A5101621571- Distributed and Parallel Computing Systems
- Advanced Data Storage Technologies
- Scientific Computing and Data Management
- Particle physics theoretical and experimental studies
- Particle Detector Development and Performance
- Parallel Computing and Optimization Techniques
- Cloud Computing and Resource Management
- Caching and Content Delivery
- Distributed systems and fault tolerance
- Dark Matter and Cosmic Phenomena
- Software-Defined Networks and 5G
- High-Energy Particle Collisions Research
- Peer-to-Peer Network Technologies
- Software System Performance and Reliability
- Research Data Management Practices
- Opportunistic and Delay-Tolerant Networks
- Neutrino Physics Research
- Cloud Data Security Solutions
- Astrophysics and Cosmic Phenomena
- Advanced Database Systems and Queries
- Mathematical and Theoretical Epidemiology and Ecology Models
- Particle accelerators and beam dynamics
- Biomedical Text Mining and Ontologies
- Evolution and Genetic Dynamics
- Radiation Effects in Electronics
Morgridge Institute for Research
2019-2024
University of Notre Dame
2021
University of Wisconsin–Madison
2021
University of Nebraska–Lincoln
2011-2020
Fermi National Accelerator Laboratory
2019
University of Pittsburgh
2017
Institute of High Energy Physics
2012
Austrian Academy of Sciences
2012
University of Nebraska System
2011
University of West Georgia
2004
Particle physics has an ambitious and broad experimental programme for the coming decades. This requires large investments in detector hardware, either to build new facilities experiments, or upgrade existing ones. Similarly, it commensurate investment R&D of software acquire, manage, process, analyse shear amounts data be recorded. In planning HL-LHC particular, is critical that all collaborating stakeholders agree on goals priorities, efforts complement each other. this spirit, white paper...
Rucio is an open-source software framework that provides scientific collaborations with the functionality to organize, manage, and access their data at scale. The can be distributed across heterogeneous centers widely locations. was originally developed meet requirements of high-energy physics experiment ATLAS, now continuously extended support LHC experiments other diverse communities. In this article, we detail fundamental concepts Rucio, describe architecture along implementation details,...
While the LHC data movement systems have demonstrated ability to move at necessary throughput, we identified two weaknesses: latency for physicists access and complexity of tools involved. To address these, both ATLAS CMS begun federate regional storage using Xrootd. Xrootd, referring a protocol implementation, allows us provide all disk-resident from single virtual endpoint. This "redirector" discovers actual location redirects client appropriate site. The approach is particularly...
Following the success of XRootd-based US CMS data federation, AAA project investigated extensions federation architecture by developing two sample implementations an XRootd, disk-based, caching proxy. The first one simply starts fetching a whole file as soon open request is received and suitable when completely random access expected or it already known that be read. second implementation supports on-demand downloading partial files. Extensions to Hadoop Distributed File System have been...
Recent gravitational-wave observations from the LIGO and Virgo observatories have brought a sense of great excitement to scientists citizens world over. Since September 2015,10 binary black hole coalescences one neutron star coalescence been observed. They provided remarkable, revolutionary insight into "gravitational Universe" greatly extended field multi-messenger astronomy. At present, Advanced can see out redshift 0.6 0.05. This probes only very small fraction volume observable Universe....
The HTCondor high throughput computing system is heavily used in the energy physics (HEP) community as batch for several Worldwide LHC Computing Grid (WLCG) resources. Moreover, it backbone of GlidelnWMS, pilot by organization Compact Muon Solenoid (CMS) experiment. To prepare Run 2, we probed scalability limits new versions and configurations with a goal reaching 200,000 simultaneous running jobs single internationally distributed dynamic pool.
The HTCondor-CE is the next-generation gateway software for Open Science Grid (OSG). This responsible providing a network service which authorizes remote users and provides resource provisioning (other well-known gateways include Globus GRAM, CREAM, Arc-CE, Openstacks Nova). Based on venerable HTCondor software, this new CE simply highly-specialized configuration of HTCondor. It was developed adopted to provide OSG with more flexible, scalable, easier-to-manage software. Further, focus not...
During 2015 and 2016, the Laser Interferometer Gravitational-Wave Observatory (LIGO) conducted a three-month observing campaign. These observations delivered first direct detection of gravitational waves from binary black hole mergers. To search for these signals, LIGO Scientific Collaboration uses PyCBC pipeline. deliver science results in timely manner, collaborated with Open Science Grid (OSG) to distribute required computation across series dedicated, opportunistic, allocated resources....
The CMS experiment will need to sustain uninterrupted high reliability, throughput and very diverse data transfer activities as the LHC operations start. PhEDEx, system, be responsible for full range of needs experiment. Covering entire spectrum is a demanding task: from critical high-throughput transfers between CERN Tier-1 centres, high-scale production among Tier-2 managing 24/7 all 170 institutions in providing straightforward access handful files individual physicists.
Electrical power requirements will be a constraint on the future growth of Distributed High Throughput Computing (DHTC) as used by Energy Physics. Performance-per-watt is critical metric for evaluation computer architectures cost- efficient computing. Additionally, performance come from heterogeneous, many-core, and high computing density platforms with specialized processors. In this paper, we examine Intel Xeon Phi Many Integrated Cores (MIC) co-processor Applied Micro X-Gene ARMv8 64-bit...
The management of security credentials (e.g., passwords, secret keys) for computational science workflows is a burden scientists and information officers. Problems with expiration, privilege mismatch) cause to fail fetch needed input data or store valuable scientific results, distracting from their research by requiring them diagnose the problems, re-run computations, wait longer results. In this paper, we introduce SciTokens, open source software help manage more reliably securely. We...
CMS will require access to more than 125k processor cores for the beginning of Run 2 in 2015 carry out its ambitious physics program with and higher complexity events. During Run1 these resources were predominantly provided by a mix grid sites local batch resources. long shut down cloud infrastructures, diverse opportunistic HPC supercomputing centers made available CMS, which further complicated operations submission infrastructure. In this presentation we discuss effort adopt deploy...
Data distribution for opportunistic users is challenging as they neither own the computing resources are using or any nearby storage. Users motivated to use expand their data processing capacity, but require storage and fast networking distribute that processing. Since it requires significant management overhead, rare resource providers allow access Additionally, in order at several distributed sites, assume responsibility maintain data.
We describe a finite element method based on piecewise pluriharmonic or pluribiharmonic splines to numerically approximate solutions partial differential equations the product SG 2 of two copies Sierpinski gasket. use this experimentally study both elliptic equations, and new class operators that we call quasielliptic, which has no analog in standard theory pde's. The existence these is observation set ratios eigenvalues Laplacian gaps. explicitly prove such gap exists around value √5.
Data distribution, storage and access are essential to CPU-intensive data-intensive high performance Grid computing. A newly emerged file system, Hadoop distributed system (HDFS), is deployed tested within the Open Science (OSG) middleware stack. Efforts have been taken integrate HDFS with other tools build a complete service framework for Storage Element (SE). Scalability tests show that sustained inter-DataNode data transfer can be achieved cluster fully loaded data-processing jobs. The...
Bosco is a software project developed by the Open Science Grid to help scientists better utilize their on-campus computing resources. Instead of submitting jobs through dedicated gatekeeper, as most remote submission mechanisms use, it uses built-in SSH protocol gain access cluster. By using common method, SSH, we are able simplify interaction with cluster, making process more user friendly. Additionally, does not add any extra be installed on cluster an attractive option for administrator....
Hadoop is an open-source data processing framework that includes a scalable, fault-tolerant distributed file system, HDFS. Although HDFS was designed to work in conjunction with Hadoop's job scheduler, we have re-purposed it serve as grid storage element by adding GridFTP and SRM servers. We tested the system thoroughly order understand its scalability fault tolerance. The turn-on of Large Hadron Collider (LHC) 2009 poses significant management challenge; been working introduce solution for...
Named Data Networking (NDN) proposes a contentcentric rather than host-centric approach to data retrieval. packets with unique and immutable names are retrieved from content store (CS) using Interest packets. The current NDN architecture relies on forwarding strategies that dependent upon on-path caching is therefore inefficient. This reduces transfer efficiency by ignoring the cached available adjacent off-path routers in network. In this paper, we propose novel distributed multipath (D-MP)...
Data access is key to science driven by distributed high-throughput computing (DHTC), an essential technology for many major research projects such as High Energy Physics (HEP) experiments. However, achieving efficient data becomes quite difficult when independent storage sites are involved because users burdened with learning the intricacies of accessing each system and keeping careful track location. We present alternate approach: Any Data, Time, Anywhere infrastructure. Combining several...
The CMS Remote Analysis Builder (CRAB) is a distributed workflow management tool which facilitates analysis tasks by isolating users from the technical details of Grid infrastructure. Throughout LHC Run 1, CRAB has been successfully employed an average 350 distinct each week executing about 200,000 jobs per day.