The Inevitable Convergence of Middleware into the Network
Today in New York our CEO Craig Betts co-keynoted the High Performance on Wall Street conference with Andy Bechtolsheim and a representative of Barclays Capital. The theme was something we talk about a lot at Solace: the convergence of middleware and the network. The need to effectively route an increasingly massive volume of data at high speed is defining competitive advantage in more and more situations. The only way we’ll ever address this trend is by making the intelligent routing of messages a robust and readily-accessible service of the network, just like the routing of IP packets. We can only shove so many servers into the increasing number of holes in the dike before the water pours through.
Craig talked about the 4 key business drivers of IT innovation today (the need for speed, the need to save, the desire to be more environmentally sensitive and the need to stay productive through M&A activity) and how they all lead down this path of convergence. Mr. Bechtolsheim dove a little deeper into the specifics of network architecture and operations to illustrate the point that raw computing power isn’t the way to crack this nut. Finally, Barclays Capitals’ head of middleware talked about the business value of consistent, predictable performance which drove them to select a (ok, our…) hardware-based middleware platform.
From business drivers, to nuts and bolts, to the bottom line, the presentation made it clear that the migration from old-fashioned software-based infrastructure to hardware is a fundamental corner we must turn.
Solace CEO to keynote HPoWS with Andy Bechtolsheim
This week the conference organizers of High Performance on Wall Street announced that Solace’s CEO Craig Betts will be joining Andy Bechtolsheim — a founder of Sun Microsystems and countless other successful startups, most recently Arista Networks — for a morning keynote at the event. They will speak on the role of technology in the ever-changing world of financial services. The conference agenda is available online here.
High frequency trading comes under fire by NY Times
Late last week the NY Times printed an article on how Goldman and others are using what they’re calling “high-frequency trading” to get ahead of the market and generate profits.
This is not new news to folks on Wall St, nor is it illegal. As opposed to “front running” scandals of the past where trading firms would use inside information to profit from coming news, all of these advantages are achieved after information has been made public. They just get to see it before everyone else by having faster infrastructure, even if it is only a few milliseconds, or microseconds faster. This is like an astronomer getting a better view of planetary objects by having a better telescope. You dramatically increase your likelihood of discovering something good up there if you get a telescope as good as the US government scientists.
Recently, Ralph Frankel, Solace’s CTO of financial services, had an article published on this topic in Securities Industry News. If you’re interested in this topic, it goes into much more depth about how latency arbitrage in high-frequency trading actually works.
Espionage on Wall St
Last year, Tabb Group famously researched the value of a millisecond in an automated trading environment. I wonder what the value of the secret trading strategies from one of the industry’s benchmark leaders is?
A Reuters blogger is reporting that a disgruntled former employee (Sergey Aleynikov) appears to have stolen Goldman’s program trading source code and uploaded it to a server somewhere in Germany. He was arrested June 4th and is being held on charges of “theft of trade secrets”. It’s a scene more out of James Bond than Jersey City, but it appears to be for real.
Goldman’s program trading platform is a crown jewel within the massively profitable firm. It will be interesting to see if any more information is made public, if there is fallout from their trade secrets possibly being for sale, and whether they will change their trading strategies.
Cache in on your message stream
On Monday we announced a noteworthy addition to our product offering—fully-integrated in memory data caching. It’s intended for applications or services where clients need to replay historical data or lookup “last values” by topic, whether in context of market data or a sensor network. Why bother to build a cache when there are already so many to choose from? Because our customers asked us to. For many applications having a cache that is ultra-fast and pre-integrated is better than having one that one that can slice bread and read your palm if you just want it to store and retrieve messages.
That’s not to underplay its capabilities though – when it comes to message caching its performance, scalability and robustness are second to none. It can cache half a million messages a second, can be distributed for extreme performance and scalability, and it lets you decide how to handle exceptional situations such as when data is updated while a cache request is in progress.
Our solution is not, however, a full-featured in-memory application server like what you’d get from GigaSpaces or Oracle Coherence. Think of it like TiVo for your message stream—it records shows as they are broadcast (caches messages as they’re sent), and then lets you watch (retrieve) them later. Only really fast. TiVo is a special purpose Linux box that has been optimized to excel at, and seriously simplify, one function: recording and playing back TV shows. You can’t back up your PC’s files to it, you can’t edit the shows you’ve recorded, you can’t (easily) move the files from your TiVo to other computers, etc.
Many of our customers will already have a product like Gigaspaces or Coherence and in those cases, we will continue to integrate with these 3rd party products. But for customers that want a fully integrated, one-vendor solution, the Solace cache may well be the ticket.
SIFMA time. Good, I’m running low on pens.
In the financial services industry, SIFMA’s annual Technology Management Conference & Exhibit (this year from June 23-25) is the biggest trade event of the year. Anyone who’s anyone comes by for at least a day to see what’s new, network, and find out where the friends they haven’t seen in a year are working.
If you’re coming to SIFMA, please stop by and see us. We’ve been hard at work on some new technologies, and will be announcing some new partnerships as well.
You can find us in booth #1745 on the Rhinelander (main) show floor.
Market data rates march onward and upward
With news of bankruptcies and bailouts dominating the headlines and sizable layoffs leading to half-empty trading floors, it’s easy to think of Wall Street as being in a holding pattern—tightening its collective belt and holding on until the economy rebounds.
But don’t talk to the IT folks supporting trading about an “economic slowdown” – the global economic crisis is driving the rapid acceleration of market data rates. The April 2009 market data stats from FIF showed new record peaks for eight of the top exchange feeds, some of them up as much as 44%! That continues to put pressure on already overworked market data infrastructures.
Despite the misperception that leading capital markets firms are mired in budget cuts and freezes, these companies have no choice but to spend to stay competitive in their infrastructure. Without timely market data they are dead in the water and might as well turn the lights out.
Staff cuts and outsourcing are for real, and paychecks aren’t what they used to be, but this report from Aite confirms that capital markets IT budgets aren’t down as much as people think. Late last year the analyst firm Aite predicted that Wall Street IT spending wouldn’t drop more than 5% in 2009, and halfway through the year they’ve only tweaked that prediction to 6%.
This business reality is what is driving the move from software- to hardware-based messaging in earnest. Capital markets firms have been moving away from the antiquated model of building their infrastructure internally for years, but the availability of easily-deployable feed handlers and messaging appliances that outperform COTS and homegrown software by orders of magnitude is making it a competitive necessity. An investment bank building their own custom hardware would be way out of scope even for the world’s largest firms, and even in the best of times.
We’re pleased to count technology leaders like Barclays Capital among the list of financial firms that are leading this transition.
Will you and 10GigE live happily ever after?
Over the past few decades we’ve seen network technology evolve from 10 Mbit Ethernet to 10 Gbit, with 40 GigE on the horizon and 100 GigE already being discussed. Is the network outpacing the capabilities of the computer? In many ways, yes.
One of the dirty little secrets of today’s operating systems is that even in multi-CPU hosts all the interrupts from the network have to be processed on a single CPU. This means a single 1 GigE card can overwhelm even the fastest commercial processors available today. So what will 10x the throughput do for your servers if the CPU/OS is the bottleneck? In many cases, the answer is not very much. I think many administrators and architects may be prematurely falling in love with 10GigE when it won’t really solve as many of their problems as they hope that it will.
Don’t get me wrong, there are lots of applications and environments out there that will benefit greatly from 10GigE. And there’s a whole slew of cool new applications and services and capabilities that it will make possible. I just think it’s important that people recognize that 10GigE isn’t a silver bullet that will open up a new world of performance in every situation.
Read more
You think you’ve got multicast storms now? Just wait for 10 GigE!
Back in the early days of market data systems, 10 MbE networks were the standard. Multicast from the publisher to the subscriber was a clever way to optimize bandwidth to trading desks, because there just wasn’t enough bandwidth to send the data repetitively to each trader. The emergence of 100 MB and 1 GigE networks resolved some of the bandwidth issues, but the delivery bottlenecks just shifted to the software middleware, so multicast lives on.
But multicast has a dark side. If you want to see any Wall Street or internet infrastructure architect get worked up, ask them about multicast storms. Multicast storms happen when application participants request retransmits of information they have missed in the multicast stream. There are two common causes of multicast storms:
- Applications that fall behind in their rate of consumption of messages
- Network speed mismatches in the underlying network.
As market data rates accelerate and trade volumes go through the roof, many people are counting on 10 GigE to bail them out just as things get really hairy. Unfortunately, the migration to 10 GigE won’t be immediate, and it won’t be universal, so the emergence of 10 GigE will actually exacerbate the second common cause of multicast storms. In context of those increasing market data rates and trade volumes, some predict a “perfect storm” in the world of trading systems.
Read more
10GigE enables world’s fastest real world market data benchmarks
In Nov 2008, we released a set of benchmarks for ultra-low latency market data distribution. The tests, run on an all-hardware architecture consisting of 1 GigE technology from Solace, Arista and NetEffect, showed the fastest real-world messaging benchmark performance ever seen — until today.
In conjunction with the launch of the new 10 GigE version of our Network Acceleration Blade, we repeated these tests in a 10 GigE environment. In addition to a 10 GigE-equipped Solace 3260 Content Router we used NetEffect 10 GigE adapters to perform TCP offload in the clients and servers, with an Arista 10 GigE cut-through switch as the layer 2 switch.
To provide an apples-to-apples comparison with the previous benchmarks, we reran the test that simulated all US equities traffic (500K msgs/sec) and the test that simulated peak OPRA market data (1 million msgs/sec.) Then we added to the mix a 2M msgs/sec test to show how performance changes as data rates continue to go up.
Once again, the performance was superb, and the latency distribution nice and tight. The system demonstrated latency 20-25% lower than the previous 1 GigE numbers and improve upon our already best-in-class performance. Here’s a summary of the results: