Datafeed Speed and Market Structure

If you absolutely must have trading data fast, who’s your huckleberry?

Burstream, apparently. The firm claims it can serve up actionable, meaningful trading data, no matter what market mayhem, in 600 nanoseconds. That’s 600 billionths of a second. The catch? You have to trade at the Nasdaq.

Burstream’s system is being installed at the Nasdaq OMX market center in New Jersey so the exchange’s important proprietary-trading customers will have a split-second – taken to the extreme – advantage. Customers wanting to use these superfast capabilities will be able to load their algorithms onto Burstream servers parked next to boxes housing the Nasdaq’s trade-matching engines.

Burstream systems will go near the Chicago Mercantile Exchange too. The idea is to unify data streams on stocks, commodities and derivatives so decisions about trading on divergence can be made faster than ever before possible. This is, of course, arbitrage.

Burstream says at its website: “Enable your high frequency trading algorithms to hit liquidity when it is revealed. Trade through market bursts while competitors exit the market. Sustained nanosecond speeds, even during message bursts will give your latency-sensitive algorithms a performance advantage.”

What makes Burstream special is its use of field-programmable gate-array chips (FPGAs) that can perform multiple calculations simultaneously, thus delivering a speed advantage over conventional hardware-processing techniques.

We have clients in the FPGA design and manufacturing businesses whose shares are publicly traded on the Nasdaq. IR pros may appreciate the irony. Companies who use the public markets to raise capital and build growth enterprises are being arbitraged via nanosecond data at the exchange that lists their shares.

We don’t fault the exchange per se. Exchanges are for-profit businesses today. They earn money by capturing a share of the quote and trade data on the consolidated tape. The amount is determined by how often securities at their market centers (most have more than one) quote or trade close to the National Best Bid or Offer. Thus, liquidity providers like Merrill Lynch, Instinet and Morgan Stanley are paid to quote at the Nasdaq. Payments come in the form of trading rebates for posting shares at the Nasdaq that offset against fees for removing them.

If the Nasdaq can speed up trading and arbitrage, more data are generated, more trades and quotes. It’s a virtuous circle: pay liquidity providers like Merrrill and Morgan to make sure liquidity exists. Encourage arbitrage with the fastest data systems to hit it. Capture more quotes and trades. Monetize consolidated-tape data. Repeat.

This is how exchanges generally work today. While rational investment is one behavior, it’s small today (and it does not re-price shares in nanoseconds). For obvious reasons. If you think about it.