IEEE/IFIP Network Operations and Management Symposium
25-29 April 2022 // Budapest, Hungary
Network and Service Management in the Era of Cloudification, Softwarization and Artificial Intelligence

WS4 - 1st International Workshop on Technologies for Network Twins (TNT 2022)

All times listed are in Central European Summer Time (CEST)

FRIDAY, 29 APRIL 2022

09:00-09:45 Opening and Keynote

Room Boróka II, Chair: Laurent Ciavaglia

 

Keynote – Network Requirements for Digital Twins and Digital Twin Requirements for Networks

Abstract: The widescale deployment of Digital Twins depends on Network infrastructures and capabilities that consist of at least four elements. The first of these is a Network of Data underpinned by technologies that provide access to distributed raw data sources and a semantic fabric for extracting meaning and value from the data. The next three are the traditional infrastructures for Communications, Computing, and Storage. Each of these infrastructures is in turn constructed as a heterogenous hierarchy defined by timescale and by spatial distribution. Much of this is driven by the requirements for modeling and analytic techniques inherent to Digital Twins. Digital Twins push the envelope on the requirements that the Networks must deliver and are characterized by a new level of complexity for the systems that manage these Networks in concert. The first part of the talk will address the requirements for Digital Twins and will be followed by how Digital Twins can be applied to aspects of Network Management. The specific example will focus on how we currently manage resources in Cellular Wireless Networks and how the use of Digital Twins can fundamentally alter what is possible. This last point illustrates how distributed Data, Communications, Computing, and Storage are intertwined and should be thought of as a single managed infrastructure.

Adam DrobotAdam Drobot, OpenTechWorks, USA

Adam is the Chairman of the Board of OpenTechWorks, Inc. He serves on the Boards of multiple early-stage companies. His activities are strategic consulting, start-ups, non-profits, and industry associations. In the past he was the President of Applied Research at Telcordia Technologies (Bellcore) and the company’s CTO, and before that the Senior Vice President for Science and Technology at SAIC/LEIDOS.

Adam is a current member of the FCC Technological Advisory Council where he Co-Chairs the Working Group on Artificial Intelligence; the University of Texas Physics Department Advisory Council; and Chairman of FAMES USA a non-profit that organizes programs to attract disadvantaged youths to careers in STEM fields.  In the past, he was on the Boards of the Telecommunications Industry Association (TIA) where he Chaired the Technology Committee; the Association for Telecommunications Industry Solutions (ATIS); the New Jersey Technology Council; the US DoT ITS Program Advisory Committee, and the University of Michigan Transportation Research Institute External Advisory Board.

Over the years he has been active in IEEE and served as the Chair of, multiple Committees including the IEEE IoT Activities Board, and multiple major conferences such as the IEEE World Forum on IoT. He is currently a member of the IEEE Press Editorial Board.  He has published over 150 journal articles and holds 27 patents. He is a fellow of the American Physical Society. His degrees include a BA in Engineering Physics from Cornell University and a PhD. in Plasma Physics from the University of Texas.

Panel - Hot Topics in Network Digital Twins

Panelists:

Chris Janz, Huawei, Canada

Chris Janz is Director of Huawei’s Optical Systems Competency Centre in Ottawa, Canada.  He has more 25 years of experience in the optical networking industry and has made contributions in many areas including III-V optical components, optical transmission and switching and network operations software.  He has worked for several companies, including Alcatel and Ciena, in the UK, France and Canada.

Chris has been a prominent figure in the SDN area for many years and was an early shaper and advocate of industry thinking on SDN implications and possibilities for transport networks and of the potential for automation and intelligence to improve operations costs and outcomes on increasingly software-driven networks.

Chris has a Ph.D. in electrical engineering from the University of Alberta, a M.B.A. from Queen’s University in Kingston, Ontario and a B.Eng. in engineering physics from the Royal Military College of Canada. He has authored or co-authored more than 100 papers or presentations in journals and conferences and holds many patents.

Albert Cabellos, Universitat Politècnica de Catalunya (UPC)

Prof. Albert Cabellos (PhD 2008), Full Professor at Universitat Politècnica de Catalunya.

In 2019 he co-founded the Barcelona Neural Networking (BNN) Center (https://bnn.upc.edu) where he is the Director. BNN’s has been created with the main goals of carrying fundamental research in the field of Graph Neural Network applied to Computer Networks and educating and training the new generation of students. He has also founded the Open Overlay Router open-source initiative (http://openorverlayrouter.org) along with Cisco.

He has been a visiting researcher at Cisco Systems and Agilent Technologies and a visiting professor at the Royal Institute of Technology (KTH) and the Massachusetts Institute of Technology (MIT), National Institute of Informatics (Tokyo) and UC Berkeley. He has participated in several national, EU (FP7, H2020), USA (NSF) and industrial R&D projects.

Cheng Zhou, China Mobile

Cheng ZHOU, project manager at China Mobile Research Institute. He has been engaged in research on technologies in data communication filed for 18+ years. His current research interest includes AI enabled network, network digital twin, intelligent routing, and other related networking technologies.

Tao Sun, China Mobile (panel moderator)

Dr. Tao Sun received the B.S. degree with the Department of Automation, in 2003 and Ph.D. degree in Control Science and Engineering, in 2008, both from Tsinghua University, Beijing, China. He jointed China Mobile Research Institute from 2008, serving as Chief Expert since 2021. Dr. Sun acted as China Mobile 3GPP delegates since 2009. He was rapporteurs of many 3GPP work items such as Study of Next Generation Mobile Architecture, 5G System Architecture Phase 1, Enhancement of Service-based Architecture. He now is the Vice Chair of 3GPP SA2 and China Mobile’s coordinator of 3GPP SA and CT groups. Dr. Sun has authored or coauthored about 30 papers including IEEE TAC, IEEE TASE, IEEE TII, China Communication, etc. He has been a Program or Session Chair IEEE EDGE, IEEE CLOUD, IEEE DTPI, CCF ICSS. He gave invited talk in WOCC 2019


Questions:

  1. What is your definition of a network digital twin?
  2. What makes network digital twins different from traditional simulation?
  3. Beyond AI, what other enabling techniques or technologies are essential to build and operate digital twins? What are some of the major aspects impacting the development of network digital twins?
  4. What role standardization should or could play to enable network digital twins?
  5. What role may network digital twins have in education, especially for technical and professional education?

 

Additional discussion points:

A. It may happen that the effort needed to build an NDT (Network Digital Twin) that tries to perfectly replicate the physical system is higher and more complex than the required to set up the real network.

How can we cope with this issue?

Should we establish a modular architecture that allows us to build or enhance an NDT as we were using Lego bricks?

Can we assume a minimal granularity of these Lego bricks? (e.g., small bricks such as routers, or use coarser granularity such as NFVs or even slices)

Can we identify some of these basic building blocks?

Is there a minimal subset of these components that should be mandatory in any NDT?

 

B. The bidirectional communication between the digital and the physical system is crucial to allow real-time a) replication (from physical to virtual) and b) reconfiguration (from virtual to physical).

I guess we all agree that it is needed to standardize this channel to foster the interconnection of existing (or future) real networks with the emerging NDT solutions.

Should we start working in this standardization process ASAP or maybe is more efficient/practical to wait until some products dominate the market to adopt them as de-facto standard? (Pros and cons)

 

C. An alternative to the modular NDT can be to build partial views (or projections) of the real network. For example, an NDT could only consider the AI related elements and other could be focused only on high level orchestration processes.

Can we identify some of these dimensions or projections of the real network that could be useful to replicate in an NDT? (e.g., AI, orchestration, cybersecurity)

Papers:

Smart DC: An AI and Digital Twin-based Energy-Saving Solution for Data Centers
 
Authors:
Ziting Zhang, China Telecom , P.R. China
Yu Zeng, China Telecom , P.R. China
Haoran Liu, China Telecom , P.R. China
Chaoyue Zhao, China Telecom , P.R. China
Feng Wang, China Telecom , P.R. China
Yunqing Chen, China Telecom , P.R. China
 
Stopping the Data Flood: Post-Shannon Traffic Reduction in Digital-Twins Applications
 
Authors:
Caspar von Lengerke, Technical University Dresden, Germany
Alexander Hefele, Technical University Dresden, Germany
Juan Cabrera, Technical University Dresden, Germany
Frank Fitzek, Technical University Dresden, Germany
 
A Chaos Engineering Approach for Improving the Resiliency of IT Services Configurations
 
Authors:
Filippo Poltronieri, University of Ferrare, Italy
Mauro Tortonesi, University of Ferrare, Italy
Cesare Stefanelli, University of Ferrare, Italy
 
Digital Twin for the Optical Network: Key Technologies and Enabled Automation Applications
 
Authors:
Christopher Janz, Huawei Technologies, Canada
Yuren You, Huawei Technologies, Canada
Mahdi Hemmati, Huawei Technologies, Canada
Zhiping Jiang, Huawei Technologies, Canada
Abbas Javadtalab, Huawei Technologies, Canada
Jeebak Mitra, Huawei Technologies, Canada
 
Accelerating Deep Reinforcement Learning for Digital Twin Network Optimization with Evolutionary Strategies
 
Authors:
Carlos Güemes Palau, Universitat Politècnica de Catalunya, Spain
Paul Almasan, Universitat Politècnica de Catalunya, Spain
Pere Barlet Ros, Universitat Politècnica de Catalunya, Spain
Albert Cabellos Aparicio, Universitat Politècnica de Catalunya, Spain
 
B5GEMINI: a Digital Twin Network for 5G and Beyond
 
Authors:
Alberto Mozo Velasco, Universidad Politectnica de Madrid, Spain
 

Patrons

Supporters