Your choice of "the best example" of decentralised trust solving useful problems supports the original poster's contention that the blockchain industry is desperately trying to sell itself as the solution to problems it's badly suited to. Virtually nobody in the travel industry is excited about blockchain, still less as a GDS replacement.

Decentralisation and anonymity are useless to the travel industry and from the point of view of the industry players, public access to local copies of all their [current and historical] inventory and reservation data is an alarming bug, not a feature. What they actually want is for a buyer to be able to via an API (i) access the current price for each relevant item of inventory that's currently on sale from lots of providers (ii) update a single private canonical record belonging to the one [necessarily trusted] party they buy from (which may or may not affect the availability of inventory for future queries). The reason the systems are slow is because pricing and routing is incredibly complex and diverse, and the associated datasets are enormous (and often computationally expensive to process even without a layer of cryptographic crap on top). It's something of an understatement to say that turning that into thousands of local copies of an ever-growing blockchain which needs to be updated to account for a million reservations per minute is not going to be an improvement...