Hardware trading: The machines are taking over…
With each day that passes it becomes more obvious that hardware is replacing software in the critical path of high performance trading. It doesn’t take a genius to see what’s happening, as the evidence is everywhere:
- Feed handlers: All of the recent entrants into the feed handler market are using custom hardware.
- Messaging middleware: Appliances from companies like Solace are winning a larger footprint for market data delivery and back-office order routing.
- Algo engines: Leading algo trading engines are heavily leveraging FPGAs and GPUs to do ultra-fast simulations that accelerate trading decisions.
- TCP offload engines: TOEs are taking the place of software TCP stacks on servers.
- Trading apps: We’ve even seen a flurry of requests from high performance trading firms looking to execute pre-trade risk checking directly in network processors inline with the hardware messaging bus.
How latency is like a sore knee
As a running form instructor, I’ve seen many examples of how poor form leads directly to a sore or injured body. Poor alignment causes fatigue. Landing on your heels causes knee pain. And don’t get me started on how most people run up hills. Little inefficiencies don’t stop you in your tracks, but at the end of the day they can keep you from achieving your peak performance.
It’s a lot like having many little sources of latency in your trading system. Each weak link in the process keeps you from reaching optimal performance and profitability.
The best way to help people see what is wrong with their running form is to use video analysis and isolate in on alignment, landing points, hip rotation, arm swing, etc.
Finding and fixing the sources of latency in trading systems takes a keen eye and sophisticated tools, too. Today we announced a partnership with TS-Associates, a leading supplier of latency monitoring solutions. They’ve added support for Solace message routers in their TipOff product, which is itself a hardware appliance dedicated to non-intrusive latency monitoring. In what may be the logical opposite of slow-motion video analysis, their solution studies end-to-end traffic in your trading environment and gives you details on which components in your system represent opportunities to cut out latency.
We’re bringing sexy (to the) back (office)
For the last few decades, from the early days of Gordon Gekko and Liar’s Poker to today’s focus on hedge funds and high frequency trading, the sex appeal has clearly centered around the front office. The press loves to write about it and we love to read about it. Every trade is like the proverbial iceberg, though — the trading decision is what everyone sees, but the majority of work happens in the murky waters below the surface where the considerably less sexy mid-and-back-office operations occur.
Last week Greg MacSweeney wrote a good article in Wall Street and Technology highlighting the back office as a new battleground for efficiency. After years of chasing zero latency for the front-office, the back office is comparatively archaic and badly in need of updating.
Come See us at TCIP 2010
For the next few days our government team will be immersed in the world of emergency messaging and alerting at the Technologies for Critical Incident Preparedness conference in Philadelphia. We’ll be in booth #505, so please stop by if you’re in the neighborhood.
We look forward to discussing the announcements we made last week (DHS/DNDO as a customer, Thermo Fisher as a partner, and our new geospatial routing capability) and seeing what else is hot in the world of information exchange and critical incident preparedness.
Making sense of sensor networks
Sensor Networks are Coming of Age
Sensor networks have been with us for years, but the combination of ubiquitous wireless networks, higher bandwidth, cheap storage, improved battery life and solar power are driving more and more applications towards data collection using sensors of one kind or another. These systems get very complex very quickly. Just consider:
- The sensors are usually distributed and heterogeneous
- Aggregate sensor data production rates are sky high (number of sensors times sample rate per sensor) particularly when combined with images and video
- The sensors can be fixed or mobile
- The people applications interested in the meaning of the sensor data can be fixed or mobile
- What is deemed important within the sensor data is fluid and constantly changing
How do you make sense of a massive stream of information, distributed across a large geography where the relationships between sensors and their surroundings can literally be changing minute to minute?
The world’s only hardware Market Data Factory
Today Solace announced a partnership with BCC Group International, headquartered in Frankfurt, Germany. BCC has selected Solace’s message routing technology to power their Market Data Factory, a powerful solution that offers their customers greater independence from market data suppliers and lowers risk. As part of the agreement, BCC will also resell Solace equipment to European customers of their Market Data Factory.
We are delighted to add BCC Group to our growing list of partners.
Is green IT flowing down the same river as water conservation?
The push to conserve power is getting serious. Compute sprawl is out of control and some power companies are mandating geographical maximums. Some forward-looking companies we’ve talked to are tying IT executive bonuses to power reduction targets to make sure managers aren’t just paying lip service to the issue. Regulators are even getting in on the act, creating new pressures as shown by this article in Datacenter Dynamics about new taxes being attached to co-location facilities.
There may be some lessons to be learned from other areas of conservation that have been driving towards efficiency longer. There was a piece on NPR this morning about water consumption in which the US geological service released stats showing that per capita water consumption is down thirty percent over the past 30 years.
This dramatic reduction is primarily the result of improved efficiencies in industrial and farming techniques. For example, in the 1930s it took 200 tons of water to make 1 ton of steel. Today’s best steel plants use just 3-4 tons. Similarly, the evolution from flood irrigation to sprinkler irrigation and more recently to precision drip irrigation lets farmers produce more food with less water. Both of these moves to superior techniques and technologies were driven by the scarcity (and thus increasing cost) of water, along with a regulatory push in the 1980′s that established strict standards for waste water discharge. One of the least costly ways for companies to comply with those standards was to reduce the amount of water they used for their industrial and agricultural processes.
The inherent goal of business is profit, and public companies in particular have a responsibility to deliver maximum shareholder value. This means they can only afford to get really serious about “going green” when they are incented or required to do so, whether that’s in the form of market demand, increasing cost of resources, or regulation. That’s what happened with water, and it’s where we’re at with energy.
Exponential increases in market data rates, trade volumes and commerce conducted over the Internet are already causing datacenter sprawl that’s driving some insane energy requirements to power and cool the acres of servers they hold. So in addition to issues of scarcity and regulation, companies are already finding themselves highly incented to find technologies that can do more for their datacenter (whether that be in terms of storage, compute power or message routing) with a smaller “carbon footprint.”
As we continue to buy more gadgets to charge and eventually cars to plug in, will it be profitability pressures and regulations on industry that drives energy efficiency into the mainstream? I think the answer is clear.
An emergency response demo that was a real train wreck
Standards body meetings aren’t usually all that exciting, but last week’s National Information Exchange Model (NIEM) Training Event and OASIS Emergency Management Interop was an exception. The centerpiece of the Interop was a demo designed to show how NIEM, Common Alerting Protocol (CAP) and Emergency Data Exchange Language (EDXL) standards could help a dozen government agencies share information and coordinate activities in a set of simulated emergency situations. The three scenarios were a train crash/chemical spill, a tornado warning, and an Amber Alert. Not your typical dry standards body fare!
Read more
High frequency trading spotted drinking Cristal in the VIP room
High frequency trading (HFT) is the media’s new “it girl”, and is generating attention everywhere she goes.
First Forbes magazine did a cover story on The New Masters of Wall Street, highlighting the hedge funds, prop traders and big banks that are turning HFT into big business.
Then Traders Magazine did a related story about how the exchanges are responding in their Race to Zero Latency. Clearly the obsession with execution latency is being driven by the need to keep up with these ultra-fast consumers of liquidity.
And you know HFT is nearing its hype peak because last week the Daily Show did a special report on the subject.
Unlike Paris Hilton though, HFT is not making money because it is hot, it is hot because it is making money. Generating buckets of profits is one thing that never goes out of style.
Trading microseconds for nanoseconds
The co-location of market data systems near or inside exchanges is becoming big business. The ultra-low latency high frequency trading systems that you find in these facilities are niche applications to be sure, but what a niche! NYSE Euronext recently committed to build a 400,000 square foot co-location facility in New Jersey. That’s a big investment to make in something NYSE Euronext CEO Steve Rubinow describes as being for “only the most obsessive traders.”
How obsessive? Architects building these systems measure latency in microseconds, and the best applications exhibit just tens of microseconds of end-to-end latency. Shaving microseconds is like dropping weight before your prize fight weigh-in—whatever it takes, get it down.
To help these latency obsessed traders develop even faster trading systems, Solace has extended its Unified Messaging API to include a shared-memory transport based on inter-process communication (IPC). This capability lets two applications share information using Solace’s API with less than 700 nanoseconds of average latency in a shared memory environment. Yes, I said nano — billionths of a second. Remember the famous Tabb Report on The Value of a Millisecond? There are a million nanoseconds in a millisecond. 700 nanoseconds is a scant seven-tenths of a microsecond.
To be clear, IPC is a highly-specialized technique that only certain systems can leverage because it occurs within the confines of a single server. For example, when the components of a high-frequency trading system (feed handler, algo, risk assessment, order execution) have been consolidated onto a high-powered multi-core server within a collocation facility. Today these applications run on many machines and share data using low latency messaging (like Solace’s). Shared memory transport among applications running on a single server eliminates the few microseconds associated with network hops and additional time lags associated with copying memory around between applications. And since IPC is now available as part of the same API customers already use for ultra low latency and other kinds of messaging, applications get the speed they need without giving up the familiar API or the flexibility to redeploy in a networked scenario as needed.
As always, we’re not publishing some mysterious single number with no detail on what it means. A white paper describing the environment and parameters of the tests is available for download on our website so customers can dig into the facts and even reproduce the results using their own systems and data. In fact, we did all the testing a quad-core 3GHz Intel Xeon E5450 server because not everyone has the latest Intel Nehalem.
HFT architects have generally been exempt from corporate technology standards because the stakes are so high they can justify whatever makes them faster. With Solace, HFT no longer needs to be an exception. The same messaging API that is speeding up back office and front office networked trading can be used to speed up collocated HFT trading as well.