Mina Down
Traders Magazine Online News

Top 10 Altcoins for 2019

In this shared article author Mina Downs gives her projections as to what will be the top "altcoins" in 2019.

Traders Poll

As the SEC Tick Pilot is soon to expire, in your opinion, was it a solid fact finding mission or folly?

Free Site Registration

January 1, 2005

Sales and Trading for Nomura in U.S.

By Peter Chapman

To that end, Nomura's system is constantly evaluating market conditions and looking for opportunities. If it finds none, it may not trade at all. "Automated opportunism" is the Nomura philosophy.

Comerford's team spent much of its first year building an infrastructure. It needed to capture and store market data; take in customer orders; transmit orders and cancellations; and deal with reports. Connecting to the markets was made easier by piggybacking on the pipes used by Nomura's index arbitrage desk. A FIX engine from NYFIX/Javelin allows Nomura to receive orders and send reports to customers.

Much of the work was spent in gathering market data and feeding it into analytic applications. For starters, Nomura chose to take in direct feeds from the various market centers rather than use an intermediary such as Reuters or Bloomberg. That holds down latency.

The team also decided to forego the use of an off-the-shelf messaging infrastructure, such as Tibco Rendezvous, to shuttle the market data internally. Nomura built its own API, its own generalized direct port messaging system. Again, the idea was to reduce latency.

The time saved, according to Comerford, can range anywhere from 50 milliseconds to one-and-a-half seconds. Abenante adds: "You're trafficking in milliseconds. A ten-millisecond delay in four parts of your architecture could translate into a missed trade opportunity."

Nomura does use Sonic, a Rendezvous rival, for some applications where messaging with guaranteed delivery is necessary. For instance, Sonic transmits data between Nomura's FIX engine and its Portware trading system.

For storage, Nomura customized a tick database from Vhayu Technologies. A tick database, or ticker plant, is software that captures real-time market data and stores it in a central place where it can be combined with (typically) historic data for manipulation. Comerford, who sits on the Vhayu board, describes the technology as a framework that makes it easy for Nomura to run analytics. Nomura has written programs that sit inside the Vhayu memory to analyze the data.

The analytics are a critical part of Nomura's trading set-up. They provide the alerts and signals that lead to trading decisions. The application layer is called RETA, for Real-Time Alerts. RETA is also known informally as the Trader Intelligence Platform. From a screen on their desktops, Nomura's traders can see into the workings of the market.

In general, RETA sifts the market for key trading triggers such as large quotes, narrow (effective) spreads or both. It looks to see if any reserve orders have gone off. [See sidebar]. Once RETA spies an opportunity, it notifies Nomura's trading logic. The software must determine how much, if anything, is in the best interest of the customer's overall strategy to trade. At Nomura, that's known as "scheduling." For Nomura, scheduling determines how much to trade, not when to trade.

The logic supporting Nomura's platform is structured in three layers. The top layer is termed the aggregation expert.' It is concerned only with the integrity of the portfolio. The aggregation expert sets constraints such as the minimum or maximum amount of cash the portfolio can hold at any point during the day.