Search NASA⌕ Search

SEARCH · Search NASA

Results for “Multi-tenant”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

13 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Distributed Quantum Learning with co-Management in a Multi-tenant Quantum System

The rapid advancement of quantum computing has pushed classical designs into the quantum domain, breaking physical boundaries for computing-intensive and data-hungry applications with the hope that some systems may provide a quantum speedup. For example, variational quantum algorithms have been proposed for quantum neural networks to train deep learning models on qubits, achieving promising results. Existing quantum learning architectures and systems rely on single, monolithic quantum machines with abundant and stable resources, such as qubits. However, fabricating a large, monolithic quantum device is considerably more challenging than producing an array of smaller devices. In this paper, we investigate a distributed quantum system that combines multiple quantum machines into a unified system. We propose DQuLearn, which divides a quantum learning task into multiple subtasks. Each subtask can be executed distributively on individual quantum machines, with the results looping back to classical machines for subsequent training iterations. Additionally, our system supports multiple concurrent clients and dynamically manages their circuits according to the runtime status of quantum workers. Through extensive experiments, we demonstrate that DQuLearn achieves similar accuracies with significant runtime reduction, by up to 68.7% and an increase per-second circuit processing speed, by up to 3.99 times, in a 4-worker multi-tenant setting.

quantum computing↗
DOE OSTI2021

MAPA: Multi-Accelerator Pattern Allocation Policy for Multi-Tenant GPU Servers

Multi-accelerator servers are increasingly being deployed in shared multi-tenant environments (such as in cloud data centers) in order to meet the demands of large-scale compute-intensive workloads. In addition, these accelerators are increasingly being inter-connected in complex topologies and workloads are exhibiting a wider variety of inter-accelerator communication patterns. However, existing al-location policies are ill-suited for these emerging use-cases. Specifically, this work identifies that multi-accelerator workloads are commonly fragmented leading to reduced bandwidth and increased latency for inter-accelerator communication. We propose Multi-Accelerator Pattern Allocation (MAPA), a graph pattern mining approach towards providing generalized allocation support for allocating multi-accelerator workloads on multi-accelerator servers. We demonstrate that MAPA is able to improve the execution time of multi-accelerator workloads and that MAPA is able to provide generalized benefits across various accelerator topologies. Finally, we demonstrate a speedup of 12.4% for75th percentile of jobs with the worst case execution time reduced by up to 35% against baseline policy using MAPA.

Ranganath, Kiran↗
DOE OSTI2018

Virtual Net Metering Market Development Project Final Technical Report

Virtual net metering (VNEM) is a policy and billing arrangement that allows energy generation credits from a single solar system to be shared virtually among multiple subscribers. In California, VNEM is limited to onsite installations, meaning that the subscribers to that system must be located on the same property as the installation. Subscriber accounts will belong to those living or working in an onsite multi-tenant building, such as tenants within an apartment complex. VNEM enables the expansion of photovoltaic (PV) adoption beyond traditional commercial and single-family roof top systems to those who live in multitenant housing, which is an essential pathway for ensuring equal solar access for all consumers. Since 2008, VNEM has been available in the territories of California’s three large investor-owned utilities (IOUs), including San Diego Gas and Electric (SDG&E), Southern California Edison (SCE), and Pacific Gas and Electric (PG&E), and since then, has evolved into several sub-tariffs, each of which is applicable to specific building types and customer sectors. For instance, from 2008 - 2012, VNEM was only available for the affordable housing sector. In 2012, VNEM was expanded to the general market and was available to any multi-tenant, multi-metered building. It should also be noted that in 2018, an affordable-housing VNEM tariff was expanded to customers within the Liberty Utility and PacifiCorp Utility territories. The use of VNEM for market-rate buildings, since 2012, has been limited, reasons for which our project sought to address.

14 SOLAR ENERGY↗
NASA NTRS2019

Cost Controls in the Cloud

The pay-as-you-go model used by most Cloud Service Providers shifts ownership of resource procurement from finance and management teams to IT, engineering, development and other stakeholders. This requires individual contributors to adopt a more "cost conscious" approach to development, and for projects to invest time into the creation of automated cost controls and cost-focused monitoring. This talk will presents some patterns and examples of cost controls based on our experience managing a multi-tenant cloud environment.

circuit breaker↗
DOE OSTI2024

Flexible silicon photonic architecture for accelerating distributed deep learning

The increasing size and complexity of deep learning (DL) models have led to the wide adoption of distributed training methods in datacenters (DCs) and high-performance computing (HPC) systems. However, communication among distributed computing units (CUs) has emerged as a major bottleneck in the training process. In this study, we propose Flex-SiPAC, a flexible silicon photonic accelerated compute cluster designed to accelerate multi-tenant distributed DL training workloads. Flex-SiPAC takes a co-design approach that combines a silicon photonic hardware platform with a tailored collective algorithm, optimized to leverage the unique physical properties of the architecture. The hardware platform integrates a novel wavelength-reconfigurable transceiver design and a micro-resonator-based wavelength-reconfigurable switch, enabling the system to achieve flexible bandwidth steering in the wavelength domain. The collective algorithm is designed to support reconfigurable topologies, enabling efficient all-reduce communications that are commonly used in DL training. The feasibility of the Flex-SiPAC architecture is demonstrated through two testbed experiments. First, an optical testbed experiment demonstrates the flexible routing of wavelengths by shuffling an array of input wavelengths using a custom-designed spatial-wavelength selective switch. Second, a four-GPU testbed running two DL workloads shows a 23% improvement in job completion time compared to a similarly sized leaf-spine topology. We further evaluate Flex-SiPAC using large-scale simulations, which show that Flex-SiPAC is able to reduce the communication time by 26% to 29% compared to state-of-the-art compute clusters under representative collective operations.

Wu, Zhenguo (ORCID:0000000322847985)↗
DOE OSTI2022

A System Approach to Deep Heating Savings Through Measurement, Management, and Motivation

Across multi-tenant commercial office and multifamily buildings, centrally metered fuel use represents a substantial fraction of whole-building energy use. Energy audit practitioners understand that improving heating distribution efficiency is typically more of an opportunity than combustion efficiency and that differing thermal comfort preferences between tenants are the bane of operators across these building typologies. There is an unmet market need for retrofit technologies that allow for the delivery of the right amount of heat to the right spaces, at the right time. The Energy Management and Information System (EMIS) package fills this gap through enhanced controls and metering, incorporating low-cost sensors and wireless communication infrastructure to provide a platform for ongoing commissioning and tenant feedback, including heat cost allocation. With support from the US DOE Building Technologies Office, Steven Winter Associates, Inc. (SWA) partnered with Sentient Buildings, E Source, building owners, and utility and policy stakeholders, to demonstrate a market viable EMIS that achieves a reduction in space heating energy use by reducing heating load, improving control, and positively impacting behavior while providing an acceptable financial return. In this study, EMIS packages were implemented in two New York City multifamily rental buildings. Both buildings conducted basic mechanical work (e.g., repairing steam traps) to ensure the heating system was operating well before any tenant feedback was layered in. Heating Energy Use Reports (HEUR) were created to provide tenants with social comparisons and energy savings tips to influence their behavior; these were provided monthly to all tenants in both buildings. Additionally, one building allocated heating costs to a portion of the tenants. Heat cost allocation (HCA) has a long history in the European Union (EU), although it is not common in the US or in steam-heated buildings. SWA leveraged existing EU best practices and stakeholder feedback to develop a Heat Cost Allocation algorithm that was considered equitable and intuitive. Energy use and tenant behavior impacts were tracked throughout the study. The basic mechanical repair work saved between 11-20% of heating energy. Those savings rose to 17-24% with the addition of tenant feedback. While it may not be possible to precisely determine the impact of COVID-19 on research studies like this, there may have been additional savings realized had the study taken place in a period of normal occupancy patterns. These types of central heating systems have been a blind spot for utilities, who have traditionally had little visibility into detailed behind-the-meter gas usage. Heating energy savings stayed consistent during the coldest months, indicating the potential for utilities to utilize EMIS packages for peak gas demand reductions or demand response programs. Tenant comfort was also improved. Post installation, room temperatures more closely matched thermostat set points. Perhaps due to this greater level of control, the vast majority of tenants being billed for heating were accepting of the allocation costs. And tenants receiving heat cost allocations were more likely to reduce their thermostat setpoints than tenants receiving behavioral feedback without financial impacts were. Variation in building specifics makes it difficult to provide precise energy and financial savings estimates. But within the range of expected conditions, the study identified a few key variables that can have the greatest impact on financial returns: the cost of fuel, the ability and willingness to allocate heating costs to tenants, and a well-functioning heating system as a starting point. This study focused on two multifamily buildings, but additional use cases, such as commercial buildings and affordable housing, should be explored to better understand the full market potential. While this type of upgrade has the potential for deep energy reductions and cost savings, future projects should take into account the balance of costs and benefits between owners and tenants, especially in the affordable, regulated, or other low-to-moderate income (LMI) segments of the market. Rent credits, utility allowances, or a shared savings program are possible options to accelerate adoption of this strategy in these market segments.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
NASA NTRS2019

Cloud Governance at Scale

Operating and maintaining a large multi-tenant ecosystem in the cloud requires scalable solutions to unique technical and process challenges. The Cloud Computing model grants significant permissions to development teams that traditionally were reserved for Data-Center Administrators and Supply-Chain Managers. Earthdata Cloud has worked to re-cast traditional data-center management into a sensible cloud-first model. This talk discusses some of our challenges, solutions, and way ahead.

financial controls↗
NASA NTRSDate not supplied

Mission Operations, Cubed: NASA Marshall Operations Support for SmallSats

SmallSats have come a long way since the Huntsville Operations Support Center (HOSC) at NASA’s Marshall Space Flight Center supported its first “minisatellite” mission in 2010. And just as SmallSats themselves have evolved in those 12 years, so too has the HOSC’s mission support for SmallSats. Marshall Space Flight Center has a long history with payload and mission operations, including support for the Apollo missions to the moon, the Space Shuttle program, and 21 years of continuous around-the-clock science operations support for research aboard the International Space Station. Today, the HOSC is a multi-tenant facility, supporting not only ISS, but also NASA’s Commercial Crew program, the Space Launch System, the Hubble and Chandra observatories and others – including multiple SmallSat missions. Two SmallSat solar sail missions will be among those taking advantage of the HOSC’s resources for planning, training for and executing mission operations – the Near Earth Asteroid (NEA) Scout and Solar Cruiser missions. One of 10 6U CubeSats manifest on the Artemis I launch of NASA’s Space Launch System rocket this year, NEA Scout’s three-year mission will be supported through a more traditional operations concept, with a dedicated Flight Controller staff operating within the HOSC. Scheduled to launch as part of the Interstellar Mapping and Acceleration Probe (IMAP) in February 2025, Solar Cruiser’s 11-month mission will take a next-generation approach to operations by utilizing a multi-mission flight controller concept, as well as Marshall’s Telescience Resource Kit (TreK). TreK provides a suite of software applications and libraries that allow the Mission Operations Center to serve as an in-house ground system which incorporates remote and automation capability options for engineers and scientists. This presentation will compare the approaches the HOSC will use to support these two missions as a way of demonstrating the array of options NASA MSFC offers for operations support for CubeSat and SmallSat missions.

Darren S Wallace↗
DOE OSTI2016

Engaging Tenants in Reducing Plug Load Energy Use

Plug and Process Loads (PPLs) account for an increasingly large percentage of commercial building energy use in the U.S. due to the rising number of energy intensive plug-in devices. In addition, buildings are becoming more and more efficient and plug load energy use has become an increasingly pertinent component to achieving aggressive energy targets and netzero energy status. For multi-tenant buildings, controlling plug loads in tenant spaces can be a significant challenge. Luckily, there are a number of PPL reduction strategies, best practices, and lessons learned from numerous commercial real estate and higher education leaders who have successfully engaged building occupants and tenants in reducing PPL energy use. This paper provides actionable PPL reduction strategies and best practices that building owners and managers can immediately apply to their own buildings.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2019

Evaluating Burst Buffer Placement in HPC Systems

Burst buffers (BBs) are increasingly exploited in contemporary supercomputers to bridge the performance gap between compute and storage systems. The design of BBs, particularly the placement of these devices and the underlying network topology, impacts both performance and cost. As the cost of other components such as memory and accelerators is increasing, it is becoming more important that HPC centers provision BBs tailored to their workloads.This work contributes a provisioning system to provide accurate, multi-tenant simulations that model realistic application and storage workloads from HPC systems. The framework aids HPC centers in modeling their workloads against multiple network and BB configurations rapidly. In experiments with our framework, we provide a comparison of representative Oak Ridge Leadership Computing Facility (OLCF) I/O workloads against multiple BB designs. We analyze the impact of these designs on latency, I/O phase lengths, contention for network and storage devices, and choice of network topology.

Khetawat, Harsh↗
DOE OSTI2023

STRATOS: A cyber-energy SAAS platform to manage real-time experiment configuration and deployment [SWR-23-16]

STRATOS is a novel management SAAS platform for achieving multi-environment, multi-tenant real-time cyber-energy experimentation. The platform empowers users to configure a wide variety of experiment environments by automating the initialization process and providing a framework for building system configurations. During run-time the platform ensures resources are allocated appropriately and the environment remains stable until the required cyber-energy events have occurred. This capability to manage this complex process in a single user-focused application significantly reduces difficulty and setup time for performing high-impact cyber-energy research.

Van Natta, Joshua↗