New challenges for distributed computing at the CMS experiment

Krammer, N.

doi:10.1088/1748-0221/15/07/c07038

Cited by 2 publications

(3 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The challenges to face were (and still are): to keep operating the system reliably enough with significantly less effort, to evolve towards new, more open and flexible computing models, and to provide the software and the resources needed by the LHC experiments. In terms of data processing, the rate of advances in hardware performance has slowed in recent years, encouraging the community to adapt and to take advantage of developments such as graphics processing unit (GPU), high-performance computing (HPC) and commercial Cloud services [44]. Cloud computing is a technology largely viewed as the next big step in the development and deployment of an increasing number of distributed applications.…”

Section: Future Challenges: Generic Detector Randd Computing and Soft...mentioning

confidence: 99%

Next frontiers in particle physics detectors: INSTR2020 summary and a look into the future

Titov¹

2020

J. Inst.

View full text Add to dashboard Cite

The physics goals of high luminosity particle accelerators, from LHC to HL-LHC and to the next generation of lepton colliders, have set quite stringent constraints on the future needs at the Instrumentation Frontier. Many technologies are reaching their sensitivity limit and new approaches need to be developed to overcome the currently irreducible technological challenges. The detrimental effect of the material budget and power consumption represents a very serious concern for a high-precision silicon vertex and tracking detectors. One of the most promising areas is CMOS sensors offering low mass and potentially radiation-hard technology for the future proton-proton and electron-positron colliders, intensity frontier and heavy-ion experiments. MPGDs have become a well-established technique in the fertile field of gaseous detectors; these will remain the primary choice whenever the large-area coverage with low material budget is required. Vacuum tube technology is inherently fast and new developments include advances in microchannel plates for photomultipliers with a potential for a picosecondtime resolution in large systems. Several novel concepts of picosecond-timing detectors will have numerous powerful applications in particle identification, pile-up rejection and event reconstruction, and serve numerous scientific goals. The story of modern calorimetry is a textbook example of physics research driving the development of an experimental method. Silicon photomultipliers have seen a rapid progress in the last decade, becoming the standard solution for scintillator-based devices. The integration of advanced electronics and data transmission functionalities plays an increasingly important role and needs to be addressed. Bringing the modern algorithmic advances from the field of machine learning from offline applications to online operations and trigger systems is another major challenge. The timescales spanned by future projects in particle physics, ranging from few years to many decades, constitute a challenge in itself, in addition to the complexity and diversity of the required accelerator and detector R&D. This paper summarizes advances and recent trends in the instrumentation techniques for particle physics experiments, largely based on the presentations given at the International Conference "Instrumentation for Colliding Beam Physics" (INSTR-20), held at BINP Novosibirsk, Russia, from 24 to

show abstract

Section: Future Challenges: Generic Detector Randd Computing and Soft...mentioning

confidence: 99%

Next frontiers in particle physics detectors: INSTR2020 summary and a look into the future

Titov¹

2020

J. Inst.

View full text Add to dashboard Cite

show abstract

“…The advent of computing and information technology, especially the widespread adoption of internet technology, digital platforms, mobile devices and cloud computing, has fueled explosive growth in Generate data in many different fields. This increase in data production manifests itself in many different types, characterized by extensive noise and high complexity [1].…”

mentioning

confidence: 99%

“…In addition, a provincial police office accumulated 20 billion pieces of road vehicle monitoring data over three years, a total of 120 TB. Forecasts from IDC, a leading computer information analysis and consulting company, predict that the global annual volume of data will reach 35 ZB by 2020 [1]. The term "big data" has emerged to summarize the nature of such large, unstructured, digitized data sets.…”

mentioning

confidence: 99%

Research and Application of Big Data Clustering Algorithm based on AI Technology in Cloud Environment

Huang,

Zhang

2024

SCPE

View full text Add to dashboard Cite

Traditional big data filling algorithms often give inaccurate results due to vulnerabilities to different data types.To solve this problem, this study presents a new big data clustering algorithm powered by AI technology in a cloud environment. The study proposes an advanced Big Data clustering algorithm that leverages AI technology in a cloud environment. It optimizes clustering based on predicted strength using parallel processing. The research focuses on optimizing the clustering algorithm based on the predicted intensity through parallel processing. Experimental results demonstrate that image clustering stability is achieved when the number of clusters exceeds 4, indicating reduced sensitivity to random factors. Although it was not possible to precisely determine the optimal number of clusters, the use of an optimization algorithm showed that at four clusters the prediction intensity reached its peak, ensuring more accurate cluster identification. Through rigorous testing, the optimal number of clusters was determined to be 4. Clustering results show that visitors characterized by certain attributes show higher interest in most columns. This algorithm makes it easier to cluster incomplete large data, improves clustering speed, and improves the accuracy of filling in missing data. Compared to existing methods, this algorithm leverages AI technology in the cloud environment to optimize clustering based on prediction intensity, providing improved accuracy and efficiency during processing in big data management.

show abstract

New challenges for distributed computing at the CMS experiment

Cited by 2 publications

References 6 publications

Next frontiers in particle physics detectors: INSTR2020 summary and a look into the future

Next frontiers in particle physics detectors: INSTR2020 summary and a look into the future

Research and Application of Big Data Clustering Algorithm based on AI Technology in Cloud Environment

Contact Info

Product

Resources

About