Notes on Foundation of Google

Add Summary:

10 years ago (1999-2000)

University of Samford (Stanford)

Web Search

Larry and Serge

Goals

Today

Over 1 billion searches a day
Highly distributed – Over 50% of queries do not come from the US
Customers expect response time of ¼ of a second. (2.5x the speed of a blink)
Queries Change Quickly. 20% of queries change every 90 days
The amount of data being updated on the web was significant. 10-20% of content changes every time they complete a query. Caching not as beneficial.
Index every 4 hours of the data stored. Data is about the size of the Library of Congress
Google Calendar alone – Over 1 million emails and SMS a day
70% of emails are spam. Less than 1% SPAM makes it into inbox
Scalability and Reliability – Because they had no money and therefore had to use cheap components that failed often, they had to build fault-tolerance and HA in the infrastructure.
Power-Consumption – Cut down on power consumption with regards to load-balancing

Gartner

Innovation

Employees can spend 20% of their time on anything that aligns with the strategies of the company.
New ideas, features, etc.
Gmail, Docs, Flight Simulator, Internal systems (Google News)
All new releases are initially used internally and provide feedback
Employees can focus on core development and don’t have to worry about every aspect (HA, backup, recovery, etc.)
Very Complex Hardware solution with software on top that masks the complexity of the HW.
- Big Tables
- Map Reduce
All apps were not designed to go into the Enterprise market

Beginning of 2007

Google realized that they apps were running their company internally and therefore could be useful to other companies.
They had the capacity to support other Enterprises
“Most innovation that happens in the world happens in the consumer space.”