Transforming IT
Showing results for 
Search instead for 
Did you mean: 

The value of a millisecond: winning the war against latency

Michael Lewis’s recent book “Flash Boys: A Wall Street Revolt” has created a storm of interest around High-Frequency Trading (HFT) and its effect on the stock market. As I noted in an earlier blog post, the market is currently dominated by computer-placed orders, trading algorithms, and high-frequency trading, which can lead to huge shifts in the market when it reacts to news and trading activity. The “flash crash” back in 2010 affected approximately $800 billion in market value.


HFT firms make money on low margins with high volumes of trading, and the difference in a cent or two in stock price can equate to millions of dollars. The trades are driven by computerized algorithms and success is measured in milliseconds and microseconds. Latency is the driver of that time delay. It’s the time between the moment a signal is sent and when that signal is received.  HP is certainly no bystander in the ongoing battle against latency at these high-stakes levels. Take a look at this case study about the National Stock Exchange (NSX), in which HP’s server and storage infrastructure solution reduced latency for critical trading applications by 50%.


Your organization may not need to operate at the microsecond levels where the trading firms live or die, but chances are you need to keep an eye on latency. Location, connection, and equipment affect latency, which in turn affects the success of the HFT firms – and, most likely, the success of your business too.




What’s true in real estate investment holds true for data centers: location is everything, especially for HFT firms. Placing the customer’s trading applications close to financial exchanges reduces latency and provides a much needed edge in trading. That’s why you’ll find a highly competitive market for data center space not in Manhattan, but in New Jersey next to the exchanges, including BATS in Weehawken, the New York Stock Exchange in Mahwah, Direct Edge in Secaucus, and NASDAQ in Carteret.


Proximity hosting, a service offered by colocation providers, is popular with many companies that either can’t afford their own data center or want the ability to interconnect to the exchanges. Proximity hosting can include anything from providing rack space in a facility close to the trading venue, to providing managed services and access to the trading venue gateway. While these sites charge a premium, the low latency aspect is so popular that there are waiting lists for prospective tenants.  The ultimate drive for lowest latency has become competition for a “cage” next to the exchange’s “cage” in the same colocation site.




Many of the colocation providers are data-circuit carrier-neutral; this allows connection to the data center from many telecommunication providers.  These telecommunication companies are owners of fiber-optic cable that serves as the connections from one spot to another, and they compete for business through providing lower latency through the most direct connection. In 2010, Spread Network provided a 13.3 millisecond connection between Chicago and the NY Metro area which shaved 3 milliseconds off their competitor’s route. In 2012, per the company’s press release announcing latency improvements, the “Ultra Low Latency Chicago-New York Dark Fiber service reduced latency to 12.98 milliseconds round trip.”


Global network providers are competing as well; as I mentioned in my blog on Iceland, Emerald Networks’ submarine cable “Emerald Express” is connecting Shirley, New York to Grindavik, Iceland.  According to their website, Hibernia Networks, which provides networks that are purpose-built for the financial community, has launched “Project Express,” which will connect New York and London in under 60 milliseconds. These network providers market latency tables, showing the expected round trip time in microseconds.




IT hardware providers compete in providing faster performance, thereby decreasing latency. Low-latency system performance is affected by the design and configuration of the hardware and software.  For financial intermediaries, the name of the game is speed to market, and latency is the driver.  Reducing latency to obtain and respond to data can equate to millions. Vendors that provide hosting, equipment, and network connection continue to advance and find ways to shave off milliseconds of latency in order to help these firms stay competitive in their field.


HP knows this territory well. Whatever industry you’re in, if you’re looking to improve your latency metrics give us a call.


Laura Cunningham is a CPA and business consultant with HP Technology Consulting. She helps CIOs and their team bridge the gap between what the CIO wants and what the CFO requires by building a comprehensive business case that can withstand financial scrutiny.

0 Kudos
About the Author

28-30 November
Madrid, Spain
Discover 2017 Madrid
Join us for Hewlett Packard Enterprise Discover 2017 Madrid, taking place 28-30 November at the Feria de Madrid Convention Center
Read more
HPE at Worldwide IT Conferences and Events -  2017
Learn about IT conferences and events  where Hewlett Packard Enterprise has a presence
Read more
View all