[Nfvrg] NFVRG interim meeting -- Model-driven & AI-Enabled Inter-Cloud Optimization

[Ramki Krishnan <[e-mail]>]

Dear All,

Please find details on the NFVRG interim meeting on "Model-driven & AI-Enabled Inter-Cloud Optimization". Looking forward to your participation.

--
Thanks,
NFVRG Co-Chairs (Ramki & Diego)

Zoom:                                vmware.zoom.us/j/3055973130
VMware address:            Promontory A, 3401 Hillview Ave, Palo Alto, CA 94304
VMware campus map:   https://www.vmware.com/files/pdf/vmw-directions-to-vmware.pdf


Title: Model-driven & AI-Enabled Inter-Cloud Optimization

  *   Introduction
     *   "Use Cases (5G/Edge Computing)"                                                     - Dilip Krishnaswamy (Reliance Jio, India)                           - 20 min.

"Architecture & Benefits"                                                                     - Sastry Isukapalli (AT&T, New Jersey, USA)                        - 25 min.

  *   Research Challenges & Solution Direction
     *   "Recent Trends in Constraint Optimization and Satisfaction"               - Nina Narodytska (VMware, California, USA)                    - 40 min.
     *   "Network Flow Optimization"                                                              - Siamak Layeghy (Univ. of Queensland, Australia)              - 40 min.


Agenda Details:
"Use Cases (5G/Edge Computing)" - Dilip Krishnaswamy (Reliance Jio, India)
Abstract: Emerging 5G network architecture will consist of a large number of distributed "clouds" with dynamically varying resource availability in terms of compute, or storage, or network bandwidth, or energy, and with support for edge computing/services. The ability to serve users with dynamic requirements across geographies using such distributed resources will require learning  and prediction of distributed models across users and infrastructure to dynamically and efficiently allocate resources. This talk will cover different use-cases that drive the need for such distributed architecture, and motivate the need for building distributed models of users and infrastructure, and the need for dynamic resource optimization across these distributed and edge clouds.

"Architecture and Benefits" - Sastry Isukapalli (AT&T, New Jersey, USA)
Abstract: This talk focuses on the Open Network Automation Platform (ONAP) Optimization Framework (OOF) open source effort, which is a declarative framework based on a Policy- and Model-driven, architecture with AI assists. While the core design framework is based on MiniZinc, it also supports optimization applications based on generic or problem-specific custom solvers. The system is presented with a set of example applications, including dynamic resource allocation under various complex constraints.



"Recent Trends in Constraint Optimization and Satisfaction" - Nina Narodytska (Vmware, California, USA)
Abstract: In this talk we overview several constraint optimization solvers and consider the main design principles behind these search procedures. First, we look at Boolean Satisfiability (SAT) and Integer Linear Programing( ILP) solvers and their extensions. These decision procedures  offer a restrictive modeling language based on Boolean logic and integer linear constrains, and powerful search procedures. Next we will talk about Constraint Programing (CP) solvers, whose rich modeling languages include expressive global constraints and a variety of decision search procedures, e.g., backtracking search, lazy-clause generation,  etc.



"Software-defined Constrained Optimal Routing (SCOR)" - Siamak Layeghy (Univ. of Queensland, Australia) (* to be finalized)
Abstract: This presentation covers an overview of SCOR (Software-defined Constrained Optimal Routing), a new Software Defined Networking (SDN) Northbound Interface for QoS routing and traffic engineering implemented in the MiniZinc modelling language.  A key feature of SCOR is that it is declarative, where only the constraints and utility function of the routing problem need to be expressed, and the complexity of solving the problem is hidden from the user, and handled by a powerful generic solver.


Dilip Krishnaswamy (Reliance Jio, India) Bio:
Dilip Krishnaswamy is currently serving as Vice President (New Technology Initiatives) at Reliance Jio Infocomm Ltd. Previously, he has worked as a senior scientist at IBM Research, as a platform architect at Intel, and as a senior staff researcher in the office of the chief scientist at Qualcomm. He has a PhD in electrical engineering from the University of Illinois at Urbana-Champaign. He is an inventor on 61 granted US patents, and has 70+ publications.  His research interests include distributed information processing, distributed function virtualization, distributed resource management, blockchain technology, edge services, NFV/SDN,  NB-IoT, AI/ML, and 5G architecture & systems.



Sastry Isukapalli (AT&T, New Jersey) Bio:
Sastry Isukapalli is Principal Inventive Scientist at AT&T Labs Research. Previously, he was part of the Machine Learning and Data Mining groups in AT&T Research. Prior to the AT&T tenure, he was Assistant Professor at the RW Johnson Medical School. He obtained his PhD from Rutgers University for development of the Stochastic Response Surface Method under the direction of Prof. Panos Georgopulos. He is also an Alumnus of the Indian Institute of Technology, Madras, India. He is an alumni of Savera Systems, Virtual Gold, and Kirusa Inc. He is a committer for the open source ONAP Optimization Framework Project (onap.org).



Nina Narodytska (VMware, California, USA) Bio:
Nina Narodytska is a researcher at VMware Research. Prior to VMWare Research, she was a researcher at Samsung Research America, a postdoctoral researcher in the Carnegie Mellon University School of Computer Science and the University of Toronto. She received her PhD from the University of New South Wales in 2011. Nina works on developing efficient search algorithms for decision and optimization problems.  She was named one of "AI's 10 to Watch" young researchers in the field of AI. She also received an Outstanding Paper Award at AAAI 2011 and an outstanding program committee member award at the Australasian Joint Conference on Artificial Intelligence 2012.



Siamak Layeghy (Univ. of Queensland, Australia) Bio: (* to be finalized)

Siamak Layeghy is a PhD student at the School of Information Technology & Electrical Engineering at the University of Queensland, Brisbane, Australia. He received his Bachelor in Telecommunication Engineering and his Master in Electronics Engineering from K.N.Toosi University of Technology, Tehran, Iran.  He has several years' experience in Network Engineering in Iran's National Infrastructure Company. His main research interests include Software Defined Networks, Routing and Traffic Engineering and Network.