Google Building Blocks Review Flashcards

Flashcards reviewing Google's core infrastructure technologies: MapReduce, Google File System (GFS), and BigTable.

Google's major software systems include MapReduce, GFS (Google File System), and __.

BigTable

MapReduce is a methodology for exploiting __ in computing clouds.

parallelism

In the context of search engines, MapReduce is used for building Google's Search Index and __.

Article clustering for Google News

Examples of modern Internet applications that require managing immense amounts of data quickly include Dish network click collection and __ data collection.

Tesla car usage

MapReduce solves the problems of parallelization, fault tolerance, I/O scheduling, and __ for the programmer.

Monitoring & Status updates

The Map/Reduce paradigm involves breaking records into segments, mapping to extract something of interest, grouping intermediate results, reducing to aggregate results, and generating __.

final output

In MapReduce, the __ manages the parallel execution and coordination of tasks automatically.

system

In a MapReduce computation, map tasks turn a chunk into a sequence of __.

key-value pairs

The Map function parses a document, extracts each word and uses each word as a __.

key

The Reduce function aggregates intermediate results by __.

key

The master controller knows how many __ tasks there will be.

Reduce

The Reduce function is generally __ and commutative.

associative

The Google File System (GFS) is designed for efficient, reliable access to data using large clusters of __.

commodity hardware

GFS supports automatic sharding of large files, automatic recovery from failures, and is optimized for __ access to huge files.

sequential

In GFS, files are divided into fixed-size chunks of __ megabytes.

64

In GFS, the __ server holds all metadata, like namespace, access control, and chunk locations.

Master

In GFS, data transfers happen directly between __ and chunkservers.

clients

GFS is optimized for __ files rather than rewrites.

appended

Bigtable is a compressed, high performance, proprietary data storage system built on top of __.

the Google File System

A table in Bigtable is sparse, distributed, persistent, multidimensional, and __.

sorted map

In Bigtable, data is treated as __.

uninterpreted strings

In Bigtable, rows are ordered __.

lexicographically

In Bigtable, each cell contains a unique __ version of the data for that row and column.

timestamped

In Bigtable, columns have two-level name structure consisting of family and __.

optional qualifier

In Bigtable, Timestamps are used to store __ versions of data in a cell.

different