Cisco Cisco Nexus 5010 Switch Dépliant
2 © 2011 Cisco and/or its affiliates. All rights reserved.
To establish competitive differentiation and profit in this market, firms are focusing
on the following business imperatives:
• Speed of execution
• Instant access to market data
• Ability to sustain peaks of activity
on the following business imperatives:
• Speed of execution
• Instant access to market data
• Ability to sustain peaks of activity
When peaks in traffic occur for a very short time (microseconds), they are called
microbursts. A microburst can overwhelm a non-optimized network and cause
transactions to be lost or require retransmission, which has significant negative
impact on a firm’s financial performance. It is not unusual for a microburst to exceed
a firm’s allocated network bandwidth or average link utilization by another degree of
magnitude. Because of their transient nature, microbursts can go undetected if the
monitoring tools are not granular enough.
microbursts. A microburst can overwhelm a non-optimized network and cause
transactions to be lost or require retransmission, which has significant negative
impact on a firm’s financial performance. It is not unusual for a microburst to exceed
a firm’s allocated network bandwidth or average link utilization by another degree of
magnitude. Because of their transient nature, microbursts can go undetected if the
monitoring tools are not granular enough.
The impact of microbursts on trading may be huge. What’s more, based on
projected volume increases in the markets, the frequency of occurrences is
increasing. One packet of information dropped because of a microburst that
exceeds the network bandwidth can dramatically increase trading time, requiring
retransmissions and possibly resulting in complete data loss and therefore
transaction loss. Since an HFT algorithm performs thousands of trades per second,
every retransmission can translate into significant revenue loss. The challenge is to
minimize trading latency end-to-end while handling peaks without packet loss.
projected volume increases in the markets, the frequency of occurrences is
increasing. One packet of information dropped because of a microburst that
exceeds the network bandwidth can dramatically increase trading time, requiring
retransmissions and possibly resulting in complete data loss and therefore
transaction loss. Since an HFT algorithm performs thousands of trades per second,
every retransmission can translate into significant revenue loss. The challenge is to
minimize trading latency end-to-end while handling peaks without packet loss.
Solution Overview
Cisco’s Algo Speed High-Frequency Trading (HFT) solution is part of the
comprehensive Cisco High-Performance Trading (HPT) solution portfolio and
integrated architecture, built to address the trading value chain from the front office
to the back office.
comprehensive Cisco High-Performance Trading (HPT) solution portfolio and
integrated architecture, built to address the trading value chain from the front office
to the back office.
Firms establish and maintain their competitive edge by continually evolving their
trading strategies and increasing the speed of trading. This stresses technology
architectures that are not agile enough to be easily managed and upgraded. An
agile architecture, such as the Cisco architecture underpinning Cisco HPT solutions,
integrates the latest technologies from both network and application domains. It is
modular, scalable, and manageable, providing a path to evolve each component with
minimal disruption to the overall architecture.
trading strategies and increasing the speed of trading. This stresses technology
architectures that are not agile enough to be easily managed and upgraded. An
agile architecture, such as the Cisco architecture underpinning Cisco HPT solutions,
integrates the latest technologies from both network and application domains. It is
modular, scalable, and manageable, providing a path to evolve each component with
minimal disruption to the overall architecture.
The Cisco Algo Speed HFT solution delivers robust network connectivity without
packet loss. The solution is tested and optimized for speed with the messaging
middleware, the latency monitoring, and the server network adapters from Cisco’s rich
ecosystem of partners. Built for HFT clients who have unique requirements for speed,
reliability, agility, and manageability, the solution (Figure 1) includes the following:
packet loss. The solution is tested and optimized for speed with the messaging
middleware, the latency monitoring, and the server network adapters from Cisco’s rich
ecosystem of partners. Built for HFT clients who have unique requirements for speed,
reliability, agility, and manageability, the solution (Figure 1) includes the following:
• Cisco Nexus® 5000, 3000 and 2000 Series Switches: Designed to meet the
stringent, low-latency requirements of data centers for next-generation financial
markets, these switches deliver cost-effective scalability with a transport that can
navigate the transition to 10 Gigabit Ethernet and unified fabric, while maintaining
operational continuity for an environment where system availability is assumed
and maintenance windows are rare, if not totally nonexistent.
markets, these switches deliver cost-effective scalability with a transport that can
navigate the transition to 10 Gigabit Ethernet and unified fabric, while maintaining
operational continuity for an environment where system availability is assumed
and maintenance windows are rare, if not totally nonexistent.
• The Cisco Catalyst® 4900M Switch: An ideal solution for deployments that
require high-performance services at wired-network speed, as well as high
availability, and the modular flexibility to support Gigabit Ethernet, 10 Gigabit
availability, and the modular flexibility to support Gigabit Ethernet, 10 Gigabit
“The dominant factor in determining
whether or not a high-speed network
had a latency problem was how
‘bursty’ the traffic was. In a gigabit
local area network environment, for
example, we saw that microbursts
could result in large spikes of traffic
that could induce upward of 500
milliseconds of delays.”
- Donal Byrne, CEO
Corvil
Corvil is the provider of the CorvilNet
latency monitoring system, which
detects microbursts and is used by
many firms in the HFT business.