IEEE Services 2018 Industry Program

The purpose of the industry program is to invite a number of industry experts in the important aspects of services to present and discuss their insight, best practices, practical issues, observations, and data analytics of real-world systems and applications, as well as the potential impact on the future services technologies and applications.

This year, the following three major aspects will be focused:

Big Data and Cognitive Computing, which are related to the two affiliated conferences/congresses of the Congress: Cognitive Computing and Big Data.
Cloud Computing and IoT, which are related to the three affiliated conferences of the Congress: Cloud Computing, IOT, and Edge Computing.
Micro service architectures and related technologies, which are related to the two affiliated conferences: Services Computing, and Web Services.

To facilitate the Congress participants to gain useful information in the same areas, each of these major aspects will be covered by one session, which is scheduled following the keynote and/or plenary panel sessions in the same areas.

The following are the three sessions of the program:

Session 1: Big Data and Cognitive Computing

Session Chair: Samir Tata, LG Silicon Valley Lab
Date: July 3, 2018

	Session 1 Talk 1 - Evolution of Big Data Messaging - A Look Back and the Path Forward
Summary:	In this presentation we will start in the nineties with the emergence of enterprise messaging (mqSeries, ActiveMQ, TIBCO,MSMQ etc.) and how in 20 years the industry evolved into big-data messaging (Kafka, Kinesis, EventHub etc.). In this talk, we will explore what lead us to big-data messaging and the architectural differences between enterprise messaging and big-data messaging systems. We will discuss some of the hard problems around exactly once processing, pub-sub, data mirroring and the different solutions. The evolution of big-data messaging and how it pushed a revolution in event (stream) processing frameworks will be explained. In closing, we will look into where the industry is headed and the challenges these or new systems will have to overcome to get in the next 10 years. This will include the needs that arise from the mass migration of applications from the batch processing paradigms (Hadoop/Spark) to real time stream processing.
Speaker:	Kartik Paramasivam, Director of Engineering, LinkedIn Karthik has been building messaging platforms since 2000 starting with BizTalk Server: Microsoft's application server with capabilities for rich pub-sub messaging, long running stateful apps, EAI (rich connector support). Later, he spent a few years on Microsoft .NET platform building the web services and workflow platform. After that, he worked on first gen of Microsoft's Cloud Messaging offering: ServiceBus, followed by Eventhub as a big-data messaging offering in Azure. In the last 4 years, he has been working on LinkedIn Streams Infrastructure developing Kafka/Samza and similar technologies.
	Session 1 Talk 2 - Data Science and the Art of Producing Entertainment at Netflix
Summary:	Netflix has released hundreds of Originals and plans to spend $8 billion over the next year on content. Creators of these stories pour their hearts and souls into turning ideas into joy for our viewers. The sublime art of doing this well is hard to describe, but it necessitates a careful orchestration of creative, business and technical decisions. In this talk I will focus on the latter two — business & technical decisions that surround a production and how machine learning, optimization and data analytics are being leveraged to achieve unprecedented logistical scale and operational efficiencies at Netflix Studio.
Speaker:	Ritwik Kumar, Director, Science & Analytics, Netflix Ritwik Kumar is the Director of Studio Production Science & Analytics at Netflix, where he and his team drive descriptive & predictive analytics, machine learning and algorithm development to support original content production (Netflix Studio) and video encoding verticals. Prior to joining Netflix, Ritwik worked at Apple & IBM Research, focusing on building machine learning system that worked with image, video, text and real-time transactional data. He has built and led multiple data science teams at Netflix & Apple. Ritwik holds a Ph.D. in Computer Engineering from the University of Florida and completed his postdoctoral fellowship at Harvard University.

Session 2: AI and Optimization in Service Management

Session Chair: Maja Vukovic, IBM
Date: July 5, 2018

	Session 2 Talk 1 - Monitoring Services in the Internet of Things: An Optimization Approach
Summary:	Devices in Internet of Things (IoT) often offer services that allow tenants to access data of different metrics collected from sensors connected to these devices. Given that such monitoring services are usually invoked within devices that have limited IT resource capacities, it is impossible to collect data of all metrics in the application’s context with a very high frequency. In this talk, we propose a framework that determines which metrics to monitor, monitoring start times, the optimal allocation of metrics to devices, and the optimal monitoring frequency of these metrics, without exceeding different device-specific time-varying resource capacities. Our approach is also adaptive; it gives updated solutions whenever a trigger happens in the system necessitating the need for a change in the previous optimal decisions. We provide an implementation of our approach and present numerical results showing its usage and limitations. At the heart of our approach is an integer programming optimization model that might be hard to solve for large-sized IoT systems. Thus, we present another predictive model that predicts for the user whether our optimization-based approach would be appropriate for her system or not. That is, whether the optimization model is predicted to give optimal solutions within some user-given optimality gaps in a time less than or equal to some user-given maximum allowed time. We also present extension ideas for our solution approach.
Speaker:	Aly Megahed, Research Staff Member, IBM Aly Megahed is a research staff member at IBM’s Almaden Research Center in San Jose, CA. His current research interests span over building analytical tools for complex service engagements, cloud computing, and IoT, and advancing research in analytics, machine learning, and operations research. Dr. Megahed got his Ph.D. in Industrial Engineering from Georgia Tech. He has done multiple analytical research/consultancy projects for 6 companies in the past and has his work published in several academic journals and conferences, in addition to filing multiple patent disclosures and winning multiple IBM internal awards as well as external ones.
	Session 2 Talk 2 - AIOps: Experiences and Challenges
Summary:	Service Management provides a set of processes for providing IT services to customers, such as incident and change management. In this talk, we discuss challenges and opportunities for AI and automation in the service management processes. Specifically, challenges arise from distributed knowledge about the operating environment, coupled with resource constraints and human error and complexity and heterogeneity of the IT environments to name a few. We present a system that employs AI to process and automate service requests, coupled with a chatbot interface. We discuss how we manage and process natural-language based request, coming from ticketing systems, emails and chats and map them to automation offerings. We further present our methodology for in-context parameter extraction and recommendation to help the user refine their request to the point where approvals and automatic executions can be made against a backend automation execution engine. To this end, we provide - intelligent service request dispatch and assistance (against federated service catalogs and automation APIs)- in-context recommendation and validation of user request for the specific automation offering the user is interested in, reducing error and confusion from user - orchestrate and manage approval of change requests across multiple parties in a secure and consistent way (applying blockchain technology). We demonstrate our prototype in action and discuss research agenda in this domain.
Speakers:	Anup Kalia, IBM Research Anup Kalia is a Research Staff Member of the Cognitive Service Management organization at IBM T. J. Watson Research Center, NY. Anup’s research interests include service computing, multiagent systems, cognitive science, and software engineering. In IBM Research, he is exploring different techniques in the areas of text mining, natural language processing, machine learning, deep learning, and transfer learning to solve problems in the area of service analytics and automation. Anup received his MS (2013) and PhD (2016) from North Carolina State University. He was advised by Professor Munindar P. Singh. Anup has received awards such as the best paper award at ICSOC, 2015, US Army research pre-doctoral fellowship, 2015, and several travel grants (AAAI, SIGAI CNC, ICSOC). During his PhD, he interned at HP Labs, Palo Alto, and US Army Research Labs.Jin Xiao, IBM Research Jin Xiao is a Research Staff Member of the Cognitive Service Management organization at IBM T.J. Watson Research Center. Jin’s research expertise is in service analytics and automation, cloud security management, network analytics and software-defined enterprise networks. Jin has helped IBM global technology services business to establish next-generation enterprise network architecture and services for large venues, which also lead to the IBM business practice in sports, media and entertainment where Jin continues to support with thought leadership and innovations in network analytics. Jin is currently driving key IBM projects on cognitive service solutions for IT service management and automation. Jin is a member of IEEE. Jin received his PhD fro University of Waterloo, Canada, for his work on service-driven networks. Jin is active in the research community, especially in network and service management areas.Maja Vukovic, IBM Research Maja Vukovic is a Research Manager of the Cognitive Service Management organization at IBM T.J. Watson Research Center. Maja’s research expertise is in service automation, cloud transformation, crowdsourcing technologies and API ecosystems. Maja has pioneered the area of enterprise crowdsourcing. Maja has received four IBM Outstanding Technical Achievement Awards and three IBM Research awards for her technical leadership and contributions to enterprise crowdsourcing and innovations in IT service management. Maja is an IBM Master Inventor and a Member of IBM Academy of Technology. She has over 80 publications in top international venues. Maja is a Senior Member of IEEE. Maja received her PhD from University of Cambridge, UK, for her work on context aware service composition using AI planning. Maja is active in the research community. In 2016, her team won best paper award at ICSOC. She has organized several workshops at leading international conferences, including a tutorial at ICSOC in 2010. She was also a Web-site Chair for ICSOC in 2010. She is currently the program chair for ICSOC in 2018.

Session 3: Cloud Computing and IoT

Session Chair: Hemant Jain, The University of Tennessee at Chattanooga
Date: July 6, 2018

	Session 3 Talk 1 - Modeling and Simulation of IoT and 5G Applications
Summary:	IoT/M2M and Vehicular communications (V2X) emerged as two killer applications of 5G standards. Using Intel Simulation Tools and Technologies (CoFluent, Simics and Docea), we show how to model and simulate connected cars scenarios and E2E IoT deployments. Using Intel CoFluent helps make early stage architectural analysis on the modem side and helps simulating and understanding HW and SW interactions. These early stage decisions can help design more efficient communication protocols optimize TCO and predict power consumption.
Speaker:	Wael Guibene, Sr. Systems Engineer, Intel Corporation Wael Guibene is a Sr. Wireless Systems Engineer at Intel, based in Santa Clara, CA. Wael is working on novel RF architectures and IoT systems/protocols design. Before joining Intel, he worked for Semtech as a wireless protocol engineer on LoRa/LoRaWAN. Prior to that, he worked at EURECOM on EU projects related to 4G/5G. His research interests include 5G, M2M/IoT, and E2E system architecture and design.
	Session 3 Talk 2 - Microservices: How loose is loosely coupled?
Summary:	Microservice architecture is a popular design pattern for DevOps deployments of cloud native applications. Its single purpose, loosely coupled, bounded context design lends itself to the independent life cycle required to quickly deploy and scale in the cloud. Docker containerization of these services further aids in the zero down-time deployments of these horizontally scalable services. But how do you keep them loosely coupled? How do they communicate without knowing about each other? And how do you keep all of those containers patched from new vulnerabilities that are being discovered every day? This talk discusses the implementation of a Container Vulnerability Remediation Services that itself is designed as a collection of loosely coupled microservices that communicate via publish/subscribe messaging model using Kafka, Could Functions (OpenWhisk), and REST APIs implemented in Python Flask. This design keeps each microservice independent and replaceable, while enabling expandability for new services to participate in business functions without any pre-determined knowledge of the business workflow.
Speaker:	John Rofrano, Research Staff Member, IBM John Rofrano is a Senior Technical Staff Member at IBM T.J. Watson Research Center where he leads several projects from natural language processing of structured data, to cloud native migration technologies leveraging microservices architecture, and was instrumental in the adoption of DevOps at IBM Research. John has spent 34 years at IBM on various software engineering projects, and is also an Adjunct Professor at New York University teaching a masters class on DevOps.