Google BigTable

Google Bigtable
Kasim Mothana
Christian D. Ozoria

 Bigtable is a compressed, highly distributed,
high performance data storage system
 Is used by other Google products, including
Google Search, Google Analytics and Google
Earth, and a part of the Google’s Platform as a
Service (PaaS)
What is Bigtable?

 Bigtable is primarily designed to scale petabytes of data (for
commercial databases such scale is difficult and or expensive to
handle) across thousands of commodity machines
 Scalability is accomplished with the flexibility to add more resources
to the system on the fly without the need to reconfigure the system
 Its model and implementation allow to maintain good performance
on such volumes of data
 The model makes it widely applicable
 Cost-effective
 Self-managing
Why Bigtable?

 Bigtable can be loosely compared to a spreadsheet that maintains
versions of cells, each with a timestamp. Often it is referred to as a
distributed multidimensional sorted map – the map where a row key,
a column key, and a timestamp are mapped to a value
 Key model elements:
 Row key (stored as a string)
 Column key
 Timestamp
 Value that is an uninterrupted array of bytes
 Bigtable is a semi-structured store: for the same row key we can
have different columns (similar to the spreadsheet)
Data Model

 Storage is organized by row key (ordered
alphabetically), column key and timestamp
 Columns are grouped into column families to improve
storage characteristics
 Each cell can hold multiple version of data
Data Model (cont.)

Example 1
Below is an example of a social network for United States
presidents. Each president can follow posts from other
presidents. The following shows a Bigtable table that tracks
who each president is following on Prezzy:

 The user name (in this case, the president name) is
used as the row key
 This table has one column family containing multiple
column qualifiers
 Because new column qualifiers can be added
dynamically, it is easy to add new followers
 Illustrates the mapping
(row:string, column:string, time:int64) → string
Example1 -- Explanation

 The row key is a reversed URL (explained in the following
slides)
 The Contents column family contains the page contents
 The Anchor column family contains the text of any anchors
that reference the page
 CNN’s home page is referenced by both the Sports Illustrated
and the MY-look home pages, so the row contains columns
named anchor:cnnsi.com and anchor:my.look.ca
 Each anchor cell has one version; the contents column has
several versions
Example 2 – Explanation

 The Bigtable implementation has three major
components:
 A library that is linked into every client
 One master server
 Many tablet servers
Implementation

 Data is dynamically partitioned based on the row key;
each row range creates a tablet, which is the unit of
distribution and load balancing
 Clients can exploit this property by selecting their row
keys so that they get good locality for their data
accesses. In Example 2, pages in the same domain are
grouped together by reversing the hostname
components of the URLs: hadoop.apache.com and
hbase.apache.com are stored next to each other as
org.apache.hadoop and org.apache.hbase
Dynamic Tablet Partitioning

 Google Analytics is a service that helps webmasters
analyze trafﬁc patterns at their web sites
 Tracks website traffic and makes it available to
webmasters
Application: Google Analytics

 Personalized Search is a service that records user
queries and clicks when using Google search
 Personalized Search stores each user’s data in
Bigtable. Each user has a unique user id and is
assigned a row named by that user id
 Enables a more personalized search experience
Application: Personalized Search

 Incredible scalability. The Cloud Big Table is designed
to scale in direct proportion to the machines in the
cluster
 Simple administration. The Cloud Big Table handles
upgrades, restarts, and replication transparently
Benefits

 Bigtable is a distributed system for storing data at
Google
 Bigtable clusters have been in production use since
April 2005
 In August 2006, more than sixty projects were using
Bigtable.
 Users like the performance and high availability
provided by the Bigtable implementation
Conclusion

 Users like that they can scale by simply adding more
machines to the system as their resource demands
change over time
 Several additional Bigtable features, such as support
for secondary indices and infrastructure for building
cross-data-center replicated Bigtables with multiple
master replicas, are in the process of development.
Conclusion cont.

 Bigtable: A Distributed Storage System for Structured Data
http://static.googleusercontent.com/media/research.google.com/en//archive/bigtable-osdi06.pdf
 Under the covers of the App Engine Database.
 http://snarfed.org/datastore_talk.html
 Wikipedia: BigTable
 https://en.wikipedia.org/wiki/BigTable
 Cloud BigTable Beta
 https://cloud.google.com/bigtable/docs/
 Introducing Google Cloud BigTable
 http://googlecloudplatform.blogspot.com/2015/05/introducing-Google-Cloud-Bigtable.html
 CS262B Advanced Topics in Computer Systems. Spring 2009.
http://www.eecs.berkeley.edu/~culler/cs262b/summary/bigtable.html
 Tudorica, Bogdan George, and Cristian Bucur. "A comparison between several NoSQL databases
with comments and notes." In Roedunet International Conference (RoEduNet), 2011 10th, pp. 1-5. IEEE,
2011.
 Lee, Ken Ka-Yin, Wai-Choi Tang, and Kup-Sze Choi. "Alternatives to relational database: comparison
of NoSQL and XML approaches for clinical data storage." Computer methods and programs in
biomedicine 110, no. 1 (2013): 99-109.
 Nayak, Ameya, Anil Poriya, and Dikshay Poojary. "Type of NOSQL Databases and its Comparison
with Relational Databases." International Journal of Applied Information Systems 5, no. 4 (2013): 16-
19.
Sources

 Vicknair, Chad, Michael Macias, Zhendong Zhao, Xiaofei Nan, Yixin Chen, and Dawn Wilkins. "A comparison of
a graph database and a relational database: a data provenance perspective." In Proceedings of the 48th
annual southeast regional conference, p. 42. ACM, 2010.
 Chang, Fay, Jeffrey Dean, Sanjay Ghemawat, Wilson C. Hsieh, Deborah A. Wallach, Mike Burrows, Tushar
Chandra, Andrew Fikes, and Robert E. Gruber. "BigTable: A distributed storage system for structured
data." ACM Transactions on Computer Systems (TOCS) 26, no. 2 (2008): 4.
 Leavitt, Neal. "Will NoSQL databases live up to their promise?." Computer 43, no. 2 (2010): 12-14.
 Stonebraker, Michael. "SQL databases v. NoSQL databases." Communications of the ACM 53, no. 4 (2010): 10-
11.
 Pallis, George. "Cloud computing: the new frontier of internet computing." IEEE Internet Computing 5 (2010):
70-73.
 Vossen, Gottfried. "Big data as the new enabler in business and other intelligence." Vietnam Journal of
Computer Science 1, no. 1 (2014): 3-14.
 Zicari, Roberto V. "Big data: Challenges and opportunities." Big data computing (2014): 103-128.
 Moniruzzaman, A. B. M., and Syed Akhter Hossain. "Nosql database: New era of databases for big data
analytics-classification, characteristics and comparison." arXiv preprint arXiv:1307.0191 (2013).
 Thankachan, Tomcy. "Google Big Table" November 27th 2012. PowerPoint presentation
 Jacotin, Romain. "Lecture: The Google Big Table" October 9th 2014. PowerPoint presentation
Sources cont.

Google BigTable

More Related Content

What's hot (20)

Similar to Google BigTable (20)

More from New York City College of Technology Computer Systems Technology Colloquium (12)

Recently uploaded (20)

Google BigTable