Low Latency Trade-Order and Market Data Considerations for the Financial Marketplace

Unprecedented growth in electronic trading is redefining system boundaries and accentuating the need for increased data transmission speeds. The scope of this document is to provide insight for addressing the latency challenge.

FINANCIAL MARKETPLACE PERSPECTIVE

One of the most pressing issues impacting financial industry firms is how to increase speed and reduce latency when accessing market data and trade execution venues. Growth in the size and quantity of orders is requiring that both the buy and sell sides develop solutions that will increase data speeds and capacity. This is necessary to realize faster and better trade execution capabilities in the highly competitive securities business.

What marketplace circumstances are driving systems technologists to satisfy an increased need for speed? Buy side firms continue to gravitate to rules based systems and brokers are using improved models for analyzing market data for subsequent trade execution. In addition, new investment vehicles such as hedge funds, with heavy reliance on algorithmic trading, make the need for speed an industry priority.

Algorithmic Trading

Rules-based trade execution is based on mathematical modeling. These models provide analysis on quotes and trades and identify liquidity opportunities that result in intelligent automated trading. This “black box” trading activity relies on the quality and timeliness of data fed into systems. Opportunities can be lost when conditions have changed before a trader has had the ability to execute an order, or, when someone else was quicker to seize the opportunity. To quote an old adage, “time is money”.

Quality Market Data

As automated algorithmic trading demands continue to grow, market data quality is also imperative in order to avoid poor trade executions. You cannot trade quality for speed. You need both. Data consolidation service providers can serve an important role by providing mechanisms to flag and correct data discrepancies. So what are firms doing to address the requirement for data integrity without losing speed? Many firms are deploying dedicated data feed handlers on their premises to deliver raw feeds to traders via optimized reduced latency network systems. Such transitions to direct raw feeds can ultimately serve to reduce latency, in some cases, by several hundred milliseconds which can be significant to the profitability of the business. Vendors are packaging direct lower latency services with significant valued add features. In concert with the push for quality is the push and adoption of sophisticated direct market services.

Direct Market Access

Buy side firms with access to a myriad of available best-of-breed algorithms are positioned to trade directly in the market. They want to gain a competitive advantage by executing trades more quickly. Direct Market Access (DMA) provides buy side traders with improved liquidity by providing access to multiple trade execution venues. DMA enables institutional traders to use broker infrastructure and clearing services thereby reducing their costs on a per traded share basis. Major broker dealers have acquired DMA vendors and as such have extended their services beyond equities into fixed income and derivatives. The growth in DMA is expected to further expand to the international arena.

TECHNICAL CONSIDERATIONS & CHALLENGES

What are some of the technical latency challenges faced by systems developers and network engineers?

Latency Variables

Latency is the amount of time it takes for a response to return from a request. It is measured in a time value, typically milliseconds. As a reference, note that 1,000 milliseconds equals 1 second.

From a systems perspective end-to-end latency incorporates both the host system server and the network (LAN & WAN). Latency will vary depending on characteristics such as message size, volume of messages in and out of the servers, traffic volumes and network speed, the physical distance between hosts and the efficiency of network drivers on the various hosts.

IP network latency is generally caused by four delay categories: queuing delay, switching delay, propagation delay and serialization delays.

• Queuing delays reflect data bottlenecks and relate to the traffic profile. Bursty applications and resulting data spikes impact this delay. In addition, if the connection is not sized appropriately data packets will begin to get backlogged until a queue is cleared.
• Switching delays are related to time delays associated with moving the data packets through network gear-routers, hubs & switches. Fewer network “hops” result in reduced latency.
• Propagation delays reflect timing delays for the data signals to traverse the telecommunications footprint. In general .7ms of propagation delay for every 100 km can be used to approximate this latency component.
• Serialization delays relate to the time it takes to move the data into a circuit. Typically the serialization delays get smaller as the bandwidth of the circuit increases. As an approximation, these delays can range from approximately 3ms (for circuits at 1.5Mbps) down to .05 ms (for circuits at 100 Mbps).

As the distance between the end points increases, the latency levels can also increase. This is explained purely by the laws of physics relative to the speed of light.

Tools are needed that can accurately measure latency, while uncovering opportunities for latency reduction. They must measure both network and application level aspects of the transaction. The latency measuring system should also include: a tap into the network, robust analytical functions and a dashboard to display events and monitor performance. This is a key challenge for most firms.

Latency Reduction Methods

1. Optimize the market data and trading system networks for reduced latency. Ethernet services ranging from 100 Mbps to Gig bps and 10Gig bps network fabrics provide for relatively cost effective transport while providing for higher bandwidth and lower latency capabilities. High industry market data volumes require high capacity transport. These requirements apply to local area networks (LAN), metropolitan area networks (MAN) and wide area networks (WAN) configuration environments. Communications carriers and extranet vendors serving the industry do offer such high capacity transport services.
2. React quickly to market changes by getting closer to market data sources and exchanges. Consider building direct market service feeds which can save anywhere from 100-200 milliseconds. Consider hosting trading systems with service companies that offer reduced “hop” network services coupled with high service levels will give firms significant improvement in latency reduction.
3. Utilize efficient and low speed messaging platforms and middleware products. Deploy strategies and tools to accurately measure latency. Some off the shelf stream processing engines handle high volumes of real time data without custom code.

CONCLUSION

Milliseconds count. Whether you are a buy side or a sell side firm, latency will affect your competitive position in the marketplace when it comes to executing trades. As such, the issue of reducing latency must be priority as the need for speed continues to grow in an expanding marketplace.

ABOUT SECTOR, INC.®

Sector provides high value managed services, connectivity and market data solutions that help increase the effectiveness of the financial services industry. With Sector, you benefit from strategic thinking, ultra low latency, advanced technology and top-quality service honed by meeting the exacting demands of the industry for over 35 years. Our services utilize the strengths of our parent company SIAC®– the technology subsidiary of the NYSE GroupSM with responsibility for the design, development, implementation and operation of the exchanges’ computer systems and communications networks.

Sector’s services operate out of the highly secure, redundant and resilient SIAC data centers that are among the very best in the industry. Our Managed Services offerings include co-location, facilities management, hosting, email and IM archival, Business Continuity plan development and review, Sector Contingency Web Site, relocation management, network/systems design, extranet and backup storage. Telecommunications solutions include voice and data lines, SFTI® and SFTI B2BSM, ARDs, SectorNet® financial services extranet, Ethernet, frame relay, ISDN, long distance, teleconferencing, Internet, VPNs and more. Market Data services include mortgage-backed securities, MBS pricing & analytics, a corporate and agency database, market data, ratings, corporate actions, tax reporting and more.

More information can be found at www.sectorinc.com. TRUST THE STRENGTH THE MARKETS TRUST®
This White Paper was written by Vincent Lanzillo, Vice President at Sector. He is responsible for Technical Services and has over 12 years experience in the financial industry working in various technical, business and sales development, and project management roles.

Disclaimer:

Please note that the information on these pages does not constitute financial, technical or other regulated advice. Any statements referring to these subjects are to be considered opinion only. The information presented on these pages is intended to provide a general overview. It is recommended that before making any decisions based on the information provided, you should take independent financial, legal or other advice from duly appointed advisors.