Slashdot Mirror
London Stock Exchange Price Errors 'Emerged At Linux Launch'
DMandPenfold writes "Within the first 20 seconds of the London Stock Exchange's new matching engine going live on Monday, price data vendors began displaying incorrect prices, blank prices and wrong trading volumes, according to Computerworld UK sources. Thomson Reuters, Interactive Data and Netbuilder are among the largest data vendors, providing share prices to traders, that have been displaying pricing problems on some stocks throughout the week. Even the LSE's own data vendor, ProQuote, experienced problems. Concerns are being raised that there could be mistakenly setup connections or incorrect software interfaces at some of the large data vendors. Alternatively, there may be a data caching issue at the LSE that means data going out is not properly synchronised between different systems."
I work on large scale air traffic control systems which run Linux and I don't envy the LSE in their task. Most of our interfaces are relatively simple and go out to organisations with a good history of validating interfaces. This trading system seems to have to interface to a lot of little offices around the place running various implementations. Its no surprise some of the interfaces weren't tested to the point where they are known to work 100%, though they may be 100% correct.
http://michaelsmith.id.au
Hi,
The NYSE Arca FIX docs are not quite relevant to this - as what we're talking about here is the market data API.
The FIX protocol (although it does support market data transactions), is typically not the protocol of choice for market data. Typically it'll be non standard provider specific protocols which are optimised for delivering partial updates of information in a very bandwidth efficient manner.
In the parent post someone mentions that Reuters, Netbuilder etc fetch quotes - which is also not entirely correct either. They accept push feeds from the exchange, which contain tick updates to the information about the instruments traded on that exchange. The generation of these ticks is triggered when a match occurs on the order book - which may (or may not) involve a change in price data. They then typically augment this information with other data - such as other identifiers, or derived information such as previous days close value / percentage etc.
It is with these derived fields that the ongoing issues seem to lie, although for a certain period of time, the exchange was actually pushing out zero values for the bid/offer fields. As you can imagine, these plays merry hell with systems that attempt to drive automatic execution based on current price.
I am informed by Netbuilder that they will have pushed a final fix to live within a subset of their servers - but it will probably be Tuesday before this is available across their whole system. This is over a week since the original identification of the issue, not good!
As I work for one of the vendors cited in the article, I can assure you that the reason is freaking simple: the systems that consume the feeds to move them into the trader's desktop are absolute mess of epic proportions riddled with mind boggling wtfs. Missing data, blank values, wrong numbers are common on those systems (root cause are mainly race conditions everywhere, and assumptions about missing data), assembled from random left and right acquisitions. There is no way to really know how they will behave under stress, and corporate culture will not allow managers to actually fix the issues (extremely costly, and corporate bonuses are not aligned to quality).
In fact, it is much cheaper for everyone to just pretend this is a one time occurrence, to fire a couple of middle managers and continue frantically fixing bugs in the mess until it sorta works.