Storing and processing data with the wso2 platform

Storing and processing data
with the WSO2 Platform
Deependra Ariyadewa
Wathsala Vithanage

WSO2
• Founded in 2005 by acknowledged leaders in XML, Web
Services Technologies & Standards and Open Source

• Producing entire middleware platform 100% open source
under Apache license

• Business model is to sell comprehensive support &
maintenance for our products

• Venture funded by Intel Capital and Quest Software.

• Global corporation with offices in USA, UK & Sri Lanka

• 150+ employees and growing.

Introduction to Data Problem
• Information explosion
o Rapid growth of published data.
o Managing large amounts of data is difficult (this leads to
an information overload)
o Difficulties include
 Capture
 Storage
 Search
 Sharing
 Analytics
 Visualization
o We need new tools to deal with BIG DATA.

The Well Known Data Solution
RDBMS
• For many years this has been the choice

• Scaling up RDBMS
o Put it in a bigger computer
o Replicate database over 2 - 3 nodes. This does not work well
with more than 2 - 3 nodes.
o Partition data over several nodes. Although JOIN queries are
hard across many nodes, may require custom code and
configuration. Transactions may not scale well.

CAP Theorem and RDBMS
• RDBMS has two key features
o Relational Model with SQL
o ACID transactions (Atomic, Consistent, Isolation &
Durable)
• CAP theorem states that in distributed systems it is only
possible to have two properties out of the properties
Consistency, Availability & Partition Tolerance at any given
time.
o Once you have picked two properties you will loose the
remaining one.
• But there are some applications that do not need all the
properties of RDBMS. Once these are dropped system
scales. (e.g. Google Big Tables)

Rise of NoSQL
• Large internet companies hit the problem first, they build
systems that are specific to their problem, and they did
scale.
o Google Big table
o Amazon Dynamo

• Soon many others followed, and most of them are free and
open source.

• Among advantages of NoSQL are
o Scalability
o Flexible schema
o Designed to scale and support fault tolerance out of the
Box

Finding the right Data Solution

• Data Types
o Unstructured Data
 Files

o Semi Structured Data
 XML Databases, Queues, Graphs and Lists

o Structured Data
 DBMS

Handling Unstructured Data
• Storage Options
o Key - Value storages for small data items
o Distributed file systems for other cases
o Metadata Registries (Nirvana, SDSC Resource broker)
• Scalability
o Key - Value storages are highly Scalable (e.g. Amazon
Dynamo)
o Distributed File Systems are generally scalable (HDFS,
Lustre)
o Metadata Registries are also highly scalable
• Search
o Each of above provide key based retrieval
o Metadata registries provide property based search.
o It is possible to build a index for content using tools like
Lucence and use that for search.

Handling Semi-Structured Data
• Storage Options
o Answer depends on the type of structure. (e.g.
XML = XML Databases,
Graphs = Graph Databases, List = Data structure servers, work
items = Queue)
o If there is a server optimized for a given type, it is often much more efficient than
using a DB. (e.g. Graph databases can support fast relationship search)

• Scalabilty
o XML databases can shared data across nodes, so usually scalable, but others are not
that scalable

• Search
o Very much custom. E.g. XML or any tree = XPath
o Graph can support very fast relationship search

Handling Structured Data (1-3 nodes)

• In general using DB here
Small (1-3 nodes)
for every case might
Loose Operation Transactions
Consistency Consistency work.
Primary Key DB/ KV/ CF DB/ KV/ CF DB • Reason for using options
other than DB
Where DB/ CF/Doc DB/ CF/Doc DB
• When there is
JOIN DB DB DB potential need to scale
Offline DB/CF/Doc DB/CF/Doc DB/CF/Doc
later.
• High write throughput
• KV is 1-D where as other
two are 2D
*KV: Key-Value Systems, CF: Column
Families, Doc: document based
Systems

Handling Structured Data (10 nodes)

• KV, CF, and Doc can easily handle
Scalable (10 nodes)
this case.
Consistency Consistency • If DBs used with data shredded
across many nodes.
Primary
Key
KV/CF KV/CF Partitioned
DB?
• Transactions might work with
Where CF/Doc CF/Doc Partitioned
given that participants on one
DB? transaction are not too many.
JOIN ?? ?? Partitioned • JOINs might need to transfer too
DB??
much data between nodes.
Offline CF/Doc CF/Doc No
• Also should consider in Memory
DBs like Vault DB
• Offline mode will work
• Most systems let users choose
*KV: Key-Value Systems, CF: Column consistency, and loose consistency
Families, Doc: document based can scale more. (e.g. Cassandra)
Systems

Highly Scalable System
• Transactions does not work in this scale.
Highly Scalable (1000s nodes) (CAP theorem).
• Same for the JOIN. Problem is sometime
Consistency Consistency too much data needs to be transferred
Primary KV/CF KV/CF No between nodes to perform the JOIN.
Key
• Offline case handled through Map-
Where CF/Doc CF/Doc No Reduce. Even JOIN case is OK since
there is time.
JOIN No No No


Systems

Highly Scalable Systems + Primary Key Retrieval

• This is (comparatively) the easy one.
Highly Scalable (1000s nodes)

Loose Operation Transactions • Can be solved through DHT
Consistency Consistency (Distributed Hash table) based solutions
Primary KV/CF KV/CF No or architectures like OceanStore.
Key
Where CF/Doc(?) CF/Doc(?) No
• Both Key-Value Storages(KV) and
JOIN No No No Column Families (CF) can be used. But
Key-Value model is preferred as it is
more scalable.

Systems

Highly scalable systems + WHERE
• This Generally OK, but tricky.
Highly Scalable (1000s nodes)

Loose Operation Transactions • CF work through a Secondary index that
Consistency Consistency
do Scatter-gather (e.g. Cassandra).
Primary KV/CF KV/CF No
Key
• Doc work through Map-Reduce views
Where CF/Doc(?) CF/Doc(?) No
(e.g. CouchDB).

JOIN No No No
• There is Bissa, which build a index for all
possible queries (No range queries)

• If you are doing this, you should do pilot
runs and make sure things work.

Systems

Hybrid Approaches

• Some solution have many types of data and hence need
more than one data solution (hybrid architectures).

• For example
o Using DB for transactional data and CF for other data.
o Keeping metadata and actual data separate for large data
archives.
o Use GraphDB to store relationship data while other while
other data is in Column family storage.

• However, if transactions are needed, transactions have to
be handled outside storages (e.g. using Atomicas,
Zookeeper ).

Other Parameters

• Above list is not exhaustive, and there are other parameters
o Read/Write ratio - when high, easy to scale.
o High write throughput.
o Very large data products - you will need a file system.
May be keep metadata in Data registry and store data in
a file system.
o Flexible schema.
o Archival usecases
o Analytical usecases
o Others ...

WSO2 Data Solutions

• Data Service Server - DSS

• Relational Storage Service - RSS
• Column Store Service - CSS
• File System as a service ( FSaaS) - HDFS

• DSS and RSS
• DSS and CSS

WSO2 Data Service Server (DSS)

WSO2 Data Service Server (DSS)
 Support for large XML outputs
 Content Filtering based on User's role
 Support for named parameters
 Ability to configure schema type for output elements
 Mixing multiple data sources in nested queries
 Distributed transaction support
 Oracle Ref Cursor support
 Support for multiple data source types
 Clustering support for High Availability and High Scalability
 Full support for WS-Security, WS-Trust, WS-Policy and WS-Secure Conversation and XKMS
 JMX and Web interface based monitoring and management
 WS-* and REST support
 Data validations
 UDT (User Defined Type) Support
 Complex Results
 Auto Generated Keys Support
 Boxcarring Support
 Batch Request Support
 Scheduled Tasks
 Registry Integration for Excel,CSV,XSLT
 Web Scraping Support
 Multiple SQL Dialect Support
 DB -> DS Generation
 Service Group/Hierarchy Support
 Database Explorer
 Data as a Service Features - DSS Stratos Service
o Cassandra Integration
o RDS Provisioning

Data Services Description Language - DSDL

WSO2 Stratos Support for Relational Data

• Offering a “database as as service” for tenants
WSO2 Relational Storage Service
• Users create database and receive JDBC URL
• Database is allocated from Amazon RDS (MySQL) horizontal cluster
• Tenants are isolated from each other and integrated with platform
security model

WSO2 Relational Storage Service

• Use your own database server (anywhere)

• Register database connection as a datasource
Use RSS to allocate a database

WSO2 Column Store Service - CSS

Users can log in to the Web Console and create
Cassandra key spaces.

Column Store Service (Contd.)

• Key spaces will be allocated from a Cassandra clusters

• Users can manage and share his key spaces through Stratos
Web Console and use those key spaces through Hector
Client (Java Client for Cassandra)

• In essence we provide Cassandra as a part of Stratos as a
Service with Multi-tenancy support and Security integration
with WSO2 security model

WSO2 CSS Admin Console
Left Menu

Keyspace View

WSO2 CSS Admin Console

Keyspace Connection Details

File System as a Service - FSaaS

File System as a Service - FSaaS

The volume will be allocated from a HDFS cluster they are
isolated from other tenants in Stratos it is integrated with WSO2
Security model.

Users can manage and share his File system through Stratos
Web Console and use the file system like any other file
system.

Data Processing - Mapreduce

• Mapreduce is inspired by map and reduce functions used in
functional programming.
o Initially introduced by Google with some parts being
patented.

• Hadoop is a Mapreduce implementation that comes under
Apache license agreement.

• WSO2 provides Mapreduce as a service.

• WSO2 Business Activity Monitor (BAM2) is an example use-
case for WSO2's Mapreduce as a service.

WSO2 Mapreduce
• WSO2 Mapreduce is secure.

• WSO2 Mapreduce can use both FSaaS and DSS.
o HDFS (FSaaS)
o Cassendra (DSS)

Selected Customers

https://ail.google.com/mail/u/0/?ui=2&i
k=ad9ae58f41&view=att&th=1331a70
983344a32&attid=0.1&disp=thd&reala
ttid=f_gtxto6mk0&zw

WSO2 engagement model

• QuickStart
• Development
Support
• Development
Services
• Production
Support
• Turnkey Solutions
• WSO2 Mobile Services Solution
• WSO2 FIX Gateway Solution
• WSO2 SAP Gateway Solution

Storing and processing data with the wso2 platform

More Related Content

What's hot (20)

Similar to Storing and processing data with the wso2 platform (20)

More from WSO2 (20)

Recently uploaded (20)

Storing and processing data with the wso2 platform