Sponsors
The IAP Carnegie Mellon Workshop on the Future of Cloud Computing was conducted on Friday, April 8, 2016 at the Gates Center for Computer Science, Room 6115 Gates Center, Carnegie Mellon University in Pittsburgh, PA. The workshop was organized by Prof. Onur Mutlu and Dr. Saugata Ghose.
Agenda - Videos of Presentations
8:30-8:50AM Prof. Onur Mutlu – Carnegie Mellon – “Rethinking Memory System Design for Data-Intensive Computing”
8:50-9:15AM Robert Broberg – Cisco – “Trust and Integrity Assurance from IoT to the Cloud Tenant”
9:15-09:40AM Prof. Jimmy Zhu – Carnegie Mellon – “All Spin Logic Devices and Circuits for Future Electronics”
09:40-10:10AM Dr. Pankaj Mehra – SanDisk – “Contemplating a new memory hierarchy for the Next Generation Data Center”
10:10-10:30AM Dr. Khalid Elgazzar – Carnegie Mellon – "Leveraging Edge Computing to Support Internet of Things Scenarios"
10:30-11:00AM Lightning Round of Student Posters
11:00-1:00PM Lunch, Cloud Poster Viewing, and Career Fair
1:00-1:30PM Prof. Yuvraj Agarwal – Carnegie Mellon – “Towards Building a Safe, Secure Internet-of-Things Infrastructure”
1:30-1:50PM Dr. Michael Gaidis – Samsung – “Novel Nonvolatile Memories in the Memory/Storage Hierarchy"
1:50-2:10PM Prof. Vyas Sekar – Carnegie Mellon – “Software-defined network security for next-generation networks”
2:10-2:30PM Dr. Sam Fineberg – Hewlett Packard Enterprise – “Software Defined Storage"
2:30-3:00PM Break - Refreshments and Poster Viewing
3:00-3:30PM Suresh Ravindran – Cavium – “Xpliant - Programmable Data Plane Switch for Software Defined Data Centers”
3:30-4:00PM Prof. Priya Narasimhan – Carnegie Mellon – "The Next-Generation Stadium Experience"
4:00-4:30PM Prof. Prof. Swarun Kumar – Carnegie Mellon – "Wireless Protocols and Security for Smart Cities”
4:30-5:00PM Dr. Hui Huang – Google – "PerfKit Benchmarker: Performance Benchmarking for Cloud"
5:00-6:00PM Reception - Refreshments and Poster Awards in 6th Floor Commons Area
Abstracts and Bios
Prof. Onur Mutlu – Carnegie Mellon – "Rethinking Memory System Design for Data-Intensive Computing"
Abstract: The memory system is a fundamental performance and energy bottleneck in almost all comput- ing systems. Recent system design, application, and technology trends that require more capacity, band- width, efficiency, and predictability out of the memory system make it an even more important system bottleneck. At the same time, DRAM and flash technologies are experiencing difficult technology scaling challenges that make the maintenance and enhancement of their capacity, energy-efficiency, and reliability significantly more costly with conventional techniques. In this talk, we examine some promising research and design directions to overcome challenges posed by memory scaling.
Bio: Onur Mutlu is the Dr. William D. and Nancy W. Strecker Early Career Professor at Carnegie Mellon University. His research interests are in computer architecture, systems, and bioinformatics, especially in the interactions between languages, operating systems, compilers, and microarchitecture. He enjoys teaching and researching important and relevant problems in computer architecture, including problems related to the design of memory systems, multi-core architectures, and scalable and efficient systems. He obtained his PhD and MS in ECE from the University of Texas at Austin (2006) and BS degrees in Computer Engineering and Psychology from the University of Michigan, Ann Arbor. Prior to Carnegie Mellon, he worked at Microsoft Research (2006-2009), Intel Corporation, and Advanced Micro Devices. He was a recent recipient of the IEEE Computer Society Young Computer Architect Award, CMU College of Engineering George Tallman Ladd Research Award, Intel Early Career Faculty Honor Program Award, Microsoft Gold Star Award, IBM and HP Faculty Partnership Awards, best paper awards at ASPLOS, VTS and ICCD, and a number of "computer architecture top pick" paper selections by the IEEE Micro magazine.
Dr. Pankaj Mehra – SanDisk – “Contemplating a new memory hierarchy: persistent memory technologies for the next generation data center”
Abstract: This talk takes a whole system view of programming and persistence in disaggregated systems, ideas I first espoused in a 2001 keynote. Since then, we have seen multiple solid state memory technologies transform the performance of database, filesystem, and virtualization workloads. As the next wave of technologies that have been in development for decades begin to roll out, many of the systems and software insights developed earlier still apply. The roles and motivations around existing technologies must now change. We survey traditional approaches to I/O optimization and much work that was done around the introduction of flash memory as fast disk. Focusing on some pioneering work done first in my persistent memory projects at HP then at Fusion-io, I will describe how this work which was the first to view solid state flash memory as memory rather than disk, now needs to evolve in order to address the rise of storage class memories. The talk will introduce new memory technologies from physics, through system attach points, to what threads, processes and CPU pipelines are likely to see. It will examine the state of the art up close, and end with a call to action for the CMU community on accelerating progress toward compiler optimization of I/O.
Bio: Pankaj Mehra, a Top 50 CTO (ExecRank'14 #39) at Fusion-IO, now VP and Senior Fellow in CTO Office and SanDisk Technology Council chair. Founder (Whodini, IntelliFabric). HP Distinguished Technologist and founding Chief Scientist HP Labs Russia ('04-'10). 10+ years creating Persistent Memory first ('02-'06) at the confluence of InfiniBand (1.0 co-author) and NonStop Advanced Architecture (co-inventor), and again ('12-'15). Built Context Engines ('08-'13). 3 books + 48 papers (many on Machine Learning since '89); 25+ patent filings, including Virtual Switch, Persistent Memory, and Compression Enhancing Routing Algorithms. Ph.D. Illinois-CS, B.Tech. IITD-CS.
Prof. Vyas Sekar - Carnegie Mellon – "Software-defined network security for next-generation networks"
Abstract: The state of network security today is quite abysmal. Security breaches and downtime of critical infrastructures continue to be the norm rather than the exception, despite the dramatic rise in spending on network security. Attackers today can easily leverage a distributed and programmable infrastructure of compromised machines (or botnets) to launch large-scale and sophisticated attack campaigns. In contrast, the defenders of our critical infrastructures are fundamentally crippled as they rely on fixed capacity, inflexible, and expensive hardware appliances deployed at designated "chokepoints". These primitive defense capabilities force defenders into adopting weak and static security postures configured for simple and known attacks, or otherwise risk user revolt, as they face unpleasant tradeoffs between false positives and false negatives. Unfortunately, attacks can easily evade these defenses; e.g., piggybacking on popular services (e.g., drive-by-downloads) and by overloading the appliances. Continuing along this trajectory means that attackers will always hold the upper hand as defenders are stifled by the inflexible and impotent tools in their arsenal.
The goal of our research is to change the dynamics of this attack-defense equation. Instead of taking a conventional approach of developing attack-specific defenses, our work focuses on empowering defenders with the right tools and abstractions to tackle the constantly evolving attack landscape. To this end, we envision a new software-defined approach to network security, where we can rapidly develop and deploy novel in-depth defenses and dynamically customize the network's security posture to the current operating context.
In this talk, I will give an overview of our recent work in this space.
Bio: Vyas Sekar is an Assistant Professor in the ECE Department at CMU, where he is part of CyLab. His research interests lie at the intersection of networking, security, and systems. He received his Ph.D. from the Computer Science Department at Carnegie Mellon University in 2010. He earned his bachelor's degree from the Indian Institute of Technology Madras, where he was awarded the President of India Gold Medal. His work has been recognized with best paper awards at ACM SIGCOMM, ACM CoNext, and ACM Multimedia.
Suresh Ravindran – Cavium – “Xpliant - Programmable Data Plane Switch Architecture for Software Defined Data Centers”
Abstract: The programmable switching hardware data plane is a critical component of Software Defined Networking (SDN) for true innovation and real life network optimization. SDN is not really about remote feature configuration of a fixed pipeline; it truly achieves its purpose only when software can define a new behavior in the hardware data plane. True innovation is possible only when smart minds can develop their own ideas and define it on hardware to function at high throughput. Often an innovative new networking protocol invention fails because its definition makes it complex to implement in hardware. A programmable network pipeline needs to be part of these tools for successful innovation. Different data centers can customize hardware to optimize for their feature services using the same ASIC at different places in their network.
Bio: Suresh Ravindran is the head of the software engineering division and application engineering of the switch platform group at Cavium Networks. A networking industry veteran, he built enterprise and datacenter networking switches and routers for many years, and worked for Brocade, Foundry Networks, Marconi Communications, FORE Systems and Wipro in the past. His engineering degree is in Computer Science and Engineering from TKMCE, University of Kerala, India in 1996.
Prof. Swarun Kumar – Carnegie Mellon – "Cloud-Assisted Networks of Wireless Sensors”
Abstract: Recent years have seen much interest in networking wireless sensors to the cloud. In this talk, I will describe the challenges and opportunities of doing so in the context of driverless cars and multi-robot networks. I will present a cloud-assisted system for autonomous driving. Our solution enables the cloud to have access to sensor data from autonomous vehicles as well as the roadside infrastructure. The cloud assists autonomous vehicles that use this system to avoid obstacles such as pedestrians and other vehicles that may not be directly detected by sensors on the vehicle. Further, it enables vehicles to plan efficient paths that account for unexpected events such as road-work or accidents. We evaluate our system on an outdoor testbed including an autonomous golf car and six iRobot Create robots. Results show a reduction in the average time vehicles need to detect obstacles such as pedestrians by 4.6x compared to today's systems that do not access the cloud.
Bio: Swarun Kumar is an assistant professor at CMU where he works on wireless networks and systems. He designs and builds novel systems that leverage a deep understanding of the wireless physical layer to enable faster wireless networks and new services. His work has been featured as research highlights in the Communications of the ACM (CACM) and the International Journal of Robotics Research (IJRR). Swarun is a recipient of the George Sprowls Award for best Ph.D thesis in Computer Science at MIT and the President of India gold medal at IIT Madras.
Prof. Yuvraj Agarwal – Carnegie Mellon – “Towards Building a Safe, Secure and Easy-to-Use Internet-of-Things Infrastructure”
Abstract: A world enabled by the Internet of Things (IoT) promises to lead to truly connected environments, where people and things collaborate to improve the overall quality of life. The IoT will give us actionable information at our fingertips, without us having to ask for it or even recognizing that it might be needed. However, in order to realize this vision, and build out a successful IoT infrastructure, there are a number of challenges that must be addressed. These challenges include safety, security and privacy in a world of pervasive IoT, scalability in terms of data storage and processing architectures, new programming abstractions to allow ease of programming, etc. At Carnegie Mellon, we are currently building an end-to-end IoT stack, called GIoTTO, to address some of these challenges. In this talk I will give an overview of our ideas and progress over the past year on this IoT Expedition (www.iotexpedition.org). If time permits I will specifically talk about some of the security and privacy challenges that IoT presents.
Bio: Yuvraj Agarwal is an Assistant Professor of Computer Science in the School of Computer Science at Carnegie Mellon University. His research interests are at the intersection of Systems and Networking and Embedded Systems, and he is particularly excited about research problems that benefit from using hardware insights to build more scalable and energy efficient systems. In recent years, his work has focused on the domain of Green Computing, Mobile Computing, Energy Efficient Buildings and the Internet of Things with a cross cutting interest on security and privacy issues.
Prof. Jimmy Zhu – Carnegie Mellon – “All Spin Logic Devices and Circuits for Future Electronics”
Abstract: Today, electronic devices are charge-based and operate by moving, accumulating, or draining charges for data processing. The flow of electric charges results in Joule heating by Ohmic loss, causing exhaustion of a substantial percentage of the power consumed while limiting the device performance. In this talk, an alternative scheme is proposed with utilization of electron spins by generating charge-free spin current to eliminate/reduce Joule-heating. A new switching device and circuit technology, mLogic, based on current signals and pulse voltage supplies for power and clocking enables implementation of complete large-scale systems without semiconductors, a potential candidate for future low power electronics.
Bio: Jian-Gang (Jimmy) Zhu is the ABB Professor of electrical and computer engineering at Carnegie Mellon University and the director of the Data Storage Systems Center. He received his Ph.D. in Physics from University of California at San Diego in 1989. Prior to coming to Carnegie Mellon in 1997, he had been a faculty in the Department of Electrical Engineering at the University of Minnesota since 1990. Some of the awards that he received include the McKnight Land Grant Professorship from the University of Minnesota in 1992, the NSF Presidential Young Investigator Award in 1993, the R&D Magazine Top 100 Invention Award in 1996, and the IEEE Magnetic Society Achievement Award in 2011. He was IEEE Magnetic Society Distinguished Lecturer in 2004. He has published over 290 refereed journal papers along with seven book chapters and has given over 90 invited papers at major international conferences. He holds 20 U.S. patents. He is a Fellow of IEEE.
Robert Broberg – Cisco – “Trust and Integrity Assurance from IoT to the Cloud Tenant”
Abstract: A look at Trustable systems and Cisco's Advanced Security Research efforts to future proof the Internet against attacks.
Bio: Robert Broberg is a Distinguished Engineer at Cisco Systems. Robert started his career in networking at Ungermann-Bass in 1984 and since then has been leading innovations in IP networking spanning everything from device drivers, applications to large system design. His career has spanned many roles. Some highlights include, 2 years in Tokyo leading network the industry transition from XNS to TCP, 3 years at Bell-Labs in Physical Systems research as IP transitioned from a TDM based phone client to owning Optical transport and 3 years at AYR netkworking, a startup using Linux as an OS for high performance ASIC assisted routers.
Robert is currently part of the newly formed Advanced Security Research Group at Cisco. This group has a vision to address security in the 5,10 to 15 year horizon by partnering worldwide with Academics, Industry consortiums and leading research institutions.
Dr. Michael Gaidis – Samsung – “Novel Nonvolatile Memories in the Memory/Storage Hierarchy for Cloud and IoT Applications”
Abstract: Advances in flash memory bandwidth, latency, and capacity have led to system architectures with potential for much higher performance than predecessors using hard disk drives. A resulting conundrum is the need for large-scale software redesign to realize the full advantages offered by the new high-performance storage. This "ecosystem" development for flash has been underway for several years already, and is far from complete. The situation becomes even more complex when emerging nonvolatile memories are considered. This presentation will cover the existing memory/storage hierarchy and the tradeoffs that will dictate what role different types of emerging nonvolatile memories will play in future system architectures. We will address questions such as: where nonvolatile memories will have the greatest impact; how they might best be implemented as tiering between, or as replacement for DRAM and flash; and where one needs to be careful of excessive hype.
Bio: Michael Gaidis is a Senior Manager with Samsung Device Solutions in San Jose, California, focused on university relationships and memory strategy. Michael works closely with the Memory Solutions Lab (flash, DRAM, memory systems), the Device Lab (modeling of advanced logic and memory devices), and the New Memory Technology Lab (MRAM). He has previously worked with device-level research including resistive RAM, MRAM, and terahertz components. This included work at HGST's San Jose Research Lab, IBM's TJ Watson Research Center, and NASA's Jet Propulsion Lab. His education included a postdoctoral fellowship at Caltech, PhD at Yale, and MS/BS degrees at MIT. He can be reached at [email protected]
Dr. Sam Fineberg – Hewlett Packard Enterprise – “Software Defined Storage – from industry standard appliances to hyper-converged systems”
Abstract: Software defined storage is an emerging concept that enables storage to be provisioned dynamically, as software running on a pool of industry standard severs or even virtual machines. This talk will present an overview of this market, as well as the major storage types and how they are driving the storage industry. Topics will also include the major deployment models for software defined storage such as hardware appliances, cloud, VSA, and hyper-converged. The talk will also describe HPE’s approach to software defined storage and key technologies that address the market.
Bio: Sam Fineberg is a Distinguished Technologist in the Hewlett Packard Enterprise (HPE) Storage Chief Technologist Office. Dr. Fineberg has over 25 years of experience in areas including storage, cloud computing, archiving, non-volatile memory, high performance computing, and data preservation. Sam leads cloud and big data technical strategy for HPE’s multi-billion dollar storage division, driving participation in the OpenStack cloud storage community and solutions for Hadoop/Spark. Prior to his role in HPE Storage, Dr. Fineberg was chief technologist of HPE’s Information Management Software business, where he architected an archival storage system, and led the acquisition and integration of Tower Software. He also developed a record breaking Terabyte sort application and early network attached non-volatile memory technology as part of HPE’s NonStop Labs. Prior to HPE, Dr. Fineberg worked NASA Ames Research Center where he led key research in high performance computing, message passing, and parallel I/O. He earned a PhD in ECE from U of Iowa, and MSEE and BSEE from Purdue. He can be reached at [email protected]
Dr. Hui Huang – Google – "PerfKit Benchmarker: Performance Benchmarking for Cloud"
Abstract: With the increasing popularity of cloud computing, we've seen a growing demand for living benchmarking of cloud infrastructures. In this talk we will introduce PerfKit Benchmarker, which is an industry effort led by Google to establish a Cloud benchmarking suite. The goal is to represent how cloud developers are building applications and evaluating varied public/private clouds. We will also cover the design methodology, architecture of PerfKit Benchmarker, and a couple of performance challenges we've seen with cloud computing.
Bio: Dr. Hui Huang is a performance engineer in Google Cloud/Platform team. She is responsible for public cloud performance evaluation and technical infrastructure improvement for Google Compute Engine (GCE). Hui is also the Google representative/committee member of Standard Performance Evaluation Corporation (SPEC) Cloud, and an active contributor to PerfKit Benchmarker. Before joining Google in 2015, she co-founded Falcon Computing Solutions, which is a prestigious startup targeting on providing heterogeneous accelerating services in data centers. Dr. Huang holds a PhD in Computer Science from UCLA, and BS from Peking University, China. She can be reached at [email protected].
Dr. Khalid Elgazzar – Carnegie Mellon – “Leveraging Edge Computing to Support Internet of Things Scenarios”
Abstract: The IoT vision is to make the world better connected and enable smarter interactions between devices, people, and everyday artifacts. However, the inherent resource constraints of IoT artifacts limit the quality and type of functionality that can be offered, inhibiting IoT scenarios from reaching full potential. Computation offloading offers resource-constrained IoT devices the opportunity to transfer compute-intensive tasks to more resourceful computing infrastructures on the network edge (aka Edge Computing). At Carnegie Mellon, we are currently building GIoTTO, an end-to-end IoT open stack . In this talk, I will present our efforts to leverage cloud computing and cloudlets to support IoT devices to reduce the time required to deliver the end-to-end IoT scenarios, including best practices for interoperability and composability between cloud and edge architectures.
Bio: Dr. Khalid Elgazzar is a postdoctoral research fellow at Carnegie Mellon School of Computer Science and an adjunct Assistant Professor in the School of Computing at Queen's University, Canada. Dr. Elgazzar received his PhD from Queen's University in 2013. He also received the 2014 Queen's School of Computing Distinguished Research Award. His research interests span the areas of distributed systems, mobile cloud and ubiquitous computing, context-aware cyber-physical systems, and elastic networking paradigms. Dr. Elgazzar has received several recognitions and best paper awards at top international venues. He is currently a member of the Google IoT Expedition team, leading the creation of an open stack for IoT.
Agenda - Videos of Presentations
8:30-8:50AM Prof. Onur Mutlu – Carnegie Mellon – “Rethinking Memory System Design for Data-Intensive Computing”
8:50-9:15AM Robert Broberg – Cisco – “Trust and Integrity Assurance from IoT to the Cloud Tenant”
9:15-09:40AM Prof. Jimmy Zhu – Carnegie Mellon – “All Spin Logic Devices and Circuits for Future Electronics”
09:40-10:10AM Dr. Pankaj Mehra – SanDisk – “Contemplating a new memory hierarchy for the Next Generation Data Center”
10:10-10:30AM Dr. Khalid Elgazzar – Carnegie Mellon – "Leveraging Edge Computing to Support Internet of Things Scenarios"
10:30-11:00AM Lightning Round of Student Posters
11:00-1:00PM Lunch, Cloud Poster Viewing, and Career Fair
1:00-1:30PM Prof. Yuvraj Agarwal – Carnegie Mellon – “Towards Building a Safe, Secure Internet-of-Things Infrastructure”
1:30-1:50PM Dr. Michael Gaidis – Samsung – “Novel Nonvolatile Memories in the Memory/Storage Hierarchy"
1:50-2:10PM Prof. Vyas Sekar – Carnegie Mellon – “Software-defined network security for next-generation networks”
2:10-2:30PM Dr. Sam Fineberg – Hewlett Packard Enterprise – “Software Defined Storage"
2:30-3:00PM Break - Refreshments and Poster Viewing
3:00-3:30PM Suresh Ravindran – Cavium – “Xpliant - Programmable Data Plane Switch for Software Defined Data Centers”
3:30-4:00PM Prof. Priya Narasimhan – Carnegie Mellon – "The Next-Generation Stadium Experience"
4:00-4:30PM Prof. Prof. Swarun Kumar – Carnegie Mellon – "Wireless Protocols and Security for Smart Cities”
4:30-5:00PM Dr. Hui Huang – Google – "PerfKit Benchmarker: Performance Benchmarking for Cloud"
5:00-6:00PM Reception - Refreshments and Poster Awards in 6th Floor Commons Area
Abstracts and Bios
Prof. Onur Mutlu – Carnegie Mellon – "Rethinking Memory System Design for Data-Intensive Computing"
Abstract: The memory system is a fundamental performance and energy bottleneck in almost all comput- ing systems. Recent system design, application, and technology trends that require more capacity, band- width, efficiency, and predictability out of the memory system make it an even more important system bottleneck. At the same time, DRAM and flash technologies are experiencing difficult technology scaling challenges that make the maintenance and enhancement of their capacity, energy-efficiency, and reliability significantly more costly with conventional techniques. In this talk, we examine some promising research and design directions to overcome challenges posed by memory scaling.
Bio: Onur Mutlu is the Dr. William D. and Nancy W. Strecker Early Career Professor at Carnegie Mellon University. His research interests are in computer architecture, systems, and bioinformatics, especially in the interactions between languages, operating systems, compilers, and microarchitecture. He enjoys teaching and researching important and relevant problems in computer architecture, including problems related to the design of memory systems, multi-core architectures, and scalable and efficient systems. He obtained his PhD and MS in ECE from the University of Texas at Austin (2006) and BS degrees in Computer Engineering and Psychology from the University of Michigan, Ann Arbor. Prior to Carnegie Mellon, he worked at Microsoft Research (2006-2009), Intel Corporation, and Advanced Micro Devices. He was a recent recipient of the IEEE Computer Society Young Computer Architect Award, CMU College of Engineering George Tallman Ladd Research Award, Intel Early Career Faculty Honor Program Award, Microsoft Gold Star Award, IBM and HP Faculty Partnership Awards, best paper awards at ASPLOS, VTS and ICCD, and a number of "computer architecture top pick" paper selections by the IEEE Micro magazine.
Dr. Pankaj Mehra – SanDisk – “Contemplating a new memory hierarchy: persistent memory technologies for the next generation data center”
Abstract: This talk takes a whole system view of programming and persistence in disaggregated systems, ideas I first espoused in a 2001 keynote. Since then, we have seen multiple solid state memory technologies transform the performance of database, filesystem, and virtualization workloads. As the next wave of technologies that have been in development for decades begin to roll out, many of the systems and software insights developed earlier still apply. The roles and motivations around existing technologies must now change. We survey traditional approaches to I/O optimization and much work that was done around the introduction of flash memory as fast disk. Focusing on some pioneering work done first in my persistent memory projects at HP then at Fusion-io, I will describe how this work which was the first to view solid state flash memory as memory rather than disk, now needs to evolve in order to address the rise of storage class memories. The talk will introduce new memory technologies from physics, through system attach points, to what threads, processes and CPU pipelines are likely to see. It will examine the state of the art up close, and end with a call to action for the CMU community on accelerating progress toward compiler optimization of I/O.
Bio: Pankaj Mehra, a Top 50 CTO (ExecRank'14 #39) at Fusion-IO, now VP and Senior Fellow in CTO Office and SanDisk Technology Council chair. Founder (Whodini, IntelliFabric). HP Distinguished Technologist and founding Chief Scientist HP Labs Russia ('04-'10). 10+ years creating Persistent Memory first ('02-'06) at the confluence of InfiniBand (1.0 co-author) and NonStop Advanced Architecture (co-inventor), and again ('12-'15). Built Context Engines ('08-'13). 3 books + 48 papers (many on Machine Learning since '89); 25+ patent filings, including Virtual Switch, Persistent Memory, and Compression Enhancing Routing Algorithms. Ph.D. Illinois-CS, B.Tech. IITD-CS.
Prof. Vyas Sekar - Carnegie Mellon – "Software-defined network security for next-generation networks"
Abstract: The state of network security today is quite abysmal. Security breaches and downtime of critical infrastructures continue to be the norm rather than the exception, despite the dramatic rise in spending on network security. Attackers today can easily leverage a distributed and programmable infrastructure of compromised machines (or botnets) to launch large-scale and sophisticated attack campaigns. In contrast, the defenders of our critical infrastructures are fundamentally crippled as they rely on fixed capacity, inflexible, and expensive hardware appliances deployed at designated "chokepoints". These primitive defense capabilities force defenders into adopting weak and static security postures configured for simple and known attacks, or otherwise risk user revolt, as they face unpleasant tradeoffs between false positives and false negatives. Unfortunately, attacks can easily evade these defenses; e.g., piggybacking on popular services (e.g., drive-by-downloads) and by overloading the appliances. Continuing along this trajectory means that attackers will always hold the upper hand as defenders are stifled by the inflexible and impotent tools in their arsenal.
The goal of our research is to change the dynamics of this attack-defense equation. Instead of taking a conventional approach of developing attack-specific defenses, our work focuses on empowering defenders with the right tools and abstractions to tackle the constantly evolving attack landscape. To this end, we envision a new software-defined approach to network security, where we can rapidly develop and deploy novel in-depth defenses and dynamically customize the network's security posture to the current operating context.
In this talk, I will give an overview of our recent work in this space.
Bio: Vyas Sekar is an Assistant Professor in the ECE Department at CMU, where he is part of CyLab. His research interests lie at the intersection of networking, security, and systems. He received his Ph.D. from the Computer Science Department at Carnegie Mellon University in 2010. He earned his bachelor's degree from the Indian Institute of Technology Madras, where he was awarded the President of India Gold Medal. His work has been recognized with best paper awards at ACM SIGCOMM, ACM CoNext, and ACM Multimedia.
Suresh Ravindran – Cavium – “Xpliant - Programmable Data Plane Switch Architecture for Software Defined Data Centers”
Abstract: The programmable switching hardware data plane is a critical component of Software Defined Networking (SDN) for true innovation and real life network optimization. SDN is not really about remote feature configuration of a fixed pipeline; it truly achieves its purpose only when software can define a new behavior in the hardware data plane. True innovation is possible only when smart minds can develop their own ideas and define it on hardware to function at high throughput. Often an innovative new networking protocol invention fails because its definition makes it complex to implement in hardware. A programmable network pipeline needs to be part of these tools for successful innovation. Different data centers can customize hardware to optimize for their feature services using the same ASIC at different places in their network.
Bio: Suresh Ravindran is the head of the software engineering division and application engineering of the switch platform group at Cavium Networks. A networking industry veteran, he built enterprise and datacenter networking switches and routers for many years, and worked for Brocade, Foundry Networks, Marconi Communications, FORE Systems and Wipro in the past. His engineering degree is in Computer Science and Engineering from TKMCE, University of Kerala, India in 1996.
Prof. Swarun Kumar – Carnegie Mellon – "Cloud-Assisted Networks of Wireless Sensors”
Abstract: Recent years have seen much interest in networking wireless sensors to the cloud. In this talk, I will describe the challenges and opportunities of doing so in the context of driverless cars and multi-robot networks. I will present a cloud-assisted system for autonomous driving. Our solution enables the cloud to have access to sensor data from autonomous vehicles as well as the roadside infrastructure. The cloud assists autonomous vehicles that use this system to avoid obstacles such as pedestrians and other vehicles that may not be directly detected by sensors on the vehicle. Further, it enables vehicles to plan efficient paths that account for unexpected events such as road-work or accidents. We evaluate our system on an outdoor testbed including an autonomous golf car and six iRobot Create robots. Results show a reduction in the average time vehicles need to detect obstacles such as pedestrians by 4.6x compared to today's systems that do not access the cloud.
Bio: Swarun Kumar is an assistant professor at CMU where he works on wireless networks and systems. He designs and builds novel systems that leverage a deep understanding of the wireless physical layer to enable faster wireless networks and new services. His work has been featured as research highlights in the Communications of the ACM (CACM) and the International Journal of Robotics Research (IJRR). Swarun is a recipient of the George Sprowls Award for best Ph.D thesis in Computer Science at MIT and the President of India gold medal at IIT Madras.
Prof. Yuvraj Agarwal – Carnegie Mellon – “Towards Building a Safe, Secure and Easy-to-Use Internet-of-Things Infrastructure”
Abstract: A world enabled by the Internet of Things (IoT) promises to lead to truly connected environments, where people and things collaborate to improve the overall quality of life. The IoT will give us actionable information at our fingertips, without us having to ask for it or even recognizing that it might be needed. However, in order to realize this vision, and build out a successful IoT infrastructure, there are a number of challenges that must be addressed. These challenges include safety, security and privacy in a world of pervasive IoT, scalability in terms of data storage and processing architectures, new programming abstractions to allow ease of programming, etc. At Carnegie Mellon, we are currently building an end-to-end IoT stack, called GIoTTO, to address some of these challenges. In this talk I will give an overview of our ideas and progress over the past year on this IoT Expedition (www.iotexpedition.org). If time permits I will specifically talk about some of the security and privacy challenges that IoT presents.
Bio: Yuvraj Agarwal is an Assistant Professor of Computer Science in the School of Computer Science at Carnegie Mellon University. His research interests are at the intersection of Systems and Networking and Embedded Systems, and he is particularly excited about research problems that benefit from using hardware insights to build more scalable and energy efficient systems. In recent years, his work has focused on the domain of Green Computing, Mobile Computing, Energy Efficient Buildings and the Internet of Things with a cross cutting interest on security and privacy issues.
Prof. Jimmy Zhu – Carnegie Mellon – “All Spin Logic Devices and Circuits for Future Electronics”
Abstract: Today, electronic devices are charge-based and operate by moving, accumulating, or draining charges for data processing. The flow of electric charges results in Joule heating by Ohmic loss, causing exhaustion of a substantial percentage of the power consumed while limiting the device performance. In this talk, an alternative scheme is proposed with utilization of electron spins by generating charge-free spin current to eliminate/reduce Joule-heating. A new switching device and circuit technology, mLogic, based on current signals and pulse voltage supplies for power and clocking enables implementation of complete large-scale systems without semiconductors, a potential candidate for future low power electronics.
Bio: Jian-Gang (Jimmy) Zhu is the ABB Professor of electrical and computer engineering at Carnegie Mellon University and the director of the Data Storage Systems Center. He received his Ph.D. in Physics from University of California at San Diego in 1989. Prior to coming to Carnegie Mellon in 1997, he had been a faculty in the Department of Electrical Engineering at the University of Minnesota since 1990. Some of the awards that he received include the McKnight Land Grant Professorship from the University of Minnesota in 1992, the NSF Presidential Young Investigator Award in 1993, the R&D Magazine Top 100 Invention Award in 1996, and the IEEE Magnetic Society Achievement Award in 2011. He was IEEE Magnetic Society Distinguished Lecturer in 2004. He has published over 290 refereed journal papers along with seven book chapters and has given over 90 invited papers at major international conferences. He holds 20 U.S. patents. He is a Fellow of IEEE.
Robert Broberg – Cisco – “Trust and Integrity Assurance from IoT to the Cloud Tenant”
Abstract: A look at Trustable systems and Cisco's Advanced Security Research efforts to future proof the Internet against attacks.
Bio: Robert Broberg is a Distinguished Engineer at Cisco Systems. Robert started his career in networking at Ungermann-Bass in 1984 and since then has been leading innovations in IP networking spanning everything from device drivers, applications to large system design. His career has spanned many roles. Some highlights include, 2 years in Tokyo leading network the industry transition from XNS to TCP, 3 years at Bell-Labs in Physical Systems research as IP transitioned from a TDM based phone client to owning Optical transport and 3 years at AYR netkworking, a startup using Linux as an OS for high performance ASIC assisted routers.
Robert is currently part of the newly formed Advanced Security Research Group at Cisco. This group has a vision to address security in the 5,10 to 15 year horizon by partnering worldwide with Academics, Industry consortiums and leading research institutions.
Dr. Michael Gaidis – Samsung – “Novel Nonvolatile Memories in the Memory/Storage Hierarchy for Cloud and IoT Applications”
Abstract: Advances in flash memory bandwidth, latency, and capacity have led to system architectures with potential for much higher performance than predecessors using hard disk drives. A resulting conundrum is the need for large-scale software redesign to realize the full advantages offered by the new high-performance storage. This "ecosystem" development for flash has been underway for several years already, and is far from complete. The situation becomes even more complex when emerging nonvolatile memories are considered. This presentation will cover the existing memory/storage hierarchy and the tradeoffs that will dictate what role different types of emerging nonvolatile memories will play in future system architectures. We will address questions such as: where nonvolatile memories will have the greatest impact; how they might best be implemented as tiering between, or as replacement for DRAM and flash; and where one needs to be careful of excessive hype.
Bio: Michael Gaidis is a Senior Manager with Samsung Device Solutions in San Jose, California, focused on university relationships and memory strategy. Michael works closely with the Memory Solutions Lab (flash, DRAM, memory systems), the Device Lab (modeling of advanced logic and memory devices), and the New Memory Technology Lab (MRAM). He has previously worked with device-level research including resistive RAM, MRAM, and terahertz components. This included work at HGST's San Jose Research Lab, IBM's TJ Watson Research Center, and NASA's Jet Propulsion Lab. His education included a postdoctoral fellowship at Caltech, PhD at Yale, and MS/BS degrees at MIT. He can be reached at [email protected]
Dr. Sam Fineberg – Hewlett Packard Enterprise – “Software Defined Storage – from industry standard appliances to hyper-converged systems”
Abstract: Software defined storage is an emerging concept that enables storage to be provisioned dynamically, as software running on a pool of industry standard severs or even virtual machines. This talk will present an overview of this market, as well as the major storage types and how they are driving the storage industry. Topics will also include the major deployment models for software defined storage such as hardware appliances, cloud, VSA, and hyper-converged. The talk will also describe HPE’s approach to software defined storage and key technologies that address the market.
Bio: Sam Fineberg is a Distinguished Technologist in the Hewlett Packard Enterprise (HPE) Storage Chief Technologist Office. Dr. Fineberg has over 25 years of experience in areas including storage, cloud computing, archiving, non-volatile memory, high performance computing, and data preservation. Sam leads cloud and big data technical strategy for HPE’s multi-billion dollar storage division, driving participation in the OpenStack cloud storage community and solutions for Hadoop/Spark. Prior to his role in HPE Storage, Dr. Fineberg was chief technologist of HPE’s Information Management Software business, where he architected an archival storage system, and led the acquisition and integration of Tower Software. He also developed a record breaking Terabyte sort application and early network attached non-volatile memory technology as part of HPE’s NonStop Labs. Prior to HPE, Dr. Fineberg worked NASA Ames Research Center where he led key research in high performance computing, message passing, and parallel I/O. He earned a PhD in ECE from U of Iowa, and MSEE and BSEE from Purdue. He can be reached at [email protected]
Dr. Hui Huang – Google – "PerfKit Benchmarker: Performance Benchmarking for Cloud"
Abstract: With the increasing popularity of cloud computing, we've seen a growing demand for living benchmarking of cloud infrastructures. In this talk we will introduce PerfKit Benchmarker, which is an industry effort led by Google to establish a Cloud benchmarking suite. The goal is to represent how cloud developers are building applications and evaluating varied public/private clouds. We will also cover the design methodology, architecture of PerfKit Benchmarker, and a couple of performance challenges we've seen with cloud computing.
Bio: Dr. Hui Huang is a performance engineer in Google Cloud/Platform team. She is responsible for public cloud performance evaluation and technical infrastructure improvement for Google Compute Engine (GCE). Hui is also the Google representative/committee member of Standard Performance Evaluation Corporation (SPEC) Cloud, and an active contributor to PerfKit Benchmarker. Before joining Google in 2015, she co-founded Falcon Computing Solutions, which is a prestigious startup targeting on providing heterogeneous accelerating services in data centers. Dr. Huang holds a PhD in Computer Science from UCLA, and BS from Peking University, China. She can be reached at [email protected].
Dr. Khalid Elgazzar – Carnegie Mellon – “Leveraging Edge Computing to Support Internet of Things Scenarios”
Abstract: The IoT vision is to make the world better connected and enable smarter interactions between devices, people, and everyday artifacts. However, the inherent resource constraints of IoT artifacts limit the quality and type of functionality that can be offered, inhibiting IoT scenarios from reaching full potential. Computation offloading offers resource-constrained IoT devices the opportunity to transfer compute-intensive tasks to more resourceful computing infrastructures on the network edge (aka Edge Computing). At Carnegie Mellon, we are currently building GIoTTO, an end-to-end IoT open stack . In this talk, I will present our efforts to leverage cloud computing and cloudlets to support IoT devices to reduce the time required to deliver the end-to-end IoT scenarios, including best practices for interoperability and composability between cloud and edge architectures.
Bio: Dr. Khalid Elgazzar is a postdoctoral research fellow at Carnegie Mellon School of Computer Science and an adjunct Assistant Professor in the School of Computing at Queen's University, Canada. Dr. Elgazzar received his PhD from Queen's University in 2013. He also received the 2014 Queen's School of Computing Distinguished Research Award. His research interests span the areas of distributed systems, mobile cloud and ubiquitous computing, context-aware cyber-physical systems, and elastic networking paradigms. Dr. Elgazzar has received several recognitions and best paper awards at top international venues. He is currently a member of the Google IoT Expedition team, leading the creation of an open stack for IoT.