Every day Google processes and serves large amount of data. The talk briefly introduces Google infrastructure and provides a more detailed description of two key technologies that make processing at such a scale possible: BigTable and MapReduce. BigTable is a large scale, high performance, self managing storage system many Google services are built upon. MapReduce facilitates the development of parallel programs by providing simple yet powerful programming model and handling such issues like load balancing and hardware/software failures.
Slides and audio not available in line with Google policy.