« Google's Colossus Makes Search Real-time by Dumping MapReduce | Main | 6 Scalability Lessons »

How did Google Instant become Faster with 5-7X More Results Pages?

We don't have a lot of details on how Google pulled off their technically very impressive Google Instant release, but in Google Instant behind the scenes, they did share some interesting facts:

  • Google was serving more than a billion searches per day.
  • With Google Instant they served 5-7X more results pages than previously.
  • Typical search results were returned in less than a quarter of second.
  • A team of 50+ worked on the project for an extended period of time.

Although Google is associated with muscular data centers, they just didn't throw more server capacity at the problem, they worked smarter too. What were their general strategies?

  • Increase backend server capacity.
  • Add new caches to handle high request rates while keeping results fresh while the web is continuously crawled and re-indexed.
  • Add User-state data to the back-ends to keep track of the results pages already shown to a given user, preventing the same results from being re-fetched repeatedly.
  • Optimize page-rendering JavaScript code to help ensure web browsers could keep up with the rest of the system.

We see here the merging of various caching strategies with the real-time web. A major evolution from the early days of long web crawling cycles and disk dominated storage. The world is faster now and fast requires memory and events working in perfect harmony. Hopefully over time we'll learn more of the nitty gritty details. It should be fascinating. 

Related Articles

  • Google search index splits with MapReduce. With Caffeine, Google can update its index by making direct changes to the web map already stored in BigTable.
  • Our new search index: Caffeine. Caffeine lets us index web pages on an enormous scale. In fact, every second Caffeine processes hundreds of thousands of pages in parallel. If this were a pile of paper it would grow three miles taller every second. Caffeine takes up nearly 100 million gigabytes of storage in one database and adds new information at a rate of hundreds of thousands of gigabytes per day. You would need 625,000 of the largest iPods to store that much information; if these were stacked end-to-end they would go for more than 40 miles.

Reader Comments (9)

How to serve up 5-7x more results?

How indeed.

Caching is great, when you get a lot of cache hits. Google Instant ups the cache hits by perturbing everyone's search patterns to be more similar and reducing the Long Tail searches:




September 9, 2010 | Unregistered CommenterBob Warfield

For sure they do not use the same Google Analytics and AdSense we, poor mortals, all use. ;)

September 9, 2010 | Unregistered Commenteridont

Easy. They switched to MongoDB.

September 9, 2010 | Unregistered CommenterGregor J. Rothfuss

@Gregor - because MongoDB is web-scale? ;)

September 9, 2010 | Unregistered CommenterMatt Passell

I think they are writing most of the search queries to /dev/null - as we all learned a few days ago. It scales like hell! ;)

The instant search is just impressive!

September 10, 2010 | Unregistered CommenterJust me

Totally agree. "The world is faster now and fast requires memory and events working in perfect harmony." Cloudscale's Realtime Data Warehouse http://cloudscale.com provides an in-memory, massively parallel architecture to meet the challenges of this new post-MapReduce world.

As reported yesterday, Google http://bit.ly/9812cv agrees with Cloudscale http://bit.ly/c10sT8 that MapReduce and Hadoop are no use for this new era of realtime data. Google's view on MapReduce "You can't do anything with it that takes a short amount of time". If speed/latency matters as well as scale then you need to look elsewhere http://bit.ly/cy1CEs


Oh, and like MongoDB we're web scale. Loved the video.


September 10, 2010 | Unregistered CommenterBill McColl

I think so that the cache is not ever the exelent solution...maybe in this case only for a few seconds...and flush again the cache.

September 10, 2010 | Unregistered CommenterAlfred

Sure it's impressive that they can do it but honestly I'm not sure I actually like it better than the old version.

I mean i guess we don't have to press the enter key but often I'll look at old search results to tailor my new searchs which isn't really possible here.

September 10, 2010 | Unregistered CommenterMatt

hundreds of thousands of pages per second, sounds crazy... must be using the software that can map the human genome lol. Technological advances are spreading like wild fire, I'm so ready to be flying around in my new car like on the Jetsons.

September 10, 2010 | Unregistered CommenterRobert K.

PostPost a New Comment

Enter your information below to add a new comment.
Author Email (optional):
Author URL (optional):
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>