Organizer: Dawei Chen, Toyota Motor North America, Ziran Wang, Purdue University

Description: The recent development of cloud computing and edge computing bring forward numerous novel technologies that could play a critical role in realizing the vision of future Connected and Automated Vehicles (CAVs). By enabling CAVs to offload their massive on-board data and heavy computing tasks to road side units (RSUs) at the edge and computing clusters in the cloud, future automotive computing could become more efficient, collaborative, and lower cost. With more communication and sharing via cloud and edge infrastructure, different entities of the future transportation system (e.g., vehicles, road infrastructure, traffic management centers) would become more situation-aware, thus improving the safety, mobility, and environment impact of the entire system. However, to realize this vision, there remain significant challenges to be addressed. This special session will bring four experts from academia and industry to present some of the pressing challenges in cloud- and edge-computing of CAVs and discuss promising directions to address them, including federated learning with multi-access edge computing for CAVs, data management with smart caching at vehicles and wireless network management with dependability constraints, digital twin based design automation tools for development efficiency, and end-to-end methodologies for addressing safety and security in edge-cloud computing of CAVs.

Presentations: Yanzhi Wang, Northeastern University, Linghao Song, University of California, Los Angeles

Title: Federated Learning and Analysis with Multi-access Edge Computing for Connected and Automated Vehicles
Authors: Zhu Han, University of Houston

Title: Vehicle as a Cache - Edge Computing for Automated Data Centric Vehicles
Authors: Rolf Ernst, Technische Universität Braunschweig

Title: Autonomous Driving Digital Twin Empowered Design Automation: An Industry Perspective
Authors: Bo Yu, BeyonCa

Title: Waving the Double-Edged Sword: Building Resilient CAVs with Edge and Cloud Computing
Authors: Qi Zhu, Northwestern University

Organizers: Deming Chen, University of Illinois at Urbana-Champaign, Priya Nagpurkar, IBM

Description: Cloud computing provides on-demand availability of computing resources for users. It allows enterprises to run their applications with improved manageability, less maintenance, and reduced overall cost. In addition, cloud computing is able to meet fluctuating computing demand through data/compute "bursting” across cloud boundaries. According to data reported by leading market analysis companies, the global spending on cloud computing services has reached $706 billion now and expected to reach $1.3 trillion by 2025. Important new cloud computing paradigms have emerged as well, including hybrid cloud and multi-cloud.

Designing efficient, performant, and secure cloud systems requires a holistic and systematic approach and novel interdisciplinary research. This is due to the unique challenges faced by future clouds: high system complexity, difficulty for test, evaluation, and monitoring, vulnerability from security attacks and data leakage, unpredictable latency, non-ideal system configurations, high energy and carbon emission, and complicated communicating interfaces across multi-clouds. Each of these challenges, if not solved well, will become a roadblock for the future growth and promises of cloud computing.

In this special session, four prominent industrial leaders with excellent cloud computing expertise from large cloud or computing hardware vendors will share their unique insights, vision, and solutions for the design of the next-generation cloud computing systems. They will focus on grand future opportunities, challenges, and promising research solutions and directions. The special session will provide a holistic and interdisciplinary coverage for the design of the next-generation cloud computing systems in areas including application and algorithm design, AI for system design, distributed computing, big-data storage, cloud computing infrastructure, computer architecture and accelerator design, security and privacy, and design automation across different system stacks.

Presentations:

Title: AI Software Is Leading the Hardware Development in Cloud Computing
Authors: Ivo Bolsens, AMD

Title: Architecting Products for the Cloud
Authors: Annie Foong, Intel

Title: Architecture and Software Stack for the Next-Generation Cloud Computing Systems
Authors: Wen-Mei Hwu, NVIDIA

Title: Engineering the Future of Cloud Software with AI
Authors: Ruchir Puri, IBM

Organizer: Jingtong Hu, University of Pittsburgh, Peipei Zhou, University of Pittsburgh

Description: Cloud computing is an increasingly popular way to use computing and storage technologies, and it is changing the way businesses deliver services. In order to provide scalable, reliable, secure, and cost-effective services, including machine learning, Searching, FinTech, etc., service providers need to constantly innovate the hardware infrastructure design in the Clouds. In this Special Session, we have three speakers from Intel, Microsoft, and Google to talk about their latest progress in achieving energy-efficient, secure, and reliable hardware infrastructures, which covers Interconnect designs, Infrastructure Processing Units (IPUs), and TPUs.

Presentations: 

Title: Compute Express Link (CXL*): Open Interconnect for Cloud Infrastructure
Authors: Debendra Das Sharma, Intel

Title: Moving the Center of Cloud Servers from Processors to IPUs
Authors: Derek Chiou, Microsoft and UT Austin

Title: Codesign from Semiconductors to AI
Author: Cliff Young, Google