Introduction to NoSQL
NoSQL Seminar 2012 @ TUT
Arto Salminen
What is NoSQL?
Class of database management systems (DBMS)
"Not only SQL"
Does not use SQL as querying language
Distributed, fault-tolerant architecture
No fixed schema (formally described structure)
No joins (typical in databases operated with SQL)
Expensive operation for combining records from
two or more tables into one set
Joins require strong consistency and fixed
schemas
Lack of these makes NoSQL databases more flexible
It's not a replacement for a RDBMS but compliments it
Database Scaling
RDBMS are "scaled up" by adding hardware
processing power
NoSQL is "scaled out" by spreading the load
Partitioning (sharding) / replication
Users
A-K
Users
L-O
App
App
Load
balancer
Load
balancer
Users
P-S
Users
T-Z
Users
A-Z
Users
A-Z
Users
A-Z
Users
A-Z
Relational DB Scaling
At certain point relational database won't scale
NoSQL DB Scaling
Scaling horizontally is possible with NoSQL
Scaling up / down is easy
Supports rapid
production-ready
prototyping
Better handling of
traffic spikes
Where NoSQL Is Used?
Google (BigTable, LevelDB)
LinkedIn (Voldemort)
Facebook (Cassandra)
Twitter (Hadoop/Hbase, FlockDB, Cassandra)
Netflix (SimpleDB, Hadoop/HBase, Cassandra)
CERN (CouchDB)
History of NoSQL
MultiValue databases at TRW in 1965.
DBM is released by AT&T in 1979.
Lotus Domino released in 1989.
Carlo Strozzi used the term NoSQL in 1998 to name his lightweight,
open-source relational database that did not expose the standard SQL
interface.
Graph database Neo4j is started in 2000.
Google BigTable is started in 2004. Paper published in 2006.
CouchDB is started in 2005.
The research paper on Amazon Dynamo is released in 2007.
The document database MongoDB is started in 2007 as a part of a open
source cloud computing stack and first standalone release in 2009.
Facebooks open sources the Cassandra project in 2008.
Project Voldemort started in 2008.
The term NoSQL was reintroduced in early 2009.
Some NoSQL conferences
NoSQL Matters, NoSQL Now!, INOSA
CAP Theorem 1/2
It is impossible for a distributed computer
system to simultaneously provide all three of
the following guarantees:
Consistency (all nodes see the same data at the same
time)
Availability (a guarantee that every request receives a
response about whether it was successful or failed)
Partition tolerance (the system continues to operate
despite arbitrary message loss or failure of part of the
system)
A distributed system can satisfy any two of these guarantees
at the same time, but not all three.
Gilbert and Lynch, Brewer's conjecture and the feasibility of consistent, available, partition-tolerant web services, 2002
CAP Theorem 2/2
In other words, CAP can be expressed as "If
the network is broken, your database wont
work"
"won't work" = down OR inconsistent
In RDBMS we do not have P (network
partitions)
Consistency and Availability are achieved
In NoSQL we want to have P
Need to select either C or A
Drop A -> Accept waiting until data is consistent
Drop C -> Accept getting inconsistent data
sometimes
NoSQL Systems and CAP
http://blog.nahurst.com/visual-guide-to-nosql-systems
ACID vs BASE
Scalability and better performance of NoSQL is
achieved by sacrificing ACID compatibility.
Atomic, Consistent, Isolated, Durable
NoSQL is having BASE compatibility instead.
Basically Available, Soft state,
Eventual consistency
ACID -- Requirement for SQL DBs
Atomicity. All of the operations in the transaction
will complete, or none will.
Consistency. Transactions never observe or
result in inconsistent data.
Isolation. The transaction will behave as if it is the
only operation being performed upon the database
(i.e. uncommitted transactions are isolated)
Durability. Upon completion of the transaction, the
operation will not be reversed (i.e. committed
transactions are permanent)
BASE -- Basically Available
Use replication and sharding to reduce the
likelihood of data unavailability and use
sharding, or partitioning the data among
many different storage servers, to make any
remaining failures partial.
The result is a system that is always
available, even if subsets of the data
become unavailable for short periods of
time.
BASE and Availability
The availability of BASE is achieved through
supporting partial failures without total
system failure.
Example. If users are partitioned across five database
servers, BASE design encourages crafting operations in
such a way that a user database failure impacts only the
20 percent of the users on that particular host.
This leads to higher perceived availability of the
system. Even though a single node is failing, the
interface is still operational.
BASE -- Eventually Consistent
Although applications must deal with
instantaneous consistency, NoSQL systems
ensure that at some future point in time the
data assumes a consistent state.
In contrast to ACID systems that enforce
consistency at transaction commit, NoSQL
guarantees consistency only at some
undefined future time.
Where ACID is pessimistic and forces consistency at
the end of every operation, BASE is optimistic and
accepts that the database consistency will be in a
state of flux.
BASE and Consistency
As DB nodes are added while scaling up, need
for synchronization arises
If absolute consistency is required, nodes
need to communicate when read/write
operations are performed on a node
Consistency over availability -> bottleneck
As a trade-off, "eventual consistency" is used
Consistency is maintained later
Numerous approaches for keeping up "distributed
consistency" are available
Amazon Dynamo - consistent hashing
CouchDB - asynchronous master-master replication
MongoDB - auto-sharding+replication cluster with a master server
BASE -- Soft State
While ACID systems assume that data
consistency is a hard requirement, NoSQL
systems allow data to be inconsistent and
relegate designing around such
inconsistencies to application developers.
In other words, soft state indicates that the
state of the system may change over time,
even without input.
This is because of the eventual consistency model
(the acronym is a bit contrived).
Some breeds of NoSQL solutions
Key-Value Stores
Column Family Stores
Document Databases
Graph Databases
In addition: Object and RDF databases as
well as Tuple stores
Key-Value Stores
Dynamo, Voldemort, Rhino DHT ...
DeCandia et al. "Dynamo: Amazons Highly Available Key-value Store", 2007
Key-Value is based on a hash table where
there is a unique key and a pointer to a
particular item of data.
Mappings are usually accompanied by
cache mechanisms to maximize
performance.
API is typically simple -- implementation is
often complex.
Column Family Stores
BigTable, Cassandra, HBase, Hadoop ...
Chang et al. "Bigtable: A Distributed Storage System for Structured Data", 2006
Store and process very large amounts of
data distributed over many machines.
"Petabytes of data across thousands of servers"
Keys point to multiple columns
Age
Name
Gender
Phone
25
Jim
123456
Age
Name
Gender
Phone
22
Jill
654321
jim_87
jill_90
Cassandra example
Document Databases (Stores)
CouchDB, MongoDB, Lotus Notes, Redis ...
Documents are addressed in the database
via a unique key that represents that
document.
Semi-structured documents can be XML or
JSON formatted, for instance.
In addition to the key, documents can be
retrieved with queries.
Redis is sometimes referred to as data
structure server since keys can contain
strings, hashes, lists, sets and sorted sets.
Graph Databases
Neo4J, FlockDB, GraphBase, InfoGrip, ...
Graph Databases are built with nodes,
relationships between nodes (edges) and the
properties of nodes.
Nodes represent entities (e.g. "Bob" or "Alice").
Similar in nature to the objects as in object-oriented
programming.
Properties are pertinent information related to nodes (e.
g. age: 18).
Edges connect nodes to nodes or nodes to properties.
Represent the relationship between the two.
Scaling graph DBs is problematic
Neo4J: cache sharding, sharding strategy heuristics
Some NoSQL Challenges
Lack of maturity -- numerous solutions still in
their beta stages
Lack of commercial support for enterprise
users
Lack of support for data analysis
Maintenance efforts and skills are required -experts are hard to find
References and Material
Michael Stonebraker, Samuel Madden, Daniel J. Abadi, Stavros Harizopoulos, Nabil Hachem, and Pat
Helland. 2007. The end of an architectural era: (it's time for a complete rewrite). InProceedings of the
33rd international conference on Very large data bases (VLDB '07). VLDB Endowment 1150-1160.
Werner Vogels. 2008. Eventually Consistent. Queue 6, 6 (October 2008), 14-19. DOI=10.
1145/1466443.1466448
Dan Pritchett. 2008. BASE: An Acid Alternative. Queue 6, 3 (May 2008), 48-55. DOI=10.1145/1394127.
1394128
Seth Gilbert and Nancy Lynch. 2002. Brewer's conjecture and the feasibility of consistent, available,
partition-tolerant web services. SIGACT News 33, 2 (June 2002), 51-59. DOI=10.1145/564585.564601
Fay Chang, Jeffrey Dean, Sanjay Ghemawat, Wilson C. Hsieh, Deborah A. Wallach, Mike Burrows,
Tushar Chandra, Andrew Fikes, and Robert E. Gruber. 2006. Bigtable: a distributed storage system for
structured data. In Proceedings of the 7th USENIX Symposium on Operating Systems Design and
Implementation - Volume 7 (OSDI '06), Vol. 7. USENIX Association, Berkeley, CA, USA, 15-15.
Avinash Lakshman and Prashant Malik. 2010. Cassandra: a decentralized structured storage system.
SIGOPS Oper. Syst. Rev. 44, 2 (April 2010), 35-40. DOI=10.1145/1773912.1773922
Couch DB - The Definitive Guide, http://guide.couchdb.org/index.html
HP whitepaper: There is no free lunch with distributed data
Long list of NoSQL papers: http://nosqlsummer.org/papers
Possible Presentation Topics
NoSQL architectures
Key-Value Store
Graph
Big Table (Columnar)
Document Store
...
Implementations
MongoDB
HBase
Cassandra
CouchDB
Google App Engine Datastore
Hadoop
BigData
Redis
Riak
Neo4j
Hosted services
Freebase
OpenLink Virtuoso
Datastore on Google Appengine
Amazon DynamoDB
Cloudant Data Layer (CouchDB)
...
Technologies / misc
MapReduce
Fault tolerance
Taxonomy
Challenges / Limitations
Tools
Use Cases
Distributed databases
Parallel systems
...
