Seminar, Department of Computer Science
Seminar 1: Java in the real world
1:00 - 2:00 pm
The modern enterprise development environment is an exciting place. The
ability to exploit innovation in hardware and software is a critical
differentiator for commercial success.
This talk will discuss some of the core technologies that Goldman Sachs
is currently using to build java applications. It will focus
particularly on the ways in which the firm is taking these technologies
to the limits of their capabilities and how it works with vendors and
the open-source community to ensure stability and scalability.
The talk will also discuss some of the skills that the next generation
of enterprise-development practitioners will need in order to build
applications that can scale to meet the ever-growing demands of the
Paul Murray is a Technology Fellow at Goldman Sachs. His primary areas
of research are in application stability, performance and scalability.
Paul's team acts as the firm's authority on fundamental aspects of java
- such as the virtual machine and the java language - and they provide
consultancy and support to the entire Goldman Sachs developer community.
Paul has been with Goldman Sachs for seven years. He previously worked
in the field of dynamic compiler technologies, spending for four years
in the compiler group at Silicon Graphics, where he specialized in JIT
compiler design. Paul also spent seven years at Insignia Solutions where
he was responsible for the design and implementation of the JIT-based
CPU emulation technology used in their flagship "SoftPC" product.
Seminar 2: How to design networks for low latency trading applications
2:00 - 3:00 pm
Nowadays, the vast majority of trades are being executed
electronically, and human traders are increasingly being replaced by
computers pre-programmed to trade accordingly to complex mathematical
and statistical algorithms. These systems are able to calculate a
multitude of variables and execute thousands of trades in a
split-second. They make their money by placing their order before the
competition, thus every microsecond counts.
This talk will discuss briefly how trading by computers, known as program trading, works. We will then delve into a discussion
of how the underlying networks are built to facilitate this trading, where the sources of latency lie, how we can measure it,
and some of the approaches we can take to minimize latency. We will discuss the various network components and common
architectures of those to understand how we can better design networks to offer low latency for use in trading environments.
Matthew Liste is a Vice President with Goldman Sachs, where he manages
the Network Architecture team. Matthew has more than seventeen years of
experience in networking, and possesses a strong understanding of
network architectures for both enterprise and service provider
Prior to joining Goldman Sachs in 2004, Matthew consulted on
leading-edge network technologies at financial services companies such
as UBS, Deutsche Bank and JPMorganChase. Before that, Matthew worked at
Schlumberger and Nortel Networks, where he received two patents on
network management systems.
In addition to his work at Goldman Sachs, Matthew currently sits on the
board of Nethope, a nonprofit IT consortium of leading international
NGOs focusing on connectivity, and has chaired the program committee
chair for the infrastructure conference at Interop, the premiere global
networking conference, Matthew received his Bachelors of Science degree
from University of Oslo in 1990.
September 19, 2007
Sage 2707 - 1:00 p.m. to 3:00 p.m.
Last updated: September 13, 2007