Latency – What does it mean for your infrastructure and business
What matters in networking is the speed at which data is sent or received. In a dynamic business environment, where speed is the essence, mission critical systems and applications need greater bandwidth, throughput and capacity. While we can see examples of this in data intensive sectors like trading, stockbroking, capital markets, and media, the other industry sectors are not far behind and can definitely benefit from improved wide-area network performance. This holds true in the case where the WAN connects multiple sites or links the man office to the data center. In situations such as these, a fast and reliable network link not only ensures overall system performance but also assures quick recovery in an exigency that may require failover.
Closer Links For Better Connectivity
The financial sector and in particular brokerages and trading/stock exchanges, were the first to adopt ultra high-speed networking. In a trading/stock exchange, speed to transact is what matters and a few milliseconds can make all the difference. Over the past decade brokerages and stock exchanges have spent millions in networks infra i.e., faster servers and Gigabit Ethernet infrastructure to score an edge over their competitors.
The TABB Group, a financial consultancy, recently surveyed CTO’s at several financial firms, and found that reducing latency and expanding network were some of their top priorities. Other key areas for consideration included improving the use of data center space and colocation. The report also found that proximity to the local exchange (48%), space for expansion (33%), reliable power supply (29%) and cost (57%) were some of the key drivers behind the choice of a data center by brokerage and trading/stock exchanges.
The findings reflect the sentiment among most CTOs in other industry verticals as well. Key priority areas include transparent network, performance, lower cost, scalable and reliable data center hosting options.
Colocation As An Option
Reducing the distance between two sites has often been employed as a measure by organizations to improve network performance between them. This is probably the reason why brokerages and exchanges look for data centers close to financial centers and in some cases, within the stock exchanges.
A case in point here is Thomson Reuters. In 2010, Thomson Reuters launched Elektron, a real-time network and hosting environment, to meet strong demand for fast and cost-effective access to the emerging cash equities market in India. Elektron enables hedge funds, asset managers, banks, brokerage houses, exchanges and other participants to connect to the world’s largest financial community and securely reach trading partners over the network globally.
This global service needed an ultra-high speed, low latency network and hosting environment that was carrier neutral, and had proximity to Bombay Stock Exchange (BSE) to help financial firms access and share information faster and cost effectively. Being the first company to launch such a service in India, Thomson Reuters needed a highly available IT environment for Elektron. In a fast paced trading environment where decisions are made on a fly, a few seconds of down time could result in potential monetary loss to its clients and customers. Hence, Thomson Reuters chose Netmagic’s datacenter located within the Bombay Stock Exchange (BSE) to host Elektron. Netmagic’s colocation hosting facility at Bombay Stock Exchange (BSE) ensured low latency and enabled high frequency trading for Thomson Reuters’ customers.
Being hosted close to/within data centers in financial centers, though a desirable option for most financial institutions creates shortage of space, leads to a strain on power supply and increases costs over a long term. With limited space and resources available within these facilities, colocating to a data center close to financial centers puts a strain on these facilities. Making use of newer and faster networks can solve this problem of overcrowding and also lead to enhanced performance.
Enhancing Performance, Reducing Latency With High Speed Networks
40/100 GB networking standard was introduced in 2010 and the first roll out of the 40/100 GB infrastructures is currently going on. This is an advancement over the conventional Gigabit and 10GB network links that reduce the network latency to less than 2msec even over distances of 100 miles or more.
Using a higher speed 40/100 GB infrastructures has its own advantages. The first one is reducing latency, which essentially means faster and error free data processing – one of the key requirements of the financial sector.
Secondly, a high network speed is able to support recurrent, periodic bursts of data generated by data intensive applications. If the rate at which the data is generated exceeds the capacity of the network link to accommodate the excess at any given point of time, queuing of data happens leading to unwanted delays and system crashes. A higher speed 40/100 GB link ensures that the network capacity always exceeds the maximum data burst rate and the network remains congestion free assuring seamless flow of data without any delays.
Thirdly, a high-speed infrastructure also helps financial institutions to spread their data center footprint rather than limiting themselves to locations closer to stock exchanges or their head offices.
Colocation It Is
A high-speed network infrastructure reduces latency; hence, proximity of data center does not matter. Organizations can benefit from lower costs for space and power while still being able to ensure performance for data intensive, mission critical applications.
The Jones Lang LaSalle Data Centre Barometer of Autumn 2013 found that power continues to be the most important factor in choosing a new data center, with over half of respondents ranking availability of power as their absolute priority.
A low latency data center interconnect can significantly improve international data traffic and organizations can benefit from shortened data transit times and improved network performance and robustness. This in turn leads to increased performance, reduced costs, a wider range of colocation options, a scalable infrastructure and efficient use of power for growth.
Nilesh Rane is the Associate Vice President - Product and Service at Netmagic Solutions. Nilesh is an expert in the data center domain, specifically in areas such as Disaster Recovery, DR-as-a-Service, IDC and Bandwidth. He has extensive experience of over 10 years within the data center domain, out of his total work experience of 20 years. Nilesh has been with Netmagic for 6 years handing key roles and responsibilities within areas such as DR, DRaaS, IDC and Bandwidth.