Priceline uses Zing® to ensure jitter-free operation in their hotel room rates cache
Their hotel room cache infrastructure holds over 7.5 billion room rates, is updated at over 42,000 rates per second, and delivers over 28,000 rates per second.
Priceline first tried to solve their GC pause issues by shrinking down the size of the individual JVMs, but they didn’t want their hardware footprint to become unmanageable. They just wanted to find a way to get through their garbage collection events.
So they deployed Zing.
Priceline went live with Zing in the Summer of 2013 after a few months of testing, and director of engineering Nasreen Ali said that this was the first time that software engineers had not been called in over a holiday weekend to deal with memory issues in the JVMs as conditions in the hotel market changed. The two software engineers that were “fussing around with garbage collection and heap memory” are now focusing on new application development.
"We love Java as a language and it has served us well. But we got into a garbage collection nightmare.”
Travel sites face many challenges in a rapidly evolving marketplace. One very important challenge is meeting customer demand for accurate, timely data. In Priceline’s case, this meant hotel pricing data from thousands of properties that was then distributed to end customers as well as Priceline’s travel industry partners. The best platform with the freshest and most accurate data wins.
Priceline’s custom in-memory cache needed to be able to handle massive amounts of data – and Zing made it possible and practical.
Garbage collection is no longer a factor for Priceline’s operation
Highly-skilled engineers could be freed from JVM tuning, and able to create new value
With a highly performant Java-based cache subsystem, Priceline continues to delight its customers
By deploying a better JVM, Priceline met their performance targets