If I were writing a movie script,
Id call this development The Empire Strikes Back. OK, thats not an
original title, but it seems as though the technology debate under
way among the GDS companies and their emerging competitors is a
rerun, too.
The empire that is
striking back is the four GDSs -- Sabre, Galileo, Worldspan and
Amadeus -- the folks who used to hold a monopoly in setting
suppliers booking fees for distribution services.
They feel they are
getting a raw deal in the battle against new entrants like ITA
Software, G2 SwitchWorks and Farelogix. The Big Four believe that
their technology has been misunderstood and maligned in the heated
contest for the hearts and minds of the countrys largest travel
agencies and suppliers.
So, the GDSs, all of
whom are members of the Interactive Travel Services Association,
have signed on ITSA to set the record straight.
In the process, it
should be noted, they also are proclaiming that the new entrants
are blowing smoke.
ITSA Executive
Director Art Sackler and Worldspan Chief Information Officer Sue
Powers briefed me the other day on changes taking place in those
large computer centers at the GDS companies.
Just dont call it
legacy technology, Powers implored. Really, nothing can be further
from the truth.
Powers said Worldspan
takes a hybrid approach to technology, using updated, modern
mainframes when they are deemed the optimal business solution and
distributed platforms for others. (Unlike mainframes, where
applications are centralized, distributed systems break out
applications in a network of servers.)
And, she added, the
tech changes that Worldspan has implemented are generally
representative of whats been going on at Sabre, Galileo and
Amadeus.
So Powers estimates
that about half of the Worldspan system is on mainframes and half
on distributed platforms these days.
Take Worldspans fare
shopping system, its most-used application, according to Powers.
About 90% of the application -- that is, U.S. and Canadian points
of sale for destinations worldwide -- has for the past three years
run on a server-based Microsoft .Net platform, she said.
By the end of the
year, the newer technology will be deployed for all global points
of sale, Powers said.
And Worldspans
browser-based desktop application for travel agents, Worldspan Go,
similarly uses the .Net platform sans mainframes, as does its
corporate booking tool, Trip Manager.
Worldspan
applications still on legacy (sorry, Sue) systems include
ticketing, car, hotel, rail, ferry and tours, Powers
said.
But these are not
your fathers mainframes. Powers argued that they are modern and
flexible, and that other industries, such as banking, have not
found a solution that performs better than the TPF (transaction
processing facility).
Im not willing to
concede that mainframes are fundamentally bad, Powers
said.
Regarding the
mainframes of this era, Worldspan is beta testing z/TPF, a spiffier
mainframe operating system from IBM. With the evolution of
mainframes, Worldspan developers can use the same language to write
programs for mainframes and the .Net platform, Powers
said.
The point of this
technology exercise is that the GDSs believe that their technology
is reliable and they are adapting it to take costs out so they can
better service suppliers and agencies and compete with the
alternative GDSs.
Sackler and Powers
argue that the GDSs and their new-entrant competitors are trying to
tackle different business problems.
The GDSs, they say,
are global systems processing 10,000 transactions per second while
the new entrants are focusing on simple air itineraries from North
American points of sale.
In that regard, as
the GDS companies content agreements with airlines expire over the
next 18 months, they wouldnt have a problem competing if they too
developed a model for simple transactions, where they dont pay
inducements to agencies and dont provide equipment and certain
support services, Powers said.
And, Sackler said the
new entrants have not solved issues involving scale, reliability,
security and privacy. So far weve seen assertion and no real
demonstration or very limited [demonstration], he said.
Some of this
technology debate -- discussions about GDS mainframes versus the
upstarts scalable, PC-based systems -- should sound very
familiar.
In 2000, Orbitz was
running tests using the then-revolutionary ITA Software search
engine, with its new computational methods for analyzing flight,
fare and seat availabilities using scalable PCs instead of the
mainframes that the GDSs used.
Sabre, at the time,
scoffed at the ITA engine, with one official claiming that
mainframes are thousands of times more capable than any PC and that
the GDSs have proven their reliability over decades whereas ITA had
never booked a fare.
An Expedia official
noted at the time that some of the flight connections retrieved
through ITA search results were wacky and couldnt be
booked.
Well, five years
later, the GDSs are using fare-shopping systems that have much in
common with the ITA platform, and all concede that ITA changed the
air-search landscape.
The new entrants
today are similarly spurring innovation. Will they be able to
accomplish their goals?
As Sackler said, The
proof will be in the pudding, wont it?