Vargas polyglot-persistence-cloud-edbt

POLYGLOT PERSISTENCE AND MULTI-CLOUD DATA
MANAGEMENT SOLUTIONS
GENOVEVAVARGAS SOLAR
FRENCH COUNCIL OF SCIENTIFIC RESEARCH, LIG-LAFMIA, FRANCE
Genoveva.Vargas@imag.fr

http://www.vargas-‐solar.com

DATA ALL AROUND
2
30 billion
pieces of content
are shared on Facebook
every month
40 billion+
hours video
are watched onYouTube
each month
As of 2011 the global size
Data in healthcare was
estimated to be
150 Exabytes
(161 billion of Gigabytes)
40 millionTweets
are sent per day about 200
monthly active users
By 2014 it is anticipated
there will be
400 million
wearable wireless
health monitors
Web 2.0 sites where millions of users may both read and
write data, scalability for simple database operations has
become important
Data collections available through front ends managed
through public/private organization, available through the
Web, e.g., Sloan Sky Server

STORING AND ACCESSING HUGE AMOUNTS OF DATA
Peta
1015

Exa
1018

Zetta
1021

Yota
1024

RAID

Disk

Cloud

3
•  Data formats
•  Data collection sizes
•  Data storage supports
•  Data delivery mechanisms

THIS TALK IS NOT ABOUT
5
http://nosql-‐database.org

Debate on whether NoSQL stores and relational systems are better or worse …
that is not the point
Of course we can surf on these waves
at the end of the talk and during EDBT School!

THIS TALK IS ABOUT
6
alternative for managing multiform and multimedia data collections
according to different properties and requirements

AGENDA
¡  Dealing with multiform and multimedia data collections:
it’s all about management requirements
¡  Polyglot persistence general principle
¡  The NoSQL plethora
¡  Polyglot database
¡  Design
¡  Deployment
¡  Querying, update and maintenance
¡  Conclusions and outlook
7

POLYGLOT PERSISTENCE
¡  Polyglot Programming: applications should be written in a mix of languages to take
advantage of different languages are suitable for tackling different problems
¡  Polyglot persistence: any decent sized enterprise will have a variety of different data
storage technologies for different kinds of data
¡  a new strategic enterprise application should no longer be built assuming a relational
persistence support
¡  the relational option might be the right one - but you should seriously look at other
alternatives
9
M.
Fowler
and
P.
Sadalage.
NoSQL
Distilled:
A
Brief
Guide
to
the
Emerging
World
of
Polyglot
Persistence.
Pearson
Education,
Limited,
2012

(Katsov-2012)
Use the right tool for the right job…
How do I know which is the
right tool for the right job?
10

SCALING DATABASE SYSTEMS
¡  A system is scalable if increasing its resources
(CPU, memory, disk) results in increased
performance proportionally to the added
resources
¡  Improving performance means serving more units
of work for handling larger units of work like when
data sets grow
¡  Database systems have been scaled by buying
bigger faster and more expensive machines
11
¡  Vertically (SCALE UP)
¡  Add resources (CPU, memory) to a single node in a system
¡  Horizontally (SCALE OUT)
¡  Add more nodes to a system

NOSQL STORES CHARACTERISTICS
¡  Simple operations
¡  Key lookups reads and writes of one record or a small number of records
¡  No complex queries or joins
¡  Ability to dynamically add new attributes to data records
¡  Horizontal scalability
¡  Distribute data and operations over many servers
¡  Replicate and distribute data over many servers
¡  No shared memory or disk
¡  High performance
¡  Efficient use of distributed indexes and RAM for data storage
¡  Weak consistency model
¡  Limited transactions
12
Next generation databases mostly addressing some of the points: being non-relational, distributed,
open-source and horizontally scalable [http://nosql-‐database.org]

DEALING WITH HUGE AMOUNTS OF DATA
13
Peta
1015

Exa
1018

Zetta
1021

Yota
1024

RAID

Disk

Cloud

Concurrency

Consistency

Atomicity

Relational

Graph

Key
value

Columns

DATA MANAGEMENT SYSTEMS ARCHITECTURES
Physical
model

Logic
model

External
model

ANSI/SPARC

Storage

Manager

Schema

Manager

Query

Engine

Transaction

Manager

DBMS

Customisable
points

Custom
components

Glue

code

Data

services

Access

services

Storage

services

Additional

Extension

services

Other

services

Extension
services

Streaming, XML, procedures,
queries, replication
Physical
model

Logic
model

External
model

Physical
model

Logic
model

External
model

14

15
Data
stores
designed

to
scale
simple

OLTP-‐style
applica7on
loads

•  Data
model

•  Consistency

•  Storage

•  Durability

•  Availability

•  Query
support

Read/Write
operations

by
thousands/millions
of
users

PROBLEM STATEMENT: HOW MUCH TO GIVE UP?
¡  CAP theorem1: a system can have two of the three properties
¡  NoSQL systems sacrifice consistency
16
Consistency
Availability

Fault-‐tolerant

partitioning

1
Eric
Brewer,
"Towards
robust
distributed
systems."
PODC.
2000
http://www.cs.berkeley.edu/~brewer/cs262b-‐2004/PODC-‐keynote.pdf

NOSQL STORES:AVAILABILITY AND PERFORMANCE
¡  Replication
¡  Copy data across multiple servers (each bit of data can be found in
multiple servers)
¡  Increase data availability
¡  Faster query evaluation
¡  Sharding
¡  Distribute different data across multiple servers
¡  Each server acts as the single source of a data subset
¡  Orthogonal techniques
17

REPLICATION: PROS & CONS
¡  Data is more available
¡  Failure of a site containing E does not result in
unavailability of E if replicas exist
¡  Performance
¡  Parallelism: queries processed in parallel on several
nodes
¡  Reduce data transfer for local data
¡  Increased updates cost
¡  Synchronisation: each replica must be updated
¡  Increased complexity of concurrency control
¡  Concurrent updates to distinct replicas may lead to
inconsistent data unless special concurrency control
mechanisms are implemented
18

SHARDING:WHY IS IT USEFUL?
¡  Scaling applications by reducing data sets in any single
databases
¡  Segregating data
¡  Sharing application data
¡  Securing sensitive data by isolating it
¡  Improve read and write performance
¡  Smaller amount of data in each user group implies faster querying
¡  Isolating data into smaller shards accessed data is more likely to
stay on cache
¡  More write bandwidth: writing can be done in parallel
¡  Smaller data sets are easier to backup, restore and manage
¡  Massively work done
¡  Parallel work: scale out across more nodes
¡  Parallel backend: handling higher user loads
¡  Share nothing: very few bottlenecks
¡  Decrease resilience improve availability
¡  If a box goes down others still operate
¡  But: Part of the data missing
19
Load%balancer%
Cache%1%
Cache%2%
Cache%3%
MySQL%
Master%
MySQL%
Master%
Web%1%
Web%2%
Web%3%
Site%database%
Resume%database%

SHARDING AND REPLICATION
¡  Sharding with no replication: unique copy, distributed data sets
¡  (+) Better concurrency levels (shards are accessed independently)
¡  (-) Cost of checking constraints, rebuilding aggregates
¡  Ensure that queries and updates are distributed across shards
¡  Replication of shards
¡  (+) Query performance (availability)
¡  (-) Cost of updating, of checking constraints, complexity of concurrency control
¡  Partial replication (most of the times)
¡  Only some shards are duplicated
20

NOSQL STORES: DATA MANAGEMENT PROPERTIES
¡  Indexing
¡  Distributed hashing like Memcached open source cache
¡  In-memory indexes are scalable when distributing and
replicating objects over multiple nodes
¡  Partitioned tables
¡  High availability and scalability: eventual consistency
¡  Data fetched are not guaranteed to be up-to-date
¡  Updates are guaranteed to be propagated to all nodes
eventually
¡  Shared nothing horizontal scaling
¡  Replicating and partitioning data over many servers
¡  Support large number of simple read/write operations
per second (OLTP)
¡  No ACID guarantees
¡  Updates eventually propagated but limited guarantees
on reads consistency
¡  BASE: basically available; soft state, eventually consistent
¡  Multi-version concurrency control
21

COMPARING NOSQL & NEWSQL SYSTEMS
SYSTEM CONCURRENCY
CONTROL
DATA
STORAGE
REPLICATION TRANSACTION
Redis Locks RAM Asynchronous No
Scalaris Locks RAM Synchronous Local
Tokyo Locks RAM/Disk Asynchronous Local
Voldemort MVCC RAM/BDB Asynchronous No
Riak MVCC Plug in Asynchronous No
Membrain Locks Flash+Disk Synchronous Local
Membase Locks Disk Synchronous Local
Dynamo MVCC Plug in Asynchronous No
SimpleDB Non S3 Asynchronous No
MongoDB Locks Disk Asynchronous No
CouchDB MVCC Disk Asynchronous No
22
SYSTEM CONCURRENCY
CONTROL
DATA
STORAGE
REPLICATION TRANSACTION
Terrastore Locks RAM+ Synchronous L
Hbase Locks HADOOP Asynchronous L
HyperTable Locks Files Synchronous L
Cassandra MVCC Disk Asynchronous L
BigTable Locs+stamps GFS Both L
PNuts MVCC Disk Asynchronous L
MySQL-C ACID Disk Synchronous Y
VoltDB ACID/no Lock RAM Synchronous Y
Clustrix ACID/no Lock Disk Synchronous Y
ScaleDB ACID Disk Synchronous Y
ScaleBase ACID Disk Asynchronous Y
NimbusDB ACID/no Lock Disk Synchronous Y
Key-‐Value
Document

Extended
records
Relational

Cattell,
Rick.
"Scalable
SQL
and
NoSQL
data
stores."
ACM
SIGMOD
Record
39.4
(2011):
12-‐27

AGENDA
ü  Dealing with multiform and multimedia data collections:
ü  it’s all about management requirements
ü  Polyglot persistence general principle
ü  The NoSQL plethora
¡  Polyglot database
¡  Design
¡  Deployment
¡  Querying, update and maintenance
23

DESIGNING AND BUILDING A POLYGLOT DATABASE
24

25
OBJECTIVE
¡  Build a MyNet app based on a polyglot database for building an integrated directory
of my contacts including their status and posts from several social networks
MyNet
DB
MyNet
App
Social
network

26
Analysis on contacts networks, overlapping
according to interests, posts topics
Top 10 most popular contacts
User sessions in different
Social networks
Integrating posts
from all networks
Contact graph traversal
For building groups out of
Common characteristics
Synchronizing posts to all
SNFriends network
User accounts activity
In different social networks
Directory synchronisation
Integrating contacts’ information
From all SN

NOSQL DESIGN AND CONSTRUCTION PROCESS
¡  Data reside in RAM (memcached) and is eventually replicated and stored
¡  Querying = designing a database according to the type of queries / map reduce model
¡  “On demand” data management: the database is virtually organized per view (external schema) on cache and some
view are made persistent
¡  An elastic easy to evolve and explicitly configurable architecture 27
Database
querying
Database
population
Database
organization
INDEX
Memcached
Replicated
Stored

GENERATING NOSQL PROGRAMS FROM HIGH LEVEL
ABSTRACTIONS
28
UML class diagram application classes
Spring Roo
Java

web

App
Spring Data
Graph database
Relational database
High-level abstractions
Low-level abstractions
http://code.google.com/p/model2roo/

DEPLOYING A POLYGLOT DATABASE
29

DEPLOYING A DATABASE APPLICATION ON THE CLOUD
Have a first experience on
¡  creating a database on a relational DBMS deployed on the cloud,
¡  building a simple database application exported as a service,
¡  deploying the service on the cloud and implement an interface.
30
MyNet
DB

MyNet
App

Contact'
idContact:"Integer"
firstName:"String"
lastName:"String"
Service

App

SHARDING AND DEPLOYING SHARDS ON THE CLOUD
Have a first experience on
¡  sharding a relational DBMS
¡  dealing with shards synchronization
¡  deploying the service on the cloud and implement an interface
31
MyNet
DB

MyNet

Service

webSite:(URI(
socialNetworkID:(URI((
(
Basic(Info(
Contact'
idContact:(Integer(
lastName:(String(
givenName:(String(
society:(String( <<'hasBasicInfo'>>'
1' *'
postID:(Integer(
timeStamp:(Date(
geoStamp:(String(
Post'
contentID:(Integer(
text:(String(
Content'
<<'hasContent'>>'
<<publishesPost>>'
*'
1'
1' 1'
street:(String,((
number:(Integer,(
City:(Sting,((
Zipcode:(Integer(
Address'
*'
email:(String(
type:({pers,(prof}(
Email'
<<'hasAddress'>>'
*'
1'
<<'hasEmail'>>'
ID:(Integer,(
Lang:(String(
Language'
*'
<<'speaksLanguage'>>'
Français

Español

English

Others

REST

JSON
documents

MyNet

MULTI-CLOUD POLYGLOT DATABASE
MyNetContacts

webSite:(URI(
(
Basic(Info(
Contact'
idContact:(Integer(
lastName:(String(
givenName:(String(
society:(String(
<<'hasBasicInfo'>>'
<<'isContactof'>>'
groupName:(String(
Group'
<<'isComposedof'>>'
1' *'
1'
*'1'
*'
postID:(Integer(
timeStamp:(Date(
geoStamp:(String(
Post'
contentID:(Integer(
text:(String(
image:(Jpeg(
video:(Avi(
Content'
<<'hasContent'>>'
<<'publishesPost'>>'
*'
1'
1' 1'
street:(String,((
number:(Integer,(
City:(Sting,((
Zipcode:(Integer(
Address'
1'
*'
number:(String(
type:({pers,(prof}(
Phone'
email:(String(
type:({pers,(prof}(
Email'
<<'hasAddress'>>'
<<'hasOtherPhones'>>'
1'
*'
1'
*'
*'
<<'hasMobilePhones'>>'
1'
<<'hasEmail'>>'
webSite:(URI(
(
Basic(Info(
Contact'
idContact:(Integer(
lastName:(String(
givenName:(String(
society:(String(
<<'hasBasicInfo'>>'
<<'isContactof'>>'
groupName:(String(
Group'
<<'isComposedof'>>'
1' *'
1'
*'1'
*'
postID:(Integer(
timeStamp:(Date(
geoStamp:(String(
Post'
contentID:(Integer(
text:(String(
image:(Jpeg(
video:(Avi(
Content'
<<'hasContent'>>'
*'
1'
1' 1'
street:(String,((
number:(Integer,(
City:(Sting,((
Zipcode:(Integer(
Address'
1'
*'
number:(String(
type:({pers,(prof}(
Phone'
email:(String(
type:({pers,(prof}(
Email'
<<'hasAddress'>>'
1'
*'
1'
*'
*'
1'
<<'hasEmail'>>'
webSite:(URI(
(
Basic(Info(
Contact'
idContact:(Integer(
lastName:(String(
givenName:(String(
society:(String(
<<'hasBasicInfo'>>'
<<'isContactof'>>'
groupName:(String(
Group'
<<'isComposedof'>>'
1' *'
1'
*'1'
*'
postID:(Integer(
timeStamp:(Date(
geoStamp:(String(
Post'
contentID:(Integer(
text:(String(
image:(Jpeg(
video:(Avi(
Content'
<<'hasContent'>>'
*'
1'
1' 1'
street:(String,((
number:(Integer,(
City:(Sting,((
Zipcode:(Integer(
Address'
1'
*'
number:(String(
type:({pers,(prof}(
Phone'
email:(String(
type:({pers,(prof}(
Email'
<<'hasAddress'>>'
1'
*'
1'
*'
*'
1'
<<'hasEmail'>>'

MANAGING A POLYGLOT DATABASE
QUERYING, INSERTING, MAINTAINING
33

CRUD OPERATIONS
34
Contact
Group
Content
Post
Id
FirstName
LastName
Society

Consistent view of data

EXAMPLE 1: SYNCHRONIZING REDIS+MYSQL
35
https://oracleus.activeevents.com/connect/sessionDetail.ww?SESSION_ID=4775

Updating REDIS #FAIL
begin
MySQL
transaction

update
MySQL

update
Redis

rollback
MySQL
transaction

begin
MySQL
transaction

update
MySQL

commit
MySQL
transaction

<<
system
crashes
>>

update
Redis

Redis has updated MySQL does not
MySQL has updated Redis does not

EXAMPLE 1: UPDATING REDIS RELIABLY
Step I Step 2
36
begin
MySQL
transaction

update
MySQL

queue
CRUD
event
in
MySQL

commit
transaction

Event
Id

Operation:
Create,
Update,
Delete

queue
CRUD
event
in
MySQL

New
entity
state,
e.g.
JSON

ACID
for
each
CRUD
event
in
MySQL
queue

get
next
CRUD
event
from
MySQL
queue

if
CRUD
event
is
not
duplicate
then

update
Redis
(incl.
eventID)

end
if

begin
MySQL
transaction

mark
CRUD
event
processed

commit
transaction

end
for
each

EXAMPLE 1: UPDATING REDIS RELIABLY
37
EntityCRUDEvent

Repository

EntityCRUDEvent

Processor

Redis
updater

ID
JSON
Processed?

INSERT
INTO
..
SELECT
…
FROM..

apply(event)

Timer

Step 1 Step 2

EXAMPLE 1:TRACKING CHANGES (HIBERNATE)
38

POLYGLOT DATABASE EVOLUTION
¡  Problem statement:
¡  Evolution of the application: modification of classes, new
classes, new relationships among classes
¡  Evolution of the “entities” managed in the polyglot
database
¡  Some change structure, change values, …
¡  The content of the stores start deriving from the
application data structures
¡  Which is the current structure of the entities stored?
¡  Are there elements that are not being accessed because
they do not longer correspond to the application data
structures?
39
Contact'
idContact:"Integer"
firstName:"String"
lastName:"String"
webSite:(URI(
(
Basic(Info(
Contact'
idContact:(Integer(
lastName:(String(
givenName:(String(
society:(String( <<'hasBasicInfo'>>'
1' *'
postID:(Integer(
timeStamp:(Date(
geoStamp:(String(
Post'
contentID:(Integer(
text:(String(
Content'
<<'hasContent'>>'
<<publishesPost>>'
*'
1'
1' 1'
street:(String,((
number:(Integer,(
City:(Sting,((
Zipcode:(Integer(
Address'
*'
email:(String(
type:({pers,(prof}(
Email'
<<'hasAddress'>>'
*'
1'
<<'hasEmail'>>'
ID:(Integer,(
Lang:(String(
Language'
*'
<<'speaksLanguage'>>'

EXTRACTING SCHEMATA FROM NOSQL PROGRAMS CODE
http://code.google.com/p/exschema/

40
Metalayer(
Struct( Relationship(
Attribute(Set(
*( *(
*(
*(
*( *(
*(
*(*(
*(
*(
*(
*(
*(
Declarations

analyzer

Updates

analyzer

Repositories

analyzer

Schema1 Schema2
Schema3
Set
Attribute
implementation : Spring Repository
Set
Attribute
name : fr.imag.twitter.domain.UserInfo
Struct
Attribute
name : userId
Attribute
name : firstName
Attribute
name : lastName
Set
Attribute
Set
Attribute
name : fr.imag.twitter.domain.Tweet
Struct
Attribute
name : text
Attribute
name : userId
Set
Attribute
implementation : Neo4j
Struct
Attribute
name : fr.imag.twitter.domain.User
Str
Attribute
name : nodeId
Attribute
name : userName
A
na
Relatio
start end
Schema1 Schema2
Set
Attribute
Set
Attribute
Struct
Attribute
name : userId
Attribute
name : firstName
Attribute
name : lastName
Set
Attribute
Attribute
Attribute
name : text
Attri
name :
Schema1 Schema2
Schema3
Set
Attribute
Set
Attribute
Struct
Attribute
name : userId
Attribute
name : firstName
Attribute
name : lastName
Set
Attribute
Set
Attribute
Struct
Attribute
name : text
Attribute
name : userId
Set
Attribute
implementation : Neo4j
Struct
Attribute
name : fr.imag.twitter.domain.User
Struct
Attribute
name : nodeId
Attribute
name : userName
Attribute
name : userId
Attribute
name : password
Relationship
start end
Attribute
name : followers
Spring Data

EXTRACTING SCHEMATA FROM A POLYGLOT DATABASE
41
Application
source
code

Metalayer(
Struct( Relationship(
Attribute(Set(
*( *(
*(
*(
*( *(
*(
*(*(
*(
*(
*(
*(
*(
Declarations

Analyzer

Updates

Analyzer

Annotations

Analyzer

Repositories

Analyzer

ExSchema

Variables/Methods
Graph
Store

Annotations/Variables

Relational
&

document
stores

AGENDA
ü  Dealing with multiform and multimedia data collections:
ü  it’s all about management requirements
ü  Polyglot persistence general principle
ü  The NoSQL plethora
ü  Polyglot database
ü  Design
ü  Deployment
ü  Querying, update and maintenance
42

CHALLENGE: POLYGLOT MEETS XPERANTO
Given a data collection coming from different social networks stored on NoSQL systems (Neo4J and Mongo) [possibly
according to a strategy combining sharding and replication techniques], extend the UnQL pivot query language considering
¡  Data processing operators adapted to query different data models (graph, document). Example query Neo + Mongo
and what about Join, Union …
¡  Assuming concurrent CRUD operations to the stores can you expect query results to be consistent ? How can you
tag your results or implement a sharding strategy in order to determine whether results are consistent?
¡  Querying data represented on different models: How can you exploit the structure of the different stores for
expressing queries ? Provide adapted operators? Give generic operators and then rewrite queries?
¡  Normally, Polyglot solutions tend to solve some data processing issues in the application code.This can be penalizing.
Discuss the challenges to address for ensuring that your queries will be able to scale as the collection grows.
43

CHALLENGE: EXPECTED RESULTS
¡  Give the principle of your proposal through a partial programming solution, of the operators of your
UnQL extension, detail the query evaluation process if U want your solution to scale
¡  We ask U to sketch the solution on the polyglot database that we provide consisting of mongo, Neo4J stores
¡  https://github.com/jccastrejon/edbt-unql
¡  Technical requirements:VMware player 5
44

WHEN IS POLYGLOT PERSISTENCE PERTINENT?
¡  Application essentially composing and serving web pages
¡  They only looked up page elements by ID, they had different needs or availability,
concurrency and no need to share all their data
¡  A problem like this is much better suited to a NoSQL store than the corporate relational
DBMS
¡  Scaling to lots of traffic gets harder and harder to do with vertical scaling
¡  Many NoSQL databases are designed to operate over clusters
¡  They can tackle larger volumes of traffic and data than is realistic with a single server
45

CONCLUSIONS
¡  Data are growing big and more heterogeneous and they need new adapted ways to be managed thus the
NoSQL movement is gaining momentum
¡  Data heterogeneity implies different management requirements this is where polyglot persistence comes up
¡  Consistency – Availability – Fault tolerance theorem: find the balance !
¡  Which data store according to its data model?
¡  A lot of programming implied …
46
Open opportunities if you’re interested in this topic!

Contact: GenovevaVargas-Solar, CNRS, LIG-LAFMIA
Genoveva.Vargas@imag.fr

http://www.vargas-‐solar.com/teaching/

http://code.google.com/p/exschema/

47
http://code.google.com/p/model2roo/

Open source polyglot persistence tools

Prof.
Dr.
Christine
Collet

Grenoble
INP

France

Juan
Carlos
Castrejón

University
of
Grenoble

France

Javier
Espinosa

University
of
Grenoble

France

Dr.
Genoveva
Vargas-‐Solar

CNRS,
LIG-‐LAFMIA

France

REFERENCES
¡  Eric
A.,
Brewer
"Towards
robust
distributed
systems."
PODC.
2000

¡  Rick,
Cattell
"Scalable
SQL
and
NoSQL
data
stores."
ACM
SIGMOD
Record
39.4
(2011):
12-‐27

¡  Juan
Castrejon,
Genoveva
Vargas-‐Solar,
Christine
Collet,
and
Rafael
Lozano,
ExSchema:

Discovering
and
Maintaining
Schemas
from
Polyglot
Persistence
Applications,
In
Proceedings
of

the
International
Conference
on
Software
Maintenance,
Demo
Paper,
IEEE,
2013

¡  M.
Fowler
and
P.
Sadalage.
NoSQL
Distilled:
A
Brief
Guide
to
the
Emerging
World
of
Polyglot

Persistence.
Pearson
Education,
Limited,
2012

¡  C.
Richardson,
Developing
polyglot
persistence
applications,
http://fr.slideshare.net/
chris.e.richardson/developing-‐polyglotpersistenceapplications-‐gluecon2013

49

DATA MODELS
¡  Tuple
¡  Row in a relational table, where attributes are pre-defined in a schema, and the values are scalar
¡  Document
¡  Allows values to be nested documents or lists, as well as scalar values.
¡  Attributes are not defined in a global schema
¡  Extensible record
¡  Hybrid between tuple and document, where families of attributes are defined in a schema, but new attributes can be added
on a per-record basis
51

DATA STORES
¡  Key-value
¡  Systems that store values and an index to find them, based on a key
¡  Document
¡  Systems that store documents, providing index and simple query mechanisms
¡  Extensible record
¡  Systems that store extensible records that can be partitioned vertically and horizontally across nodes
¡  Graph
¡  Systems that store model data as graphs where nodes can represent content modelled as document or key-value structures and arcs
represent a relation between the data modelled by the node
¡  Relational
¡  Systems that store, index and query tuples
52

KEY-VALUE STORES
¡  “Simplest data stores” use a data model similar to
the memcached distributed in-memory cache
¡  Single key-value index for all data
¡  Provide a persistence mechanism
¡  Replication, versioning, locking, transactions, sorting
¡  API: inserts, deletes, index lookups
¡  No secondary indices or keys
53
SYSTEM ADDRESS
Redis code.google.com/p/redis

Scalaris code.google.com/p/scalaris

Tokyo tokyocabinet.sourceforge.net

Voldemort project-‐voldemort.com

Riak riak.basho.com

Membrain schoonerinfotech.com/products

Membase membase.com

SELECT

name

FROM

group

WHERE

gid
IN
(
SELECT

gid

FROM

group_member

WHERE

uid
=
me()
)

54
SELECT

name,
pic,
profile_url

FROM

user

WHERE

uid
=
me()

SELECT

name,
pic

FROM

user

WHERE

online_presence
=
"active"

AND

uid
IN
(
SELECT

uid2

FROM

friend

WHERE

uid1
=
me()
)

SELECT

name

FROM

friendlist

WHERE

owner
=
me()

SELECT

message,
attachment

FROM

stream

WHERE

source_id
=
me()
AND
type
=
80

https://developers.facebook.com/docs/reference/fql/

DOCUMENT STORES
¡  Support more complex data: pointerless objects, i.e.,
documents
¡  Secondary indexes, multiple types of documents
(objects) per database, nested documents and lists, e.g.
B-trees
¡  Automatic sharding (scale writes), no explicit locks,
weaker concurrency (eventual for scaling reads) and
atomicity properties
¡  API: select,
delete,
getAttributes,

putAttributes
on documents
¡  Queries can be distributed in parallel over multiple
nodes using a map-reduce mechanism
56
SYSTEM ADDRESS
SimpleDB amazon.com/simpledb

Couch DB couchdb.apache.org

Mongo DB mongodb.org

Terrastore code.google.com/terrastore

EXTENSIBLE RECORD STORES
¡  Basic data model is rows and columns
¡  Basic scalability model is splitting rows and columns over
multiple nodes
¡  Rows split across nodes through sharding on the primary key
¡  Split by range rather than hash function
¡  Rows analogous to documents: variable number of attributes, attribute
names must be unique
¡  Grouped into collections (tables)
¡  Queries on ranges of values do not go to every node
¡  Columns are distributed over multiple nodes using “column
groups”
¡  Which columns are best stored together
¡  Column groups must be pre-defined with the extensible record
stores
58
SYSTEM ADDRESS
HBase hbase.apache.com

HyperTable hypertable.org

Cassandra incubator.apache.org/cassandra

SCALABLE RELATIONAL SYSTEMS
¡  SQL: rich declarative query language
¡  Databases reinforce referential integrity
¡  ACID semantics
¡  Well understood operations:
¡  Configuration, Care and feeding, Backups,Tuning, Failure and recovery,
Performance characteristics
¡  Use small-scope operations
¡  Challenge: joins that do not scale with sharding
¡  Use small-scope transactions
¡  ACID transactions inefficient with communication and 2PC overhead
¡  Shared nothing architecture for scalability
¡  Avoid cross-node operations
59
SYSTEM ADDRESS
MySQL C mysql.com/cluster

Volt DB voltdb.com

Clustrix clustrix.com

ScaleDB scaledb.com

Scale Base scalebase.com

Nimbus DB nimbusdb.com

NOSQL STORES:AVAILABILITY AND PERFORMANCE
60

REPLICATION MASTER - SLAVE
¡  Makes one node the authoritative copy/replica that handles
writes while replica synchronize with the master and may
handle reeds
¡  All replicas have the same weight
¡  Replicas can all accept writes
¡  The lose of one of them does not prevent access to the data store
¡  Helps with read scalability but does not help with write
scalability
¡  Read resilience: should the master fail, slaves can still handle
read requests
¡  Master failure eliminates the ability to handle writes until
either the master is restored or a new master is appointed
¡  Biggest complication is consistency
¡  Possible write – write conflict
¡  Attempt to update the same record at the same time from to
different places
¡  Master is a bottle-neck and a point of failure
61
Master'
Slaves'
all'updates'
made'to'the'master'
changes'propagate''
To'slaves'
reads'can'be'done'
from'master'or'slaves'

MASTER-SLAVE REPLICATION MANAGEMENT
¡  Masters can be appointed
¡  Manually when configuring the nodes cluster
¡  Automatically: when configuring a nodes cluster one of them elected as master.The master can appoint a new master when the master
fails reducing downtime
¡  Read resilience
¡  Read and write paths have to be managed separately to handle failure in the write path and still reads can occur
¡  Reads and writes are put in different database connections if the database library accepts it
¡  Replication comes inevitably with a dark side: inconsistency
¡  Different clients reading different slaves will see different values if changes have not been propagated to all slaves
¡  In the worst case a client cannot read a write it just made
¡  Even if master-slave is used for hot backups, if the master fails any updates on to the backup are lost
62

REPLICATION: PEER-TO-PEER
¡  Allows writes to any node; the nodes coordinate
to synchronize their copies
¡  The replicas have equal weight
¡  Deals with inconsistencies
¡  Replicas coordinate to avoid conflict
¡  Network traffic cost for coordinating writes
¡  Unnecessary to make all replicas agree to write, only
the majority
¡  Survival to the loss of the minority of replicas nodes
¡  Policy to merge inconsistent writes
¡  Full performance on writing to any replica
63
Master'
nodes'communicate'
their'writes'
all'nodes'read'
and'write'all'data'

SHARDING
¡  Ability to distribute both data and load of simple
operations over many servers, with no RAM or disk
shared among servers
¡  A way to horizontally scale writes
¡  Improve read performance
¡  Application/data store support
64
¡  Puts different data on separate nodes
¡  Each user only talks to one servicer so she gets rapid
responses
¡  The load should be balanced out nicely between
servers
¡  Ensure that
¡  data that is accessed together is clumped together on
the same node
¡  that clumps are arranged on the nodes to provide best
data access
Each%shard%reads%and%
writes%its%own%data%

SHARDING
Database laws
¡  Small databases are fast
¡  Big databases are slow
¡  Keep databases small
Principle
¡  Start with a big monolithic database
¡  Break into smaller databases
¡  Across many clusters
¡  Using a key value
65
Instead of having one million customers information
on a single big machine ….
100 000 customers on smaller and different machines

SHARDING CRITERIA
¡  Partitioning
¡  Relational: handled by the DBMS (homogeneous DBMS)
¡  NoSQL: based on ranging of the k-value
¡  Federation
¡  Relational
¡  Combine tables stored in different physical databases
¡  Easier with denormalized data
¡  NoSQL:
¡  Store together data that are accessed together
¡  Aggregates unit of distribution
66

SHARDING
Architecture
¡  Each application server (AS) is running DBS/client
¡  Each shard server is running
¡  a database server
¡  replication agents and query agents for supporting
parallel query functionality
Process
¡  Pick a dimension that helps sharding easily (customers,
countries, addresses)
¡  Pick strategies that will last a long time as repartition/
re-sharding of data is operationally difficult
¡  This is done according to two different principles
¡  Partitioning: a partition is a structure that divides a space
into tow parts
¡  Federation: a set of things that together compose a
centralized unit but each individually maintains some aspect
of autonomy
67
Customers data is partitioned by ID in shards using an
algorithm d to determine which shard a customer ID belongs to

REPLICATION:ASPECTS TO CONSIDER
¡  Conditioning
¡  Important elements to consider
¡  Data to duplicate
¡  Copies location
¡  Duplication model (master – slave / P2P)
¡  Consistency model (global – copies)
68
Fault

tolerance

Availability

Transparency

levels

Performance

à Find a compromise !

PARTITIONING
A PARTITION IS A STRUCTURE THAT DIVIDES A SPACE INTO TOW PARTS
69

BACKGROUND: DISTRIBUTED RELATIONAL DATABASES
¡  External schemas (views) are often subsets of relations
(contacts in Europe and America)
¡  Access defined on subsets of relations: 80% of the
queries issued in a region have to do with contacts of
that region
¡  Relations partition
¡  Better concurrency level
¡  Fragments accessed independently
¡  Implications
¡  Check integrity constraints
¡  Rebuild relations
70

FRAGMENTATION
¡  Horizontal
¡  Groups of tuples of the same relation
¡  Budget < 300 000 or >= 150 000
¡  Not disjoint are more difficult to manage
¡  Vertical
¡  Groups attributes of the same relation
¡  Separate budget from loc and pname of the relation
project
¡  Hybrid
71

FRAGMENTATION: RULES
Vertical
¡  Clustering
¡  Grouping elementary fragments
¡  Budget and location information in two relations
¡  Splitting
¡  Decomposing a relation according to affinity
relationships among attributes
Horizontal
¡  Tuples of the same fragment must be statistically
homogeneous
¡  If t1 and t2 are tuples of the same fragment then t1 and t2 have
the same probability of being selected by a query
¡  Keep important conditions
¡  Complete
¡  Every tuple (attribute) belongs to a fragment (without information
loss)
¡  If tuples where budget >= 150 000 are more likely to be selected
then it is a good candidate
¡  Minimum
¡  If no application distinguishes between budget >= 150 000 and
budget < 150 000 then these conditions are unnecessary
72

SHARDING: HORIZONTAL PARTITIONING
¡  The entities of a database are split into two or
more sets (by row)
¡  In relational: same schema several physical bases/
servers
¡  Partition contacts in Europe and America shards where
they zip code indicates where the will be found
¡  Efficient if there exists some robust and implicit way to
identify in which partition to find a particular entity
¡  Last resort shard
¡  Needs to find a sharding function: modulo, round robin,
hash – partition, range - partition
73
Load%balancer%
Cache%1%
Cache%2%
Cache%3%
MySQL%
Master%
MySQL%
Master%
Web%1%
Web%2%
Web%3%
Even%IDs%
Odd%IDs%
MySQL%
Slave%1% MySQL%
Slave%2%
MySQL%
Slave%n%
MySQL%
Slave%1% MySQL%
Slave%2%
MySQL%
Slave%n%

FEDERATION
A FEDERATION IS A SET OF THINGS THAT TOGETHER COMPOSE A CENTRALIZED UNIT BUT EACH
INDIVIDUALLY MAINTAINS SOME ASPECT OF AUTONOMY
74

FEDERATION:VERTICAL SHARDING
¡  Principle
¡  Partition data according to their logical affiliation
¡  Put together data that are commonly accessed
¡  The search load for the large partitioned entity can
be split across multiple servers (logical and
physical) and not only according to multiple
indexes in the same logical server
¡  Different schemas, systems, and physical bases/
servers
¡  Shards the components of a site and not only data
75
Load%balancer%
Cache%1%
Cache%2%
Cache%3%
MySQL%
Master%
MySQL%
Master%
Web%1%
Web%2%
Web%3%
Site%database%
Resume%database%
MySQL%
Slave%1% MySQL%
Slave%2%
MySQL%
Slave%n%
MySQL%
Slave%1%
Internal%
user%

NOSQL STORES: PERSISTENCY MANAGEMENT
76

«MEMCACHED»
¡  «memcached» is a memory management protocol based on a cache:
¡  Uses the key-value notion
¡  Information is completly stored in RAM
¡  «memcached» protocol for:
¡  Creating, retrieving, updating, and deleting information from the database
¡  Applications with their own «memcached» manager (Google, Facebook,
YouTube, FarmVille,Twitter,Wikipedia)
77

STORAGE ON DISC (1)
¡  For efficiency reasons, information is stored using the RAM:
¡  Work information is in RAM in order to answer to low latency requests
¡  Yet, this is not always possible and desirable
Ø  The process of moving data from RAM to disc is called "eviction”; this
process is configured automatically for every bucket
78

STORAGE ON DISC (2)
¡  NoSQL servers support the storage of key-value pairs on disc:
¡  Persistency–can be executed by loading data, closing and reinitializing it
without having to load data from another source
¡  Hot backups– loaded data are sotred on disc so that it can be
reinitialized in case of failures
¡  Storage on disc– the disc is used when the quantity of data is higher
thant the physical size of the RAM, frequently used information is
maintained in RAM and the rest es stored on disc
79

STORAGE ON DISC (3)
¡  Strategies for ensuring:
¡  Each node maintains in RAM information on the key-value pairs it stores. Keys:
¡  may not be found, or
¡  they can be stored in memory or on disc
¡  The process of moving information from RAM to disc is asynchronous:
¡  The server can continue processing new requests
¡  A queue manages requests to disc
Ø  In periods with a lot of writing requests, clients can be notified that the server is
termporaly out of memory until information is evicted
80

NOSQL STORES: CONCURRENCY CONTROL
81

MULTIVERSION CONCURRENCY CONTROL (MVCC)
¡  Objective: Provide concurrent access to the database and in programming languages to implement transactional
memory
¡  Problem: If someone is reading from a database at the same time as someone else is writing to it, the reader could
see a half-written or inconsistent piece of data.
¡  Lock: readers wait until the writer is done
¡  MVCC:
¡  Each user connected to the database sees a snapshot of the database at a particular instant in time
¡  Any changes made by a writer will not be seen by other users until the changes have been completed (until the transaction has been
committed
¡  When an MVCC database needs to update an item of data it marks the old data as obsolete and adds the newer version elsewhere à
multiple versions stored, but only one is the latest
¡  Writes can be isolated by virtue of the old versions being maintained
¡  Requires (generally) the system to periodically sweep through and delete the old, obsolete data objects
82

Vargas polyglot-persistence-cloud-edbt

More Related Content

What's hot (19)

Similar to Vargas polyglot-persistence-cloud-edbt (20)

More from Genoveva Vargas-Solar (10)

Recently uploaded (20)

Vargas polyglot-persistence-cloud-edbt