The Ninth Workshop on…
	Fault-Tolerant Spaceborne Computing Employing New Technologies, 2016
	SCHEDULE
	Location guide: All activities OUTSIDE the red outline are a the Marriott Hotel (I-40 and Louisiana). The main meeting place for the meeting is at the CSRI building near Sandia labs. The closed sessions will be elsewhere at Sandia, but meet at the CSRI building
	Tuesday, May 31, 2016		Wednesday, June 01, 2016		Thursday, June 02, 2016		Friday, June 03, 2016
8:00 AM	Region outlined in red at Sandia CSRI/90 -->		Breakfast (Sandia)		Breakfast (Sandia)		Breakfast (Sandia)
8:15 AM
8:30 AM	Legend		Intro		Lew Cohn, DoD, ATP Group RH microelectronics program		User group 7 MUG (Sandia) Maestro User Group (POC Bancroft/Crago)
8:45 AM	Administrative and meals		Bill Vanderlinde, IARPA, Computing beyond the end of Moore's Law
9:00 AM	Working group				[Remote] Ran Ginosar, Ramon Chips Ltd, RC64—High Performance Rad-Hard Manycore for Software Defined Satellites
9:15 AM	Presentation session
9:30 AM	Closed session		Peter Kogge, EMU Systems., Migrating Threads and the Emu Architecture		Ken Heffner/John Samson, Honeywell, Space 3D Qualification/DM
9:45 AM
10:00 AM			Ken Heffner, Honeywell, 3D SoC Architecture - TRUST		Thomas Llanos, JHU, Lessons Learned from Applying Risk Management Framework and DoD Survivability Concepts to a  Space Mission
10:15 AM
10:30 AM			Amine Habid, Acelium, TBD		Jon Ballast, Boeing, 14nm near threshold; 32 nm; mem
10:45 AM
11:00 AM			Jim Lyke, AFRL Kirtland, Chiplet architecture		Paul Armijo, Cobham (Aeroflex, Geisler), RHBD update
11:15 AM
11:30 AM			Ken Heffner, Honeywell/COM DEV, Quantum Crypto/Space		Brandon Eames, Sandia, Game Theory Analysis of Trust in Microsystem-based Space Systems
11:45 AM
12:00 PM	Note: Please do not go to Sandia this day; go to the hotel. There will be a registration desk from 12:30 or so		Lunch (Sandia, provided)		Lunch (Sandia, provided)		Limited lunch (Sandia, provided)
12:15 PM
12:30 PM
12:45 PM
1:00 PM	Working group 1 (hotel) Trust: Trust and Fault-Tolerant Computing (POC Heffner).	Working group 2 (hotel) Cubesats (POC John Bellardo)	Daryl Haegley, OSD, Cyber Resiliency and Trust in DoD Industrial Control Systems and Platform Information Technology Systems		Closed session (meet at CSRI building and carpool or caravan to a different location)
1:15 PM
1:30 PM			Ken O'Neill, Microsemi, Rad Hard FPGA update
1:45 PM
2:00 PM			Ashok Ghosh, NM Tech, Fluid-Filled Composite for Radiation Protection
2:15 PM
2:30 PM			Robert Geer, SUNY, 3D async FPGA project
2:45 PM
3:00 PM	Break (provided, hotel)		Break (provided)
3:15 PM
3:30 PM	Working group 1 continued	Working group 2 continued	Kangsen Huey, Xilinx, TBD
3:45 PM
4:00 PM			Kaye Ortiz, DMEA, Trust in Systems Engineering
4:15 PM
4:30 PM			Tyler Lovelly, AFRL/UF, Processor benchmarking (unofficial title)
4:45 PM
5:00 PM			Ned Hanlon, USNA, Advanced Mobile On-orbit Diagnostic System
5:15 PM	Reception (hotel)
5:30 PM			Drive to hotel; socialize in lobby		Dinner on your own
5:45 PM
6:00 PM
6:15 PM
6:30 PM
6:45 PM
7:00 PM	Working group 3 (hotel) Trust: Modeling Space Security (POC Heffner)	Working group 4 (hotel) 3D Memory and FPGAs (POC Bergman)	Dinner served		Working group 5 (hotel) Trust: Enabling Capability (POC Heffner)	Working group 6 (hotel) Neural Networks and Exploratory Issues (POC Cox)
7:15 PM
7:30 PM
7:45 PM			Dinner speaker: Konstantin Batygin, Assistant Professor of Planetary Sciences, Caltech, Planet Nine from Outer Space
8:00 PM
8:15 PM
8:30 PM
8:45 PM
9:00 PM
9:15 PM
9:30 PM
9:45 PM
10:00 PM
	1. Trust: Trust and Fault Tolerant Computing
	The threat to space mission performance by cyber attack vectors presents a complex challenge to the space industry.  A TRUST series of three working sessions comprise a converging strategy for preparing a collaborative research roadmap that meet these forthcoming challenges.  The first scientific area for discussion is resiliency and its significance to space fault tolerant computing.   Participants working in manned, un-manned, satellite and integrated ground, air and sea systems are encouraged to attend and provide guidance to the development of the R&D plan.
	Lead: Ken Heffner (Honeywell)
1:00 PM	Ken Heffner, Honeywell, 2016 Space Computing Workshop Trust Series Objectives
1:15 PM
1:30 PM	Craig Rieger, INL, “Cyber Resiliency and Trust” vs. Fault Tolerance in  the Critical Infrastructure
1:45 PM
2:00 PM
2:15 PM	Daryl Haegley, OSD, Cyber Resiliency and Trust in DoD Industrial Control Systems and Platform Information Technology Systems
2:30 PM
2:45 PM
3:00 PM	Break (provided, hotel)
3:15 PM
3:30 PM	Carl McCants (Bill Vanderlinde), IARPA, TRUST for Microelectronics through TIC, CAT and RAVEN.
3:45 PM
4:00 PM
4:15 PM	Wes Hansford (USC - ISI), Challenges in Design for Security
4:30 PM
4:45 PM
	2. Cubesats
	This CubeSat working group is intended to be an open, wide-ranging discussion of computing challenges surrounding higher risk, lower cost missions.  Seed topics for discussion include:
		• What are new advances expected in onboard computing for CubeSats?
		• How being “CubeSat" impacts the development process
		• What are some of the new mission capabilities (applications) envisioned for CubeSats that have been enabled by advanced onboard computing?
		• New software and hardware development processes for rapid prototyping and flight validation
		• CubeSat operating systems, GN&C, health management, middleware, etc
		• CubeSats pushing the technical boundaries of flight computing, such as multi-processors, power consumption, and miniaturization
	Everyone is encouraged to bring their own ideas and topics for discussion.  The format will be a mix of open discussion and prepared presentations / remarks, with the majority of time spent in open discussion.
	speaker			topic
1:00 PM	John Bellardo (Cal Poly), Larry Bergman (JPL)			Overview of WG, Introductions
1:15 PM	Kyle Kemble (AF)			AFRL CubeSat Portfolio
1:30 PM
1:45 PM	Chris Wilson (Univ of Florida)			Hybrid Space Computing with CSP (CHREC Space Processor)
2:00 PM
2:15 PM	Yutao He (JPL)			Sphinx -- The Rad-Hard CubeSat voionics for Deep Space Missions
2:30 PM
2:45 PM	Open discussion
3:00 PM	Break (provided, hotel)
3:15 PM
3:30 PM	Steve Crago (USC/ISI)			Maestro Flight Experiments
3:45 PM
4:00 PM	John Samson (Honeywell)			Update on the DM ISS Flight Experiments
4:15 PM
4:30 PM	Ned Hanlon (USNA)			Self-Repairing CubeSats (preview of Wed plenary talk)
4:45 PM	Open Discussion			Discuss future challenges and technology opportunities
5:00 PM
	3. Trust: Modeling Space Security
	In this second TRUST Working Group session, research areas for modeling the space platform and the integrated space enterprise will be identified.   Approaches for achieving a cyber system model for space systems and networks will be presented.  Participants with an interest in ientifying research topics for enabling successful deployment and validation of space, cyber system modeling tools are enclouraged to attend.
	Lead. Ken Heffner (Honeywell)
7:00 PM	Ken Heffner, Honeywell, 2016 Space Computing Workshop Trust Series - Modeling Objectives
7:15 PM
7:30 PM	H. Amine, Acelium, A Cyber Ecosystem Stability Model
7:45 PM
8:00 PM	Todd Carpenter, Adventium, Architecture Framework for Fauult Management Assessment and Design (AFFMAD)
8:15 PM
8:30 PM	Kevin Butler, UF FICS, (TBD)
8:45 PM
9:00 PM	Session workup -- output
9:15 PM
	4. 3D Memory and FPGAs
	In this working group session, we discuss opportunities, technology and flight qualification challenges, requirements, and current status on adopting 3D memory cubes for aerospace applications.  Larry Bergman (JPL), Sung Kyu Lim (Georgia Tech)
7:00 PM	Sung Kyu Lim, Georgia Tech, and Larry Bergman, JPL			Introductions, Overview, and goals of working group
7:10 PM	Muhannad Bakir, Georgia Tech			3D and 2.5D Heterogeneous System Integration Technologies: Signal, Power, and Thermal Considerations
7:50 PM	Peter Kogge, Notre Dame			Processor in Memory (PIM) Architectures for 3D Stacking
8:30 PM	Jean Yang-Scharlotta, JPL			How Might We Space Qualify 3D Memory Stacks?
9:10 PM	Open discussion
10:00 PM
	5. Trust: Enabling Capability
	Through the information gathered under the first two TRUST Working Group sessions of the workshop in place, a status of TRUST assessment tools in progress in industry and academia will be presented.  The research challenges for establishing metrics of effectiveness and common research taxonomy will discussed along with the capacity of industry to implement these solutions for Space.  This session is aimed at participants seeking a means to ensure space programs can comply with emerging government requirements for demonstrated TRUST and to have role in summarizing the research and development roadmap for the science supporting the compliance tools.
	Lead: Ken Heffner (Honeywell)
7:00 PM	Kevin Butler,UF - FICS
7:30 PM	Potential MIT Presentation, TBD
8:00 PM	Ken Heffner, SAE G19A, Supply Chain Life Cycle -  Malware Test Methods
8:30 PM	TRUST Series Roadmap Artifact and Future Plans
9:00 PM
10:00 PM
	6. Neural Networks and Exploratory Issues
	Lead: Jonathan Cox, Sandia
7:00 PM	Jonathan Cox, Sandia, Deep Learning issues
7:15 PM
7:30 PM
7:45 PM	IBM True North (via Webex)  IBM’s Neurosynaptic System & TrueNorth Architecture: A million-neuron substrate for low power, real-time, large scale neural network computation.  Enabling the algorithmic power of deep learning with the efficiency of neuromorphic processors.
8:00 PM
8:15 PM
8:30 PM	Matt Marinella, Sandia, Emerging Technologies for the Acceleration of Neuromorphic Algorithms. Neuromorphic algorithms are well suited for implementation with emerging technology, allowing the physics to perform the computation. In particular I will discuss emerging devices, such as memristors, and circuit architectures such as crossbars that are becoming increasingly relevant to neuromorphic computing.
8:45 PM
9:00 PM
9:15 PM	Frances Chance, Sandia, Retina-inspired sensor algorithms (temporary title)
9:30 PM
9:45 PM
10:00 PM
	7. Maestro User Group (MUG)
	The Maestro Users Group will be an open forum for current and potential users of the rad-hard Maestro many-core processor for space. The Maestro processor is based on the commercial Tile64 processor, has 49 cores, and provides up to 25 GFLOPS and 50 GOPS. The MTUG meeting will be an informal and interactive meeting of developers and users to discuss the current state of Maestro hardware and software technology, applications, performance, flight prospects, systems, and user experiences
	Abstracts
	Muhannad Bakir	1. 3D and 2.5D Heterogeneous System Integration Technologies: Signal, Power, and Thermal Considerations
	The aim of our research is to explore the design, fabrication, and characterization of 3D and 2.5D electronic systems and advanced interconnect networks to address the disparity in performance and energy dissipation of interconnects (data movement) relative to computation. Specifically, this presentation will address new approaches to silicon interposer design and fabrication using ultra-low loss polymer embedded copper TSVs for mixed-signal systems. Moreover, we discuss the use of innovative mechanically flexible interconnects to assemble large-scale interposers directly on an FR4 board as well as the formation of silicon-bridging between interposers and chips for large-scale silicon systems. Self-alignment structures with sub-micron accuracy for chip assembly are also described; such accuracy is critical for silicon photonic interconnected systems and silicon bridging applications. Lastly, we are explore advanced cooling approaches for high-performance 3D computing as well as nano-scale interlayer vias in silicon for monolithic-like 3D chip stacking. Thermal and power delivery design considerations are also addressed.
	Konstantin Batygin	Planet Nine from Outer Space
	What if we discovered that our solar system contains a planet no one ever knew about before?
	At the outskirts of the solar system, beyond the orbit of Neptune, lies an expansive field of icy debris known as the Kuiper Belt. The orbits of the individual asteroid-like bodies within the Kuiper Belt are not randomly oriented, and like a jigsaw puzzle, show a lot of large-scale structure when viewed collectively. Much of this structure can be attributed to periodic gravitational pulls exerted by Neptune as it traverses its orbit in a clockwork fashion. However, novel observations that have surfaced over the last two years have revealed a feature of the Kuiper Belt that has, despite numerous efforts, eluded explanation.
	All Kuiper Belt objects with orbital periods longer than 7,000 years have orbits that are physically aligned. Statistically, the chances of such alignment being a coincidence are smaller than a hundredth of a percent. What sculpts this alignment and how is it preserved? In this talk, Batygin will argue that the observed clustering of Kuiper Belt orbits can be maintained by a distant, eccentric, Neptune-like planet, whose orbit lies in approximately the same plane as those of the distant Kuiper Belt objects. In addition to accounting for the observed grouping of trajectories, the existence of such a planet naturally explains other, seemingly unrelated dynamical features of the solar system.
	Batygin and his colleague Michael Brown described their findings in a recent article published in the Astronomical Journal. Professor Batygin’s primary research interests lie in understanding our solar system’s place within its galactic context. He is on the 2015 Forbes list of 30 scientists under 30 who are changing the world.
	Lew Cohn	Advanced Technology Program Group Radiation Hardened Microelectronics Branch Program: Overview and Status
	The objectives of the briefing are to:     Provide an introduction to the Radiation Hardened Microelectronics Branch mission and objectives     Discuss present development efforts concerning RHBD microelectronics technology development efforts and an new initiative to investigate neuromorphic computing
	Kent Dahlgren	HPC for Space
	The paper discusses technology trends in High Performance Computing (HPC) and how these technologies can be leveraged in space platforms to deliver flexible, scalable, and robust computing payloads. Technologies that will be discussed include: Radiation tolerant FPGA devices with integrated high-speed serial transceivers and high performance multicore processors, the RapidIO interconnect with NGSIS extensions, and the OpenCL heterogeneous computing model. Benchmark data showing how OpenCL can be used to accelerate common signal processing and general purpose computing tasks will be presented.
	Ran Ginosar	RC64—High Performance Rad-Hard Manycore for Software Defined Satellites
	Abstract:  RC64 integrates 64 DSP cores, 4 Mbytes on-chip shared memory, telecomm accelerators and high bandwidth I/O in a rad-hard 65nm device. It is accompanies by comprehensive software development tools, sample applications, reference system design and 3U-VPX demonstration boards targeting Space-VPX systems.  The technology has recently been tested for radiation hardness. Prototype chips will be available in 2017.
	Ned Hanlon	Automomous Mobile On-orbit Diagnostic System
	The Autonomous Mobile On-orbit Diagnostic System (AMODS) utilizes CubeSats to provide conventional satellites with cost-effective on-orbit assessment and repair services. AMODS, which will service both new and legacy spacecraft, is comprised of 1) several repair CubeSats (RSats) with manipulable arms that will latch onto and locomote around a host satellite; and 2) one self-propelled transport CubeSat (BRICSat) designed to successively deliver multiple RSats to their respective host spacecraft on-orbit.  This presentation provides an overview of the AMODS system and its potential to effect a paradigm shift in space operations. It details the considerations and required capabilities that guided the design of the BRIC-Sat transfer vehicle and the robotic manipulators and end-effectors on the RSat unit. It reviews data collected from on-orbit and lab tests and concludes with a status summary, including methods of validation.
	Amine Hamdi	A Cyber Ecosystem Stability Model
	There are modeling capabilities that help resolve the issue of fast tactical response decision making during critical security events. Many security technologies focus on the detection and containment of threats as the target response. In our work, we focus on the fast modeling of response actions in real-time and their effects on the ecosystem as a whole. The presentation will outline the philosophy our approach, which centers on finding the perfect balance between accuracy of the model and speed of execution. I will present the modeling that we deploy to achieve this aim, including the multiplicity of the models and why the response needs to be an ecosystem stability model as much as it needs to be a contained precise targeted action. Current testing results show that operators can act much faster when given a smaller set of action choices based on accurate real-time models.
	Peter Kogge	Migrating Threads and the Emu Architecture
	This talk will discuss a new architecture where threads are allowed to migrate freely throughout a large shared memory system. Such an execution model is particularly valuable for applications that are not cache-friendly, especially when there are many "random" accesses without much locality. The first production implementation offers significant performance and power advantages over conventional architectures, and future implementations, especially 3D-based will grow those advantages by large factors.
	Ashok Kumar Ghosh	A Fluid Filled Composite for Radiation Protection
	Many advanced technology systems demand materials with unusual combination of properties that cannot be achieved with the conventional metal alloys, ceramics and polymeric materials, especially the materials that are needed for space applications. In this study, an effort is made to understand the applicability of a fluid filled composite (FFC) developed at New Mexico Tech for deep space exploration. Preliminary radiation analysis using a simulation model (Stopping and Range of Ions in Materials) demonstrates that the FFC can yield a mass saving of 23% and cost saving of ~80% over aluminum. When implemented to design GPS orbit, it is expected to provide significantly superior performance under the anticipated impact, thermal, acoustic, and radiation environments Stopping and Ranging of Ions in Matter (SRIM) simulation results show that different interstitial compound were able to magnify the energy absorption properties of the base composite.
	Peter Kogge	PIM Architectures for 3D Stacking
	This talk will discuss the potential use of 3D stacking to create space computing architectures with significant advantages. The basic idea is to design a "stack of stacks" where substacks perform different functions, but with standardized interfaces using 3D TSVs for interconnect. An image processing example is developed using a 2D FFT.
	Tyler Lovelly	Performance Analysis of Space-Grade Processors with Device Benchmarking
	On-board computing demands for space missions are continually increasing due to growing needs for real-time sensor and autonomous processing coupled with limited communications bandwidth with ground stations. However, the most suitable space-grade processor for any given mission is typically not obvious, with many choices including multi-core and many-core CPUs, DSPs, and FPGAs. Therefore, a study of top-performing, space-grade processors is conducted via device benchmarking to determine which architectures are most suitable for which applications and missions. Additionally, algorithmic and architectural optimizations are considered for several processors to determine which are most critical for improving performance. Preliminary results demonstrate tradeoffs between various architectures and optimizations and can be considered when comparing or designing space-grade processors for future missions.
	Jean Yang	How Might We Space Qualify 3D Memory Stacks?
	3D memories such as the hybrid memory cube and the high bandwidth memory have promised significant performance leaps over their planar counterparts. Commercial 3D NAND has gone even further and demonstrated better density scaling than 2D NAND. Therefore, 3D memory architectures seem likely for flight in the future. To that end, we review current qualification methods for flight memories and discuss potential strategies for developing qualification methods for 3D memories based on current knowledge and experience.
	Date of this document: May 29, 2016