LBNL-SEC-Comment-Final

Cipot - Djl

Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres

4 pages

English

Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres

A propos
Informations
Extrait

Description

Informations

Publié par	Cipot
Nombre de lectures	28
Langue	English

Extrait

August 8, 2010

Securities and Exchange Commission
100 F Street, NE
Washington, DC 20549 -‐1090
Re: Comments on SEC File No. S7 -‐11-‐10: Consolidated Audit Trail
1 2From: David Leinweber and Horst Simon
Computational Research Division
Lawrence Berkeley Natio nal Laboratory

We welcome the opportunity to comment on the proposed rule for a Consolidated Audit Trail.

Lawrence Berkeley National Laboratory (LBNL) is not a familiar name in financial circles, so some
introduction is appropriate. Berkeley Lab, found ed in 1931, is one of the leading scientific research
institutions in the United States and the world. The Laboratory employs over 3900 people, including
31320 scientists and engineers. The FY2010 budget exceeds $770 milion.

Eleven scientists associate d with Berkeley Lab have won the Nobel Prize. One of the most recent is
former Laboratory Director, Steven Chu, now Secretary of Energy. Thirteen Lab scientists have won the
National Medal of Science, our nation's highest award for lifetime achievement in scientific research.

In the Laboratory’s early days, its principal focus was on high energy and nuclear physics. This has
expanded to include research in chemistry, biology, climate, energy, and, most importantly in the
context of this comment, high perfo rmance large -‐scale computing (“supercomputing”). The Laboratory’s
Computational Research Division is a global center of excellence in research across all the disciplines
mentioned, and a national talent and technology resource capable of addressing the sub stantial
research issues posed by the Commission’s interest in a Consolidated Audit Trail. Berkeley Lab is
singularly capable of bringing the world’s most powerful computing machinery to bear on the complex
and dynamic problem of understanding the ever fas ter data streams and interactions in the fast,
fragmented US market.

1David Leinweber, Financial technologist, and author of “Nerds on Wal Street: Math, Machines and
Wired Markets”, Wiley 2009 http://tinyurl.com/nerd sonwallst . Lawrence Berkeley Labs Center for
Innovative Financial Technology http://www.lbl.gov/CS/CIFT.html

2 Horst Simon, Director of Computational Research at Lawrence Be rkeley Laboratory,
http://www.nersc.gov/homes/simon

3 th A retrospective of the Laboratory and its accomplishments was written for the 75 anniversary in
2006. http://www.lbl.gov/Publications/75th/ Twenty five scientific breakthroughs at Berkeley Lab that
improved the world, see http://www.lbl.gov/Publications/breakthroughs/index.html .

The Lab’s supercomputing expertise goes more than forty years. In the 1960s, the computational
challenge for the Lab was to analyze data from high -‐energy physics experiments. In 1974, a
supercomputer at Berkeley Lab was first used for computational chemistry. The NERSC (National Energy
Research Scientific Computing) Center moved to Berkeley in 1995. Originaly established solely to
support magnetic fusion research, NERSC helped pioneer many of the computing practices taken for
granted today. For the last four decades Berkeley Labs and its sister DOE (Department of Energy) labs
provided the impetus for the continued development of “supercomputers”, machines many thousands
of times more po werful than those used in ordinary setings. Many of the largest and most powerful
computers in the world are found at the Laboratory and its affiliated institutions.

Applications for these machines have grown to include complex problems in many areas of physics,
biology, chemistry, energy and information technology. Berkeley Lab’s capabilities in large -‐scale data
management, analysis, visualization, and simulation are unsurpassed. Synchronization of multiple
disparate sources, and events that happen v ery quickly, well below the millisecond level, are recurring
themes in the Laboratory’s research. These issues are central to an effective Consolidated Audit Trail
system.

Close collaboration with industry to meet the needs of the most demanding computati onal tasks has
been a hallmark of the Laboratory’s computational research. This practice can bring the same unrivaled
research and development skills to bear in the challenge of monitoring and understanding increasingly
complex, ever -‐faster interconnected markets. This could be the foundation for a strong research
component in the efforts of the SROs regarding the Consolidated Audit Trail.

With due respect to some of the other responders, we do not believe that there is a quick of -‐the -‐
shelf solution for t he Consolidated Audit Trail. In high performance computing, large problems are
characterized by their scale, in terms of data sizes and computational demands. The march of
technology has meant that yesterday’s large problems are solved on tomorrow’s deskt ops. “Mega”
problems, involving megabytes of data were overwhelming to early computers with Kilobytes of storage.
Many “giga” scale problems are today accesible to desktop machines. Coping with “tera” scale data is a
challenge in convent

Univers
Ebooks
Livres audio
Presse
Podcasts
BD
Documents

LBNL-SEC-Comment-Final

YouScribe

Le catalogue

Le service

Les conditions