What Fare caching means
Live airline shopping is slow and expensive. Each search can trigger many requests to GDSs, NDC APIs and low-cost carrier connections, and suppliers measure sellers by their look-to-book ratio. A busy website or metasearch partner can send thousands of searches for every booking.
A cache keeps the answers to recent or popular searches, such as Dubai to London next month, for a set time. When a similar search arrives, the system returns the stored result instead of polling suppliers. Caching is usually part of the travel middleware layer that sits between booking front ends and supply.
The trade-off is freshness. Airline prices and seat availability change constantly with revenue management, so a cached fare can be wrong by the time a customer clicks. Well-designed systems therefore treat the cache as a fast first answer and run a live price and availability check before the booking is confirmed.
Vendors now offer caching as a service. In February 2026 Sabre announced cache-powered shopping in SabreMosaic, using predictive models to decide whether to poll airlines live or return validated cached results. It reported up to a 28% reduction in look-to-book ratios and up to 95% accuracy against live polling in measured deployments. Ypsilon markets a fare cache that pre-fetches fares by airline and market for tour operators, metasearch sites and OTAs.
How Fare caching works
A typical design answers from the cache when it can and goes live when it must.
- Receive the searchA customer or agent searches for a route, dates and passengers. The system builds a cache key from those parameters.
- Check the cacheIf a stored result for that key exists and is within its time-to-live, the system returns it straight away.
- Fall back to live shoppingIf there is no usable result, the system queries suppliers live, returns the answer and stores it for later searches.
- Refresh in the backgroundPopular routes may be pre-fetched or refreshed on a schedule, and some systems use models to predict which results are likely to have changed.
- Revalidate before bookingWhen the customer selects a fare, the system checks live price and availability. If it has changed, the customer is shown the new price before payment.
Fare caching in practice
Verified vendor figures. Sabre's February 2026 announcement for SabreMosaic cache-powered shopping cited, from measured deployments, up to 28% lower look-to-book ratios, up to 95% accuracy versus live polling and responses under 500 milliseconds for traditional content. Ypsilon states that its cache averages around 60% hits and can be loaded several times a day.
Illustrative. An OTA caches results for 15 minutes. A search for Dhaka to Kuala Lumpur at 10:00 is answered live and stored. At 10:05 another customer gets the cached result in a fraction of a second. At 10:12 that customer selects a fare, the system re-prices it live, finds the booking class has sold out and shows the next available price. The timings are invented for illustration.
Who uses it
Advantages and limitations
Advantages
- Faster search results, which improves customer experience and conversion.
- Fewer live requests to suppliers, which lowers search costs and look-to-book ratios.
- Makes features such as fare calendars and flexible-date searches practical.
- Reduces the risk of hitting supplier rate limits during traffic peaks.
Limitations
- Cached prices can be stale, leading to price changes or failures at booking.
- Frequent price changes at checkout can frustrate customers and harm trust.
- Choosing time-to-live values is a balance between cost and accuracy for each route.
- Some suppliers restrict how long or in what form their content may be cached.
Frequently asked questions about Fare caching
Why does the flight price change when I click to book?
The price on the results page may have come from a cache or an earlier search. When you select it, the system checks the airline live, and if the booking class has sold out or the fare has changed, a different price appears. Airline revenue management changes prices and seat availability throughout the day.
How long should flight search results be cached?
There is no single right answer. Popular, volatile routes need short cache lifetimes, while quieter routes far in the future can be held longer. Many systems set time-to-live values by route and departure date and always revalidate before booking, so stale results never reach ticketing.
Does caching break airline or GDS rules?
Not in itself, but supplier contracts may set conditions on how content is stored and shown. Agencies and developers should check their GDS, NDC and aggregator agreements for caching, display and price accuracy rules before building a cache.
What is the difference between a cache hit and a cache miss?
A cache hit means a valid stored result was found and returned without contacting suppliers. A cache miss means nothing usable was stored, so the system had to search live. A higher hit rate saves cost, but only if the cached results are still accurate.
Own your travel platform outright
The Source Code Edition gives you the full codebase to host, customise and extend on your own infrastructure.
See the Source Code Edition