The Securities Information Processor, or SIP, long regarded as a default way of receiving data on Wall Street, is about to become turbocharged. And IEX customers are about to receive IEX data from the SIP 10 times faster than from the market itself.
Over the summer, data latency discussions focused on “speed bumps” as IEX prepared for its August launch as the first US stock exchange with a built-in delay, or speed bump, designed to equalize market access.
A new development coming this fall, however, may turn that market latency debate on its head. That’s because the Securities Information Processor, or SIP, long regarded as a default way of receiving data on Wall Street, is about to become turbocharged. For many trading firms that have invested heavily in systems and technology for receiving data faster than the SIP – and for IEX, whose speed bump artificially introduces old-SIP-like latency into its market – a super-fast SIP will change the playing field.
As it currently stands, the SIP adds about 400 microseconds of processing latency (not counting network latency) to markets. When NASDAQ migrates legacy SIP technology to its INET platform in October, that processing latency will reduce to an estimated 50 microseconds.
Compare that for a moment with IEX, whose trademark speed bumps famously introduce two 350-microsecond delays (one going in, and one coming out of the exchange) as a way of ensuring that slightly slower market participants won’t continually lose out to their faster peers.
IEX’s idea is that by slowing down market access ever so slightly, information can’t be received about fills immediately, and then routed via high speed networks to other venues and acted on ahead of the entity executing that order.
But as IEX is slowing things down, the SIP is speeding things up.
This is important because, according to SEC filings, IEX reports data to the SIP before its outgoing speed bump. While information exiting IEX will be subject to a 350-microsecond delay, beginning sometime in October it will be available on the SIP in just 50 microseconds. This introduces an unusual situation in which SIP data will actually be significantly more current (in microsecond terms) than data exiting the exchange through any other method.
The upgrade of SIP speed has been a long time in coming. In August 2013, a glitch in the Nasdaq-run SIP caused the exchange to shut down trading for three hours. It was reported that Nasdaq was even considering relinquishing its SIP management, but a year later the Unlisted Trading Privileges (UTP) Operating Committee, which oversees the SIP, awarded Nasdaq a new SIP contract, after approving Nasdaq’s plan to reduce SIP latency by migrating the SIP to its INET platform (INET is the proprietary core technology used across Nasdaq’s global markets).
As I see it, this raises two questions:
What effect will 50-microsecond SIP speeds have on IEX’s speed bumps?
Does Nasdaq’s control of the SIP indirectly control the impact of IEX’s speed bump?
These are issues that we plan to continue assessing. In a market where microsecond differences in access to information – even nanosecond differences – can present arbitrage opportunities, the juxtaposition of speed bumps and faster SIP speeds is not just ironic but potentially impactful.
As the SIP gets even faster – Nasdaq has suggested that SIP latency could be shaved to 25 micorseconds by 2017 – its data will potentially grow more relevant.
The implications on the high-speed trading arms race may be the single most significant issue facing trade execution.
Faster SIP data and slower speed bumps underscore the critical importance of latency analysis as a component of trading strategy.
Sources: Tabb Forum