Market data terminals, past, present and future…

Working within capital markets nearly always involves working with a lot of different types of data.  Some examples from my working life are:

  • Highly structured time series data that has been collated, cleaned and checked for accuracy
  • High frequency data such as tick-by-tick data from an exchange traded product
  • Analytic data such as “greeks” for derivatives
  • Portfolio level analytics such as percentage of a portfolio held in cash and near-cash instruments
In some consulting engagements clients have provided me with a great deal of market data in order to get the job done.  How does this get charged?

[Updated with link to source code] 
In a typical model, there is a cost for the actual market data software installation – say £1500 per calendar month per user.  Then there is a pass-through cost of exchange data, so if an exchange publishes real-time data with no delay for a market segment at £50 per user per calendar month then this will be passed on to the end-user firm.
The market data world is dominated by two big firms, ThomsonReuters and Bloomberg.  There are other firms in that sector such as Markit but let’s focus on ThomsonReuters and Bloomberg for now.
ThomsonReuters have recently revamped their product line with Eikon taking over from the Reuters 3000 Xtra.  The old flagship 300 Xtra platform was a pretty substantial beast written in a mixture of Microsoft technologies such as VC++.  There was even an embedded implementation of VBA to allow application automation. I have seen demonstrations of Eikon and it has some nice new touches but I am not sure of the technical architecture.  
Bloomberg has a single desktop product – “Bloomberg Professional”.  In terms of user interface, it’s a throwback to the early 1980s with some contemporary pieces.  The technology stack is hidden away from users, an internet search does not show any clear recent statement from someone at Bloomberg of the actual code base.  I have heard that it’s a mixture of C, Fortran, C++, Python, Ruby and Javascript.
Both systems are now software based although both vendors offer fancy keyboards.  The software applications are both pretty heavyweight.
So why this blog post?  Below is a screen capture of a very simple software based market data terminal I have been working on as a proof-of-concept for a client firm to use for a specific small market they serve.  This is of course crude and not a finished product, but it’s so simple to use modern software frameworks and tools to create a simple event driven real-time web page
For this application the technology stack includes: