Saturday, December 20, 2008

Should the financial services industry be using Flex?

For the last ten years software built by banks can be broadly divided into two very distinct categories. One type is software destined to be used internally by traders, sales people and the like. This software needs to be "connected", fast and is typically what I call "push-oriented". For example internal trading systems typically need to show streaming prices that a trader can "execute" (i.e not just indicative prices). The other type is software is designed to be used externally by the bank's customers and partners.

Today internal systems in banks are typically built using rich native desktop technologies such as Java Swing or Windows Forms.NET. On the other side of the fence software built for external use by the banks' customers is typically built using traditional web technologies such as JSP/Struts/Spring MVC or ASP.NET.

For a long time now business managers in banks have been under pressure to offer better services to their customers and partners. Hedge funds, for example, are typically staffed by the brightest and best former employees of the investments banks and represent very demanding customers. As a consequence IT managers in the banks are under pressure from the business to come up with software that provides as good an experience outside the firewall as is offered to the bank's traders inside the firewall. The last five years has seen the development of increasingly sophisticated web applications, still using traditional web technologies but with the addition of AJAX to the mix, that look more and more like the bank's internal systems.

In practice this means that most banks are developing and maintaining two software systems that increasingly offer overlapping functionality, one for internal use and one for external use. The push for more service-oriented architecture in banks is one of the responses this this duplication of effort. Clearly the parts that are common between systems that largely do the same thing can be "factored out" into common services to reduce the duplication.

Nonetheless building separate user interfaces, one for internal use and one for external use, has a significant cost implication. Given that even the smallest IT projects in the financial services industry cost in the order of tens of millions of dollars (many cost much more than this), if it's possible to build software that provides both the performance and connected-ness of traditional native desktop applications so that it can be used internally while at the same time providing the "reach" of a traditional web application so that the same software can be deployed externally, there's a huge potential cost saving for financial services institutions.

Clearly RIA technologies, such as the Adobe Flash Platform, allow banks to take advantage of this opportunity. RIA represents a huge bottom line opportunity for the financial services industry worth thousands of millions of dollars. What's surprising is that, even in the current financial climate, most IT managers in the financial services industry don't appear to have identified this opportunity yet.

Instead the visionaries in the financial services industry are the business managers. One great example of a visionary is Mike Chadney who gave up his position as head of derivatives trading at an a investment bank in London to build a business providing on-line options trading on the internet.

The company that Mike founded, CityOdds Ltd., has a traditional multi-threaded real-time pricing and quoting engine at the back-end and a front-end RIA built using Flex and Adobe LCDS.

It's been a non trivial exercise to build a system that can deliver tradable quotes, process cash trades and revalue positions in real-time to thousands of connected desktops concurrently. Nonetheless I'm convinced this is the shape of things to come in the financial services industry.

Buy ADBE and register to trade for free at!


Anonymous said...

Trading platforms based on AIR (for desktops) might be ok but still would prefer Java.

The Flash Player tends to crash a lot in browsers, hardly the thing you want especially if you're a trigger happy scalper trying to day trade.

Maybe when Flash Player is multi-threaded and more stable will we see it happened in this industry.

Anonymous said...

pushing flash player to its limits? he is not pushing data from flash player but to flash player. i think he is pushing jms to its limits if he is using one. i think flex is bad solution for highly responsive apps. flash player is really fast but flex framework just sucks. it is bloated and very slow. ajax interfaces do better job.

Rich Apps Consulting said...

I don't think Flex is matured enough to be used in Financial Services Industry. Still there is a long way to go for Flex.

spookyluke said...

Lets hammer this out.

Flash Player DOES NOT TEND to crash in most browsers, either you are viewing dodgy movies or your browser/plugin is broke by other means.

Flex is not slow that is a misconception invented by people who dont like it. It carries slightly more bloat yes, about as much as a low res photo so, unless you are on a 56K dial up and clear your browser cache every time you use it then whats the big deal?

Ajax interfaces do not do a better job if they did I would still be programming Ajax. I was reluctant when my ex company wanted to switch to Flex now I would never go back. Even the best Ajax frameworks (like Ext) will always have minor flaws in the x-browser department.

If I had £10 for every recruitment consultant that has tried to poach me for the financial services industry in the last 6 months I would have enough to take a 2 week holiday in the sun. Some people might think Flex is not mature enough, the Financial Services industry seems to be dissagreeing with you. Mature enough for what? Its been around for nearly half a decade now and based on something even older by the time you see it as mature enough computers will probably be obsolete.

Flash Magazine said...

Flash magazines are still the best option for creating digital publications from print versions. The reasons are simple. Flash platform is stable, there is a huge development community constantly churning out components and continuously improving the usability and the already-phenomenal functionality of a flash magazine.