How we keep Search fast and reliable
With Search, our goal is helping you find the information you’re looking for — quickly, and whenever you need it. While Search has become vastly more capable over the years, two things remain sacred: speed and reliability. These are simple principles, but they require new and creative solutions to achieve at a global scale. Here’s an update on how we’re keeping Search fast and reliable.
Saving you time with every search
We know you expect Search to deliver the information you're looking for faster than the blink of an eye. That’s why delivering results in a fraction of a second is our baseline, and as we improve Search and build new features, staying fast and reducing latency are priorities.
When we talk about latency, we’re measuring the time between entering a search and seeing results. Like a pit crew, our teams look at every component of Search to find ways to shave off milliseconds. Any increase in latency (from a new feature or change to Search) must be offset by making some other part of Search faster. This drives teams to continually optimize — phasing out slower code and lesser-used features and improving Search.
Trimming time off of individual queries adds up to major time savings for people using Search. Collectively, over the past two years, these latency improvements have saved users over 1 million hours every single day.
When we roll out major improvements to Search from the Knowledge Graph to AI Overviews, we focus on reducing latency. The latency improvements we’ve already made with AI Overviews have saved users another half-million hours daily.
Keeping Search running around the clock
While speed is critical, above all else, Search must be reliable and available when you need it. From record high searches during cultural events like global sports moments to critical searches related to natural disasters, Search is built from the ground up to be available to people around the world, around the clock — so you can get the information you need.
Our systems are designed to handle enormous demand and operate under pressure, even when faced with unforeseen surges in searches. Search data scientists continually evaluate subtle signals, like users refreshing a page, to identify cases where Search is not meeting people’s expectations. Engineers then use these signals to identify weaknesses in the system and build mitigations to prevent outages.
We’ve also built a first-in-class Search infrastructure and have a team dedicated to maintaining it. Our servers are built to process billions of searches every day and connect you with the most helpful results from the web, regardless of the capability of your network or device.
An average user would have to complete around 150,000 queries on Google before encountering a failure due to an error in our Search infrastructure. That means if you searched 10 times a day, it would likely take you more than forty years before encountering a server-side error.
Our teams are constantly fine-tuning and optimizing to ensure Search is the dependable, lightning-fast tool you expect, wherever you are.
