

Supported Projects:

ARC DP110101340 and LP130100324 (Swinburne, Australia), NSFC 61300042 (ECNU, China)

Cloud Workflow Scheduling

- Workflow scheduling is the process of allocating workflow activities to workflow resources. As workflows have many constraints such as time, cost, reliability, security, energy and so on, workflow scheduling is the major approach to meet those constraints and achieve better performance. Since workflow scheduling is NP problem, heuristics and meta-heuristics based algorithms are required. In our current research, we focus on the following scheduling objectives

• Dynamic workflow scheduling algorithms to improve the workflow makespan or throughput.

• Static meta-heuristics based workflow scheduling algorithms to optimise multi-objectives, for example GA, ACO and PSO.

• Workflow rescheduling algorithms to address the temporal violation problem at workflow runtime, as part of the workflow monitoring framework.

• Energy-aware cloud workflow scheduling to save the energy consumption for running cloud workflows. 

Figure: Workflow Rescheduling Strategy for Local Workflow Segment to Tackle Temporal Violations

Ref: Z. Wu, X. Liu. Z. Ni, D, Yuan and Y. Yang, A Market‑Oriented Hierarchical Scheduling Strategy in Cloud Workflow Systems (pdf - 2185K). Journal of Supercomputing, Springer, 63(1):256-293, Jan. 2013.

Ref: X. Liu, Z. Ni, Z. Wu, D. Yuan, J. Chen and Y. Yang, A Novel General Framework for Automatic and Cost-Effective Handling of Recoverable Temporal Violations in Scientific Workflow Systems. Journal of Systems and Software, Elsevier, 84(3):492-509, Mar. 2011.

Supported Projects:

Partly funded by ARC LP0990393(Swinburne, Australia), NSFC 61300042 (ECNU, China)

Social Network/Context based App/Service Recommendation

- App (Mobile Apps) and service recommendation are interesting topics in the mobile/cloud computing area. Traditional recommendation techniques only utilise the data about the Apps/services and the customer profile. With the emergence of cloud computing and big data, recommendation systems can now process much larger data from various sources such as context data (time, location, mood) and customer's social networks (friends, friends' download history, influence). 

• App recommendation based on social networks

• User influence analysis in social networks

• Developer network and developer recommendation

                        Social Network                              Tencent App Store - App Ranking based on Friends  

Ref: Q. Wang, X. Liu, S. Zhang, Y. Jiang, F. Du, Y. Yue and Y. Liang, A Novel APP Recommendation Method Based on SVD and Social Influence,15th International Conference on Algorithms and Architectures for Parallel Processing, pp. 269-281, Zhangjiajie, China, November, 2015. 

Supported Projects:

APP Recommendation based on Social Network, CCF-Tencent Open Fund, 2014-2015.