HomeNews and EventsColloquia → Colloquium

Candidate Seminar

When: 10:00AM - 11:30AM , February 27, 2018

Where: CC-3-3540

Speaker: Dr. Hanku Lee

Abstract

Dr. Hanku Lee’s talk aims to develop new cloud computing systems and technologies for data-intensive applications related to data parallel languages, compilers, high-performance computing, scientific computing, data science, big data analytics and social media. These new systems present not only challenging research problems that can enhance the future cloud computing stack, but also an opportunity to review traditional cloud computing questions (e.g. where resources are located, how services, platforms, and infrastructures are managed).

HPJava (http://www.hpjava.org) is a Java implementation of the HPspmd data parallel programming language model. It is a preprocessor-based programming language and extends the Java language with some additional syntax and with some pre-defined classes for handling distributed arrays language primitives. The distributed array model is adopted from the HPF array model. But, the programming model is quite different from HPF. It is one of explicitly cooperating processes. All processes carry out the same program, but the components of data structures are divided across processes. Each process operates on locally held segment of an entire distributed array. Also, HPJava supports the true multidimensional array, which is a modest extension to the standard Java language. The new arrays allow regular section subscripting, similar to Fortran 90 arrays. The syntax for the multidimensional array is a subset of the syntax for distributed arrays. HPJava also introduces some new control constructs such as overall, at, and on statements. Additionally, it is necessary to provide bindings from the extended language to various communication and parallel computation libraries.

Moreover, this talk presents his research accomplishments that have enabled the roots of traditional cloud computing to evolve towards one that serves as a next-generation social media cloud platform with novel systems, related to cloud stream processing, SaaS aggregation, fashion big data analytics, and data science. These research accomplishments are a major step forward compared to traditional cloud computing approaches. Especially, his SaaS aggregation services have been successfully connected with similar paradigms from the Internet of Things and smart home systems.

This talk also focuses on his new research agenda to enhance the future cloud computing stack and collaborate with scientific communities in data science and big data analytics. He is developing a next-generation cloud platform geared towards optimizing cloud computing systems by performing data processing at the “edge” (edge computing) of the network and by replacing inefficient procedural cloud systems with event-driven “FaaS” (Function as a Service) systems. Overall, merging edge computing and FaaS amplifies the synergy and adds a new dimension to cloud computing because of the explosion of IoT and data-intensive applications. There could be major challenges to build a next-generation event-driven computing model around clouds and HPC with batch and streaming data. This model, however, could analyze large volumes of data generated from edge events and provide powerful and packaged event-driven functions (FaaS) spanning from edge to multiple clouds. For example, a signal from a sensor of an unmanned ground vehicle is an edge-event. Calling a HPC linear algebra function could be event-driven and may replace traditional libraries by FaaS.

Speaker Bio

Dr. Hanku Lee is currently a full professor with tenure in the Department of Software at Konkuk University, Seoul, Korea, and a senior visiting professor in the Department of Computer Science and Engineering at University of Colorado Denver, Denver, CO. His primary research focuses on the development of new cloud computing systems and technologies for data-intensive applications related to data parallel languages, compilers, high-performance computing, scientific computing, data science, big data analytics and social media. In 2010, he built the Center for Social Media Cloud Computing (SMCC) at Konkuk University with an annual budget of $1M-$1.5M from the Korean Ministry of Science and ICT. The Korean government pledged its full support for his center and named the center as one of South Korea’s IT-Convergence Research Centers, called ITRC, for innovative research in areas of social media, cloud computing, high-performance computing, parallel languages, compilers, data science, and big data analytics. His center has been closely collaborating with institutions, the Korean government agencies, and industrial partners in a cross-disciplinary manner. His center has obtained 21 research grants, 4 US Patents, 24 Korean Patents, 14 Technology Transfers, 109 publications, seven “best research” awards from the Korean government and Konkuk University.


This page was last modified on March 08, 2018
© 2018 University of Massachusetts Boston

Template by OS Templates