The Irish internet booking engine company Datalex reported their 2009 results last week, but safer for me given my own employment just to quote the summary from elsewhere:

EBITDA (earnings before interest, tax, depreciation and amortisation) fell from $2m in 2008 to just $100,000 and total annual revenue fell by 18% to $27.1m. The company attributed the fall in revenue to a reduction in demand for its e-business and consulting services and “a dearth of new business opportunities”

Earlier in March they announced that they were putting Panamanian carrier  Copa Airlines into production, but I did pick up one interesting quote in the Tnooz article at the time:

Jim Nation, regional sales director for Datalex USA, says the common platform “imports each airlines’ fares and rules to support shopping requests without the need to hit the host system and Datalex applies the appropriate fare family/fare branding attributes as defined by each airline, customizable per route.”

I try to be as fair and balanced as possible on this blog, as you can see from previous positive things I’ve written on Datalex, but on occasion I’ve mentioned the bad news as well.  

But it was the above quote that was most interesting. Storing all the fares and their application rules is no big deal, but the real challenge is matching the inventory with the fares to find the applicable fares over a given period of time. From the two articles I’ve linked to, it is not totally clear everything that Datalex stores, as they talk about fares and rules but also about not hitting the host system. Caches are definintely flavour of the month, especially with the new search paradigm the industry will move towards,  and I assume Datalex feed a cache based on their production traffic, but it may be that they have a regular feed from airlines and their partners. In one of the articles they claim that they can also return in their searches interline and codeshare, which would mean that they potentially cache as well any interline partners. Not hitting the host system potentially raises the issue of keeping inventory up to date and of showing correct availability, but especially on the latter point, this is where any cache really proves its value – all caches are not created equal, but this must be a challenge for buyers to understand when comparing systems.