Hacker News Reader

Comment by edouard1234567 | Hacker News Reader

My guess is that they predict the most common searches and constantly query and cash results for these searches, something I don't think Kayak or hipmunk does. I don't think you can launch a search product on google if it takes more than .xx seconds to return the result...

View on HN

I'm not so sure, I looked for a few random flights. However, having built my own very rudimentary internal flight search tool from the global flight list, I can tell you there are roughly 7 million flights per month, a small data set by google standards.

It's not the size of the dataset that is relevant but more that this dataset changes in realtime : price + availability. These requests go through GDSs that are relatively slow, that's why so called real-time search engines like kayak are fairly slow. What do you exactly mean by 7M flights? Keep in mind that most search requests involve a combination of flights (leg), outbound, inbound and sometime multiple stops...

It could have something to do with the fact that Google owns ITA.

I think that owning ITA allows them to have more data to mine/predict and better access but it doesn't explain how fast the searches are. Most airfare searches still hit GDSs which are relatively slow to respond. The only way I can think of Google can deliver such a performance is by intelligently pre-caching airfare searches. Google?

Reply on news.ycombinator.com