Join RDI2 and Gustavo Portella for "Statistical and Utility-Based Analysis of Cloud Transient Pricing and Availability"
RDI2 welcomes you to attend a talk given by Gustavo Portella, "Statistical and Utility-Based Analysis of Cloud Transient Pricing and Availability." The goal of this talk is to present and discuss strategies to help cloud users to better hire IaaS spot instances, depending on costs constraints and application requirements.
Abstract: Public IaaS cloud computing has evolved with the emergence of dynamic pricing policies and transient resource allocation constraints. In this talk, I will present two analysis of Amazon EC2 Spot pricing model. The first one is a statistical analysis that uses a time-smoothed moving average considering 12-hour periods, aiming to provide price-availability tradeoff to the user. Our experiments with spot price history data, from September to November 2016, show that the user's bid can be set at 30% of the on-demand price, with an availability above of 90%, depending on instance type and purpose. The second one is a utility-based analysis that consists of a predefined strategy, so that user conflicting objectives related to the reduction of costs and the increase of resources availability are considered. Thus, the model provides bid/cost estimation values and expected availability rates of virtualized computing assets. The model also provides recommendations for the best moment to hire the service, i.e. now or a specific weekday.
Date: January 23, 2019 | 2 p.m.-3 p.m.
Location: CoRE Building room 631, Busch Campus - 96 Frelinghuysen Rd, Piscataway Township
RDI2 Associate Director Forough Ghahramani Presents at EdgeCon2019: "Transformative Tools for Advancing Collaborative Research in New Jersey"
Forough Ghahramani, Associate Director for the Rutgers Discovery Informatics Institute, presented “Transformative Tools for Advancing Collaborative Research in New Jersey” at NJEdge's EdgeCon2019 on January 11th.
NJEdge is a member-driven, non-profit technology consortium of academic and research institutions in New Jersey which aims to deliver and sustain a healthy, vibrant, and thriving technology ecosystem that is purpose-built for the communities it serves. The theme of this year's EdgeCon, NJEdge's annual conference, was Digital Transformation: From Strategy to Practice.
Dr. Ghahramani's session focused on digital transformation in New Jersey's research landscape. As high performance networking and computing capabilities have matured in New Jersey, coupled with availability of cutting edge computation technologies and the promise of big data, the landscape for collaborative and innovative research has changed for domain scientists, accelerating scientific discovery and allowing for highly productive collaborations among scientists throughout the region.
This session featured two recent additions to the portfolio of capabilities serving advanced research in New Jersey – Caliburn, the most powerful high performance computing system in the state, and the Virtual Data Collaboratory, an NSF funded initiative that is designed to drive data-intensive, interdisciplinary and collaborative research, and enable data-driven science and engineering discoveries.
RDI2 Distinguished Seminar: Magaret Martonosi, "What is the Role of Architecture and Software Researchers in Making Quantum Computing Practical?"
In the past 3-5 years, Quantum Computing (QC) has reached an interesting and important inflection point. For decades, quantum computing research was comprised of very abstract mathematical algorithms development “up high” that demonstrated some potential for future impact, and physics device demonstrations “down low” that were modest in size but that offered some hope for eventual implementations. However, with prominent QC algorithms like Shor’s factoring algorithm needing roughly a million times more physical quantum bits (qubits) than successful implementations currently provide, there has been a cavernous gap between algorithm and implementation. What is needed now are computer scientists to develop the crucial intermediate tool flows, abstraction layers, and programming languages that will help quantum computing scale through the current so-called NISQ (noisy, intermediate-scale quantum) era. Dr. Martonosi's talk will both (i) give details about our new approaches for optimal and near-optimal spatial-temporal placement of QC algorithms onto real systems, and (ii) more broadly advocate for the role that computer architecture, compiler and programming languages researchers must play in order for QC to reach its full potential.
Margaret Martonosi is the Hugh Trumbull Adams '35 Professor of Computer Science at Princeton University, where she has been on the faculty since 1994. She is also Director of Princeton University's Keller Center for Innovation in Engineering Education. Martonosi's research interests are in computer architecture and mobile computing. Her work has included the development of the Wattch power modeling tool and the Princeton ZebraNet mobile sensor network project for the design and real-world deployment of zebra tracking collars in Kenya. Her current research focuses on hardware-software interface approaches in both classical and quantum computing systems. Martonosi is a Fellow of both IEEE and ACM. Notable awards include the 2018 IEEE Technical Achievement Award, the 2010 Princeton University Graduate Mentoring Award, and the 2013 Anita Borg Institute Technical Leadership Award. Her research has earned four recent Test-of-Time Paper Awards: the 2015 ISCA Long-Term Influential Paper Award, 2017 ACM SIGMOBILE Test-of-Time Award, 2017 ACM SenSys Test-of-Time Paper award, and 2018 (Inaugural) HPCA Test-of-Time Paper award.
Friday, March 1, 2019
10:30 am—12:00 pm
Busch Campus Student Center, Center Hall
604 Bartholomew Road, Piscataway, NJ
Refreshments will be provided
RDI2 Researchers Participate in 2019 Exascale Computing Project (ECP) Annual Meeting
Congratulations to RDI2 researchers Pradeep Subedi and Philip Davis, who were invited to participate in the 2019 Exascale Computing Project (ECP) Annual Meeting, a widely recognized and well-respected gathering of computing technology leaders, industry luminaries, computational pioneers and forward-thinking application, software and hardware experts pushing the boundaries of HPC to shape the nation’s future exascale ecosystem. Additional information is available at Exascale's website and annual meeting site.
About the ECP project : The ECP is a collaborative effort of two U.S. Department of Energy organizations – the Office of Science (DOE-SC) and the National Nuclear Security Administration (NNSA).
ECP is chartered with accelerating delivery of a capable exascale computing ecosystem to provide breakthrough modeling and simulation solutions to address the most critical challenges in scientific discovery, energy assurance, economic competitiveness, and national security. This role goes far beyond the limited scope of a physical computing system. ECP’s work encompasses the development of an entire exascale ecosystem: applications, system software, hardware technologies and architectures, along with critical workforce development.
Manish Parashar Delivers Keynote Speech at "IEEE ICPADS 2018" - Extreme Scales, Big Data, and the Transformation of Science
At December 11-13's IEEE ICPADS 2018, the 24th International Conference on Parallel and Distributed Systems in Sentosa, Singapor, RDI2 Director Manish Parashar delivered a keynote speech, entitled “Extreme Scales, Big Data, and the Transformation of Science.”
Established in 1992, ICPADS has been a major international forum for scientists and engineers to exchange and share new ideas and their latest research results regarding systems that are inherently parallel and/or distributed. The conference provides an international forum for scientists, engineers and users to exchange and share their experiences, new ideas, and latest research results on all aspects of parallel and distributed systems.
Dr. Parashar's speech encompassed how extreme scales and big data have become essential to computational and data-enabled science and engineering in the 21st century, while simultaneously data-related challenges are quickly limiting the potential impact of scientific application workflows enabled by current and emerging extreme scale, high-performance, distributed computing environments. In his talk Dr. Parashar explored some of these challenges and investigate how solutions based on data sharing abstractions, managed data pipelines, data-staging service, and in-situ / in-transit data placement and processing can be used to help address them. To read more about the conference and Dr. Parashar's keynote speech, visit ICPADS's website.
Advanced cyberinfrastructure, expertise, and collaborations at RDI2 revolutionizing ocean observation and transforming ocean research
Despite covering 71 percent of the Earth’s surface, we still don’t know as much as we could about the world’s oceans. In fact, we currently have better maps of the surface of Mars than we do of the ocean floor. An enormous amount of ocean data still needs to be collected, but it will require a collaborative effort. That's why scientists are turning to large-scale operations like the Ocean Observatories Initiative (OOI). The Ocean Observatory Initiative funded by the National Science Foundation (NSF), utilizes data from a collection of sensors and platforms all over the globe that measure physical, chemical, geological, and biological properties of the ocean.
The OOI collects continuous observations of the seafloor, ocean, and atmosphere in the Atlantic and Pacific. Collected data is freely disseminated to scientists addressing research on coastal ocean dynamics, ecosystem health, plate-scale seismicity, seafloor volcanism, and more.
These systems are managed by Rutgers, the State University of New Jersey, the University of Washington, and Oregon State University. The coalition is led by the Woods Hole Oceanographic Institution.
Collecting so much data demands a robust cyberinfrastructure (CI). To learn more about how Rutgers made this happen, read the full article at Science Node
Tracking undersea eruptions. This Bottom Pressure and Tilt Meter installed at the summit of the Axial Seamount measures the rise and fall of the seafloor due to melt migration in the subsurface. Courtesy NSF-OOI/UW/CSSF. (CC BY-NC-ND)
Call for RDI2 Graduate Fellowship Applications: Rutgers Discovery Informatics Institute Graduate Fellowship for Excellence in Computational and Data Science
The Rutgers Discovery Informatics Institute (RDI2) is pleased to announce its Fellowships for Excellence in Computational and Data Science to support Graduate Students in all areas of Computational and Data-enabled Science and Engineering (CDSE), with a specific focus on research in Big Data and Extreme Scale computing for academic year 2019-2020. RDI2 will support up to three second and third year Rutgers PhD students working on multi-disciplinary collaborative computational/data-enabled research projects that address more than one element of CDSE research.
Each fellowship appointment will be for one year and comes with $30,000 towards partial Graduate Assistant support for the academic year fellowship. The fellowship has potential for renewal, and includes allocations on RDI2’s research infrastructure (including Caliburn). Prospective candidates are expected to pursue novel and creative research avenues within existing research programs and faculty expertise in collaboration with an RDI2 faculty mentor. Faculty can submit applications for consideration on an ongoing basis.
Successful candidates are expected to start/continue their research as part of the RDI2 research community and to participate in the RDI2 research and other activities. During the academic year, RDi2 fellows are required to submit two reports on progress of their research, mid-year report is due January 31 and end of year report is due May 1. In addition to the progress status, the report should include details of the science impact, presentations and publications resulting from the research. Fellows are required to acknowledge RDI2 support in presentations and publications. Please note, RDI2 will allocate the fund to the student’s respective department and the administration of the fund will be managed by the receiving department.
Applications should include: a 1-2 page project description written by Rutgers faculty, curriculum vitae of proposed student/fellow, and a statement on long term relevance of research and possible future funding agencies. Existing RDI2 Fellow applicants must also include a copy of the mid-year report with the application.
Applications should be emailed by March 1, 2019 to:
Forough Ghahramani (Associate Director, Rutgers Discovery Informatics Institute (RDI2))
Email: fellowship@rdi2. rutgers.edu
(Please include "Grad Student” in the subject line).
We look forward to hearing from you and appreciate your support for computation and data science enabled solutions.
RDI2 Hosts New Jersey Big Data Alliance (NJBDA) Research Collaboration Workshop
On Friday December 7, the New Jersey Big Data Alliance (NJBDA) Research Collaboration Workshop hosted by The Rutgers Discovery informatics institute brought together over 50 researchers representing 18 institutions from government, academia, and industry. According Forough Ghahramani, Associate Director of RDI2 and one of the NJBDA Research Committee organizers of the workshop, in addition to establishing contacts with the New Jersey research community, the objectives of the workshop included:
- An information exchange about Research Collaboration resources and tools and Big Data research opportunities
- Learning about unique data sets available to New Jersey researchers, to identify multidisciplinary and multi-institution collaboration opportunities
- Discussing the next steps in shaping the Big Data Research Collaboration Community in New Jersey.
Peggy Brennan-Tonetta, Associate Vice President of Economic Development, Rutgers University and the President of the New Jersey Big Data Alliance provided welcome remarks and information about the NJBDA. The NJBDA Research Committee organized and facilitated the workshop: J.D Jayaraman (NJCU), David Belanger (Stevens Institute of Technology), Edward Chapel (NJEdge), and Forough Ghahramani (Rutgers University).
Diving Into Big Data with Camden County Technical School
On December 6, RDI2 welcomed high school students from Camden County Technical School to tour the state-of-the-art facilities at RDI2 and learn about the fundamentals of data science and top research initiatives at the institute as part of a Diving Into Big Data and Large Scale Computing workshop. Students had a great time learning about big data, and experimented with live oceanographic data and developed a web-based dashboard. The event was well received by students. The following statement was included in her follow-up email from Linda Beluch, the teacher who accompanied the students to Rutgers. “Thank you so much for welcoming us on Thursday. We had a fun and informative visit. The students mentioned they really enjoyed the data center visit and were surprised at all the work you do with big data. I think they are still in the mindset that programmers just make video games, so this was great exposure for them to see other opportunities that are out there. I really hope we can keep in touch and collaborate in the future.” – Linda Beluch, Camden County Technical School educator
New Jersey Big Data Alliance Research Collaboration Workshop
The Research Committee of the New Jersey Big Data Alliance is convening its first ever Research Collaboration Workshop which is open to you and a select group of research professionals from the institutions that comprise the NJBDA who are interested in shaping the Big Data research collaboration landscape.
The workshop will feature a research stream that pertains to the dynamics of network operations and flow for a carrier class wide area network. Participants will have the opportunity to formulate research problems and develop collaborative strategies and techniques using big data, cyberinfrastructure and technologies to investigate them.
Seating is limited, so please register today at https://bit.ly/2JHb41y
When: Friday December 7, 2018 10:00 am–1:00 pm.
Where: Rutgers University, Busch Campus Student Center - Multipurpose Room B Address: 604 Bartholomew Road, Piscataway, NJ
RDI2 Presents at SC18, International Conference for High Performance Computing, Networking, Storage, and Analysis
RDI2 Presents at SC18, International Conference for High Performance Computing, Networking, Storage, and Analysis
RDI2 staff and researchers presented at the SC18, International Conference for High Performance Computing, Networking, Storage, and Analysis in Dallas, Texas this past week; the conference ran from November 11th to November 15th.
Philip Davis presented the paper Stacker: An Autonomic Data Movement Engine for Extreme-Scale Data Staging-Based In Situ Workflows, authored by Davis, Pradeep Subedi, Shaohua Duan, Scott Klasky, Hemanth Kolla, and Manish Parashar; the paper can be downloaded from the SC18 website. Davis also authored and led two workshops, Scaling Deep Learning for Cancer with Advanced Workflow Storage Integration and Leveraging Scalable Event Distribution to Enable Data-Driven In Situ Scientific Workflows, the second of which was also authored and led by graduate student Zhe Wang, as well as postdoc associate Pradeep Subedi. Subedi also presented the paper Stacker: An Autonomic Data Movement Engine for Extreme-Scale Data Staging-Based In Situ Workflows during a File Systems session.
Manish Parashar presented "The Future of NSF Supported Advanced Cyberinfrastructure" on behalf of the NSF at a Birds of a Feather session, in which attendees are invited to openly discuss current topics of interest to the HPC community. This presentation updated attendees on new NSF cyberinfrastructure strategies and activities, and provided a forum for discussing the latest funding opportunities. Read more about the presentation here.
To read more about the conference, visit its website here.