Tuesday, 22 September 2015

Google Unlimited Capacity

This chapter details how Google built its infrastructure to scale with the grown of the internet as a whole. Google wanted to build a scalable search model, which they accomplished by making their search algorithm based purely on mathematics. This allowed for the search model to operate unadulterated by human hands, assuring that the algorithm would scale without human interference.

Larry Page and Sergey Brin used $15,000 of their own money to build their first server that could store 1 terabyte of data, a trend of building in-house that continues throughout Google’s life as a business. The PageRank system that they employed at the company’s outset could take weeks to calculate properly, resulting in Google searches that would take up to 10 seconds to complete at the site’s outset.

To combat this, they began building data centers. Their first center was built in The Dalles, Oregon, which was mired in controversy because of Google’s choice to conduct business through private firms and private contractors, circumventing state channels. Google then built a second data center to create overlap with the Dalles center, known in the business as redundancy, allowing Google to effectively backup their first data center should it suffer from a power outage or loss of service.