Well this time it maybe.. The Times Online reports that Britain’s biggest mortgage lender has accidentally made triple withdrawls on their customer’s bank accounts due to a technical error:

Thousands of homeowners have empty bank accounts after HBOS, Britain’s biggest mortgage lender, repeatedly withdrew home loan repayments from their current accounts. About 7,000 Bank of Scotland mortgage customers have had up to three mortgage payments taken from their current accounts on consecutive days this week.

Well.. uhm, isn’t this embarrassing. We have no way of knowing what kind of technology is driving this system, but wouldn’t it be quite interesting to find out? Well if you work in IT for finance and ever feel like your efforts are somehow disconnected from daily life, let this be a reminder to you that real people out there, real unsuspecting average people, are counting on you to do the best you can. No really, I’m being serious:

Judith Lomax, 48, an opera singer who has a mortgage with Bank of Scotland, said: “I have so far paid my mortgage three times this week. I was told that the payment may be withdrawn again today and tomorrow.

See what I mean?

28 Comments

An interesting revelation (to anyone not working or following the internal development habbits over at JP Morgan) is that the bank is using Eclipse and the RCP (Rich Client Platform) technology to build their market data, pricing and research workstation terminals.

JPMorgan developers selected Eclipse largely for its RCP (Rich Client Platform) technology, which enabled the international financial powerhouse to build an internal development system known as OneBench.

OneBench is a deployment platform for RCP-based applications, said Bruce Skingle, a London-based distinguished engineer with JPMorgan’s Investment Bank Technology group.

Basically, OneBench is an alternative to a user desktop for deploying the Eclipse IDE (integrated development environment) to run business applications. In essence, OneBench is a user’s workbench, whereas Eclipse is a programmer’s workbench, Skingle said.

This is quite interesting, but even more interesting is that their choice came down to Eclipse and Java’s SWING:

Skingle said JPMorgan looked at several technologies as the basis for OneBench. Eventually, the company narrowed the decision down to the Java Swing GUI components platform and Eclipse’s RCP. Skingle said RCP won out because of its inherent plug-in nature. Now the RCP-based OneBench is viewed as the platform of choice for desktop application developers at JPMorgan.

This is worth paying attention to since any time it comes down to real-time UIs for finance, its often a tradeoff of performance and flexibility.

Yet, RCP stood out in graphical performance, Skingle said. Indeed, a proof of concept based on the Snapper UI showed that an SWT table could easily handle a 400,000-row data set with 1,000 updates per second.

Well I will reserve judgement for now since I have not had any first hand exposure to Eclipse or RCP as of late. That said I am interested in knowing how they performed their test. Was it just one grid? When it comes to UI performance I maintain that nothing should be as capable or performant as .NET’s Windows Forms because it provides the most tightly knit wrapper of the Win32 display system. How would RCP perform with say half a dozen Level II quote windows on the screen, in addition to an order blotter, real-time charting and all of the other typical 3rd party trading apps that run on desks today?

Hmmmm…

1 Comment

With the recent meltdown at Amaranth, it hasn’t taken long for other funds to look towards outside technology to mitigate risk. CMP’s Wall Street & Technology report that Finetix has joined hands with Cadence Capital Group. I’m not sure how much this is joining hands as opposed to Cadence bringing in consultants, but in any case:

“Experience shows that market risk is never completely understood and rarely properly quantified, as most recently evidenced by the debacle at Amaranth,” says Babayev. “In our partnership with Finetix we look forward to leveraging their depth of development experience in the capital markets to build an extensive analytical platform that will enhance our risk management methodology and help us control our exposure to future ‘black swan’ events.”

It seems unlikely that a one-size fits all platform could ever emmerge from these types of relationships, and it also seems unlikely that we’ll ever be able to fully evaluate the success of such relationships. Yet you can’t blame anyone for looking to take extra precaution in the aftermath of such a major financial catastrophe.

1 Comment

One thing that we like to cover here at WSTD is the demographics of financial business sectors. Recently we noted some mainstream media coverage of the trend to open up shop in Greenwich, CT. Now comes news that a major hedge fund services firm has gone a bit further into the Connecticut heartland:

A New York state firm that provides services to hedge funds has leapfrogged over Fairfield County to establish a large support operation in Hartford. In adding 150 jobs in downtown Hartford, GlobeOp Financial Services L.L.C. becomes the first company to take advantage of two job incentive programs that became effective July 1, the Job Creation Tax Credit and the Displace Worker Tax Credit.

In so doing, the company passed up on situating the facility close to Greenwich, whose large hedge fund cluster has attracted a number of supporting services firms. Greenwich also offers the advantage of being a quick drive from GlobeOp’s headquarters in Harrison, N.Y.

One might think, what gives? And the story does point out the plethora of insurance companies in Hartford and their skill-relevancy to hedge fund services. However there may be a more sinister motive at play:

“Hartford might seem … far enough way from Greenwich that maybe your employees might not as readily jump ship to go work for a hedge fund’s internal operations,” McGuire said.

Given that’s the case, this does work both ways. GlobOp may benefit from better retention, but they also deny themselves access to a talent pool that is probably smart enough to realize that moving to Hartford (or anywhere near it) is going to remove them from a safe regional workforce.

It will be interesting to see if anyone follows their move.

No Comments

In a move that you could see coming a mile away (and which I’m almost certain I recently predicted but am too lazy too look up a link for), several of the largest sell-side investment banks have announced plans to launch their own alternative trading network targeted at institutional investors and their block trade orders. Think Liquidnet but with all of the IT buearocracy and political infighting of Morgan Stanley, UBS, Goldman Sachs, Merryl Lynch and Lehman Brothers thrown into the mix.

“They’re the 33rd entry into this market,” said Robert Hegarty, an analyst with the Tower Group, said of the brokers’ system. “They’ve got a long way to go to get the pole position.”

Exactly… But then again, can anyone really blame them?

Analysts said the system was in some ways a defensive move by top Wall Street firms to keep business from institutional clients who might be tempted to trade with competing systems offering pools of anonymous buyers and sellers.

Well that ship set sail quite a while ago, but then again how soon did you expect a giant effort like this to come to fruition. In all seriousness I think the large banks are right to move in this direction, even if block trading is a small piece of their increasingly complicated profit puzzle. The problem is that it reminds me of the large media outlets that suddenly announced that they are going to have their own blogs. There’s something sadly ironic about it all and you know when you compare the blogs of an enterprising individual (or group of individuals) that it always outshines what passes for a “blog” over at the newsroom in Times Square.

Anyone who’s been on the inside of technical efforts at large sell-side firms knows first hand how complicated it can be to build enterprise-wide systems that involve the cooperation of everyone within the group. Intercine warfare abounds within one bank, let alone half a dozen of them. There’s something a little too johnny-come-lately about this announcement, but it’s worth watching over the next year nonetheless. The banks do afterall have the advantage in infrastructure, both technically and service-wise:

Still, he said the brokerages will have some big advantages, including the fact that their fixed costs are already in place.

“They’ve already got trading research, investment banking services to connect to the customer,” he said. “They’re in the perfect position to get the block order already.”

We shall see…

No Comments

Not really. We are in.. unviolent agreement it seems.

It’s always great to experience an outbreak of harmony on the web, especially having spent the years before HTML and the pretty image-laden web reading comp.* newsgroups with perhaps more than their fair share of CAPITALISED posts flying around.

I was thus delighted to see Benjamin Friedlin’s response to my gentle prod

The concept of performance, especially in the context of things like real time matrix pricing engines, is one that causes me endless fascination.

Go on, go on…

I have talked to quite a few different people about this issue, both verbally and virtually, and I have heard all ranges of opinion from people who are looking into designing ASICs to implement pricing algorithms in hardware (and I thought assembler was a scary maintenance proposition) through the likes of Benjamin who prefer to stick to languages like C/C++ to people who have implemented in java, but performed strange contortions to make sure their applications generate no garbage to avoid GC kicking in at all (some even looking for ways to turn off the collector) {Benjamin: I must admit that I have almost never seen J2EE involved in this kind of scenario}

It’s true, there do seem to be a variety of favored approaches, and I’m inclined to think they are more driven by the familiarity of their authoring practitioners than by what is the best way forward. I’d like to correct Peter in one place however. Although I favor C++ for such applications, I am an even bigger fan of even more esoteric methods, such as using SSEx assembly instructions and to go several steps further, using the GPU. My thoughts on those two methods however will be saved for another day.

The conclusion; there is no conclusion, certainly no right answer. As with every dimension of an engineering solution, you can only define performance with respect to the problem space. The only certainty is that - whatever your performance criteria - you cannot test performance into an application at the end of the development cycle. Build bits, performance test often and optimise as part of your development process (I’d love to hear people’s experiences of this cycle with the logic in an ASIC!) Sorry if that’s stating the blindingly obvious…

You’re forgiven! In all seriousness, I think Peter raises a good strategy when he suggests that one does frequent performance testing. Unfortunately in my experience I have not seen quant teams driven to do much performance testing. A lot of these desks are driven by one manager who’s in turn driven by measuring performance on returns. Certainly you’d think that would provide motivation for improving performance of the very algorithms that drives their trades, yet it’s rarely the case that big banks or even hedge funds want to step outside the line of the safe and proven technology of 3rd and 4th generation languages such as C++ and the managed choice of Java or .NET.

And clearly that is one of the great appeals of a vendor like Azul (Peter, don’t forget to send me that check). I think we’re going to be seeing a lot more innovation in this space and it might help to make this very debate somewhat of a moot issue.

1 Comment

Some smart business developers and politicians in Pennsylvania have approached the top suits of several of the biggest buy-side firms in an effort to appeal to their prudent side and their desire to prevent downtime in the event of any catastrophic events that may cripple NYC’s financial infrastructure:

Declaring “we’ve got your back,” Pennsylvania officials and a developer Tuesday sought to coax executives from major Wall Street companies to establish backup operations in the northeastern part of the state as a way to safeguard their operations against terrorists.

Three helicopters whisked executives from Goldman Sachs Group Inc., Merrill Lynch & Co., Morgan Stanley and others companies from Manhattan to a resort in the Poconos. There they dined on filet and lobster and listened to a pitch for “Wall Street West,” an ambitious effort to build millions of square feet of office space and install hundreds of miles of fiber-optic cable in as many as nine Pennsylvania counties.

Managing risk is the essence of the financial industry. And what seemingly bigger risk is there than another terrorist strike? My only concern regarding this strategy is that any solution such as this can’t simply fall back on yet another easily targeted epicenter of technical infrastructure. It’s not immediately clear how this would be addressed and how the backup systems would be protected, but it’s worth noting that the efforts do have the backing of Washington:

The state also is expected to award a contract in November for the installation of hundreds of miles of fiber-optic cable from New York to Pennsylvania, a process that could take a year to 18 months. Meanwhile, the U.S. Labor Department has awarded a $15 million grant to the Wall Street West initiative for work force training.

There are lots of funny cracks I could make about politics and business, but this is a serious issue that effects the nation’s economic strength as well as the world at large. This is a good issue to watch and I hope that the solution for the long term is more robust than having one monolithic backup center in the Keystone state, and that is no knock on Pennsylvania, but a nod to the notion that a decentralized backup infrastructure is a better defense than one that is centralized. After all, if terrorists can penetrate the well funded defenses of 21st century NYC, how hard is it going to be to distrupt a single secondary location?

2 Comments

Its not often that we get to trawl GooTube (formerly YouTube) for tech videos, but when they come up, they are great. Check out this ‘trailer’ for the launch of SOA from IBM - advertising may never be the same again.

More on IBM’s SOA here.

No Comments

FEN (Financial Engineering News) recently published their fifth annual report on Software for Financial Engineering and Risk Management. After surveying 18 companies, one of their primary findings was that speed remains a critical factor in financial systems:

Nearly all the firms we surveyed cited customer demands for greater speed (without loss of accuracy) with their software systems. And “speed” is defined not only by how quickly a result (a VaR, a price, etc.) can be calculated but also how quickly a new analytical tool and be created, prototyped and implemented across an organization. Two major trends driving this need for speed include continued financial product innovation – specifically increasingly complex products – and the regulatory requirements for real-time enterprise risk management systems that can collect, aggregate and analyze data across multiple activities and capture the combined effects various risk exposures: market, credit, operational, liquidity and so on.

Fortunately, there is a revolution occurring in both hardware and software for financial engineering that is enabling companies to meet this demand. 64-bit computing is already a reality and many firms are quickly adapting their software to take full advantage of the processing speeds afforded by this advance. But at the same time, software providers also realize that speed is as much dependent on innovative algorithms and approaches to computational tasks like optimization and other numerical techniques that converge on answers quickly and efficiently. The need for speed is also being met by providing common service architectures for entire financial management systems (i.e., enhanced performance through better integration) and other novel approaches like cluster or grid computing where complex analyses are shared across multiple central processing units (CPUs) or distinct computer systems.

It’s good to see that there is work in the financial systems space that doesn’t rely solely on more powerful hardware to improve performance. The job of squeezing out more speed is a two step phase, better algorithms that take advantage of better hardware. Go and read the rest of this interesting survey.

No Comments

It’s been years that we’ve been hearing about the eventual transformation of Linux from a wildly popular operating system for back-end and device based software into one that could pose a legitimate challenge to Microsoft’s Windows and even Apple’s OSX. A lot has been made of Linux’s uphill struggle to find its way onto the desktops of not just devotees of the OS, but of large-scale enterprise businesses as well, a place where Microsoft has never had to face even the remotest of threats. This is a fact that even Linux’s greatest proponents are willing to concede, so much so that there’s been a comprehensive and organized effort from all corners of the open source world to help evolve the OS into a capable and mature option for desktop users.

It can be argued that Linux has already reached the point where it can be used for serious desktop work. Open source desktop manager projects such as KDE and Gnome provide users with a coherent and modern windowing system that in many ways is as good as what’s available on Windows and Mac systems. With the great gains being made, one might wonder, why hasn’t Wall St. embraced Linux as a desktop option in any serious way, especially given the tremendous cost of building and supporting desktop systems within large enterprise financial firms.

Digging further however uncovers some unpleasant truths that (at least in my view) are probably significant impediments to Linux’s rise to desktop stardom in large enterprises. While Linux itself may be ready for professional desktop usage, at least in terms of the polished functionality provided by KDE and Gnome, the picture is not so clear when we look at issues of GUI API stability, commercial application compatibility, developer talent, and hardware support.

Let’s start with GUI APIs. There really is no clear winner when it comes to writing GUIs for the *nix world. This is in part because there is no clear winner when it comes to the desktop managers themselves, which each are closely tied to different underlying widget APIs. If you’re a commercial application developer, which do you support? Even if you were to throw your weight behind one over the other (and there are more than two, although KDE and Gnome are the most popular), there’s enough uncertainty about their future direction to prevent a publicly traded company from sinking millions of dollars into system redevelopment and porting efforts.

The issue of API selection leads right into another problem: programmer talent. While various Linux desktop distributions have reached functionally stability, there is a relative dearth of professional experience in the talent market to staff any serious commitment to the technology. Compared to Windows, relatively few can claim to have any serious experience developing high quality and high performance interfaces on any Linux desktop distribution whatsoever. While it may not be fair or even fun to admit, the Win32 platform has decades of technical mindshare amongst developers. Even if you cut your teeth on VB Forms and are now dealing with .NET Windows Forms or even hacking legacy MFC apps, both are simply wrappers of a well documented and slowly evolving message based windowing API. It’s not difficult to unravel the underlying connections between them since they share the same foundation.

But the problems don’t end there. One other advantage that Windows holds over Linux in the desktop space is a dedicated flow of high performance graphics hardware to power the underlying windows and graphical APIs. Bolstered by the gaming industry, Windows has benefited from rapid fire innovation in the hardware accelerated 3D hardware industry. The technology has become so prevalent that OSX has used it to power its high gloss interface for some time now, and it’s well established that Vista (and even XP by way of an addon) will follow suit. Linux on the other hand simply can’t get any luck when it comes to 3D hardware. NVidia has done an admirable job at providing performance drivers for the OS, however they still don’t get the same treatment that Windows and OSX do. And one hardware vendor cannot be counted upon to bolster a standard on an OS platform, especially when you are betting your farm. The fact is that ATI and Matrox simply won’t play along, with the former providing very low quality linux drivers and the latter simply providing none at all.

You might challenge the importance of this, after all how necessary is 3D accelerated hardware for a business desktop platform? The reality is that 3D graphics are an emerging technology that have real applicability when rendering even everyday 2D images at very high speeds, such as frequently-updated trading blotters to name just one example. If Linux can’t endear itself to 3D IHVs, I’m afraid it will always be at a disadvantage when trying to compete for the enterprise desktop.

When all has been said and done, Linux has come a long way in the desktop space. But have we evolved enough with it to use it in a way that is as powerful as its long established incumbent? I’m afraid we haven’t and I don’t see that changing any time soon. It would be nice to see the incredible wealth of innovation and flexibility that Linux affords developers find its way onto user’s desks, but Linux and its proponents will need to build a lot more consensus amongst itself on the issues of APIs, Look & Feel and application compatibility. When that’s all been sorted out, maybe then professional developers will be willing to invest their time to build the skills necessary to actually build systems around it. But even before then, graphics hardware vendors will have to show a much more consistent level of commitment.

No Comments