SOA Implementation Styles

Shivnarayan Varma
Sr. Architect

25th January 2013

What are we going to discuss

1. SOA Introduction
2. SOA vs EAI
3. SOA vs Web Services
4. Different SOA Implementation Styles
I.
II.
III.
IV.
V.
VI.
VII.

SOA with REST

SOA with JAX-WS
SOA with Messaging
SOA with ESB
SOA with Java/JEE and JMS (MessageMQ)
SOA with Java based products like Gemfire/Tangosol.
Sample use case for using SOA without ESB

2012 Fiserv, Inc. or its affiliates.

SOA Introduction
SOA allows the integration of existing systems, applications and users
into a flexible architecture that can easily accommodate changing
needs. Integrated design, reuse of existing IT investments and above
all, industry standards are the elements needed to create a robust
SOA.

2012 Fiserv, Inc. or its affiliates.

SOA vs EAI
EAI products are essentially the same as ESB products. The
difference between them is ESB's addition of and emphasis on
Web Services. Both EAI and ESB share the idea of a centralized
point of control and try to reduce everything to a common
denominator. The architecture of ESB and EAI middleware is
almost the same.

2012 Fiserv, Inc. or its affiliates.

SOA vs Web Services

SOA

Web Services

Service Oriented Architecture is the

term that designates a philosophical
approach to a solutions design

The utilization of web services is one

method by which such an architecture
might be attained.

SOA doesn't have to be limited to Web

services -- it could be built on CORBA
components .

Lot many times attempting to implement

SOA, only by implementing Web
services end up with redundant,
incompatible, and often unmanaged
services that have nothing to do with
architecture."

2012 Fiserv, Inc. or its affiliates.

SOA with JEE (Java Web Service) Architecture

2012 Fiserv, Inc. or its affiliates.

SOA with JEE (RESTful Web Service)

2012 Fiserv, Inc. or its affiliates.

SOA with JEE (RESTful Web Service)

When should I use REST based service?
Use REST when the web services are completely stateless.
Scalability and Performance
Loose Coupling
Consistency and Uniformity
REST services are more light weight (use JSON) over SOAP.
When we are decided with HTTP protocol model. (JAX-WS also
supports SOAP over JMS).
We dont need distributed transactional features.

2012 Fiserv, Inc. or its affiliates.

JAX-WS vs REST in web services

REST

Simple point-to-point communication using

well-established HTTP verbs, protocols, and
standards.

SOAP

Agnostic to language and platform.

Conceptually simpler to understand and
develop.
Less reliance on tools
Closer in design and philosophy to the web.

Locked into the HTTP transport model.

Cannot be scaled to non-HTTP
communication models.
Lack of standards support for security,
policy, messaging, etc., making apps with
richer requirements harder to develop.

2012 Fiserv, Inc. or its affiliates.

Supports a variety of transports (HTTP, JMS, etc.)

Integrates with a variety of web service standards.
Typically used to pass contractually structured
data between applications.
Support distributed Transactions protocol,
standard messaging for retry
Agnostic to language, platform, and transport.
Designed for distributed environments and
applications.
Richer, more mature, standards-based support
and integration.
Built-in error-handling (faults).
Extensible through integration with other
transports and data bindings.
Heavier abstraction and steep learning curve.
More verbose.
Difficult to develop without tools.

SOA with JEE (RESTful Web Service)

Web Services Are Web Sites
/products
/products/{id}
HEAD

PUT

DELETE

The product list (/)

List your
products

A product (/
{product})

List the
products
objects

Create the
product

Delete the
product

/product/{id}

Get the
Set the objects
objects
Get the objects
Delete the
value and
value and metadata
object
metadata
metadata

GET

10

2012 Fiserv, Inc. or its affiliates.

SOA with JEE (RESTful Web Service)

JAX-WS vs REST in web services
An advantage of REST over SOAP is that the action is specified by
the HTTP method and by the URL; this means that fewer bytes have
to be sent as compared to SOAP which relies on verbose XML. And
fewer bytes mean less latency; in this scenario REST is
recommended instead of SOAP.
JSON VS XML:
JSON is smaller and thus uses less bandwidth and memory during
parsing/generation .It supports Boolean and numeric values.

11

2012 Fiserv, Inc. or its affiliates.

SOA with JEE (RESTful Web Service)

12

2012 Fiserv, Inc. or its affiliates.

SOA with JEE Web Service(JAX-WS)

13

2012 Fiserv, Inc. or its affiliates.

SOA with JEE Web Service(JAX-WS)

14

2012 Fiserv, Inc. or its affiliates.

SOA with JEE Web Service(JAX-WS)

When should I use JAX-WS based service?

Supports a variety of transports (HTTP, JMS, etc.)
Support distributed Transactions protocol, standard messaging
for retry.
Designed for distributed environments and applications.
Richer, more mature, standards-based support and
integration.
Built-in error-handling (faults).
Extensible through integration with other transports and data
bindings.

15

2012 Fiserv, Inc. or its affiliates.

JMS Technology

16

2012 Fiserv, Inc. or its affiliates.

SOA with Java/JEE and JMS (MessageMQ)

17

2012 Fiserv, Inc. or its affiliates.

SOA Implementation with JMS 2 Cont.

18

2012 Fiserv, Inc. or its affiliates.

SOA using Messaging

Asynchronous messaging is fundamentally a pragmatic reaction to

the problems of distributed systems. Sending a message does not

require both systems to be up and ready at the same time.
Furthermore, thinking about the communication in an
asynchronous manner forces developers to recognize that working
with a remote application is slower, which encourages design of
components with high cohesion (lots of work locally) and low
adhesion (selective work remotely).

19

2012 Fiserv, Inc. or its affiliates.

SOA using Messaging (Point to Point Model)

A Point-to-Point Channel ensures that only one receiver consumes any
given message. The channel can have multiple receivers that can
consume multiple messages concurrently, but only one of them can
successfully consume a particular message. If multiple receivers try to
consume a single message, the channel ensures that only one of them
succeeds, so the receivers do not have to coordinate with each other.
Example: JMS Queue
In JMS, a point-to-point channel implements the Queue interface. The
sender uses a QueueSender to send messages; each receiver uses its
own QueueReceiver to receive messages [JMS 1.1],

20

2012 Fiserv, Inc. or its affiliates.

SOA using Messaging (Public Subscribe Model)

21

2012 Fiserv, Inc. or its affiliates.

SOA using Messaging (Public Subscribe Model)

A Publish-Subscribe Channel works like this: It has one input channel
that splits into multiple output channels, one for each subscriber.
When an event is published into the channel, the PublishSubscribe Channel delivers a copy of the message to each of the
output channels. Each output end of the channel has only one
subscriber, which is allowed to consume a message only once. In
this way, each subscriber gets the message only once, and
consumed copies disappear from their channels.
Example: JMS Topic
In JMS, a Publish-Subscribe Channel implements the Topic

interface. The sender uses a TopicPublisher to send messages;

each receiver uses its own TopicSubscriber to receive messages

22

2012 Fiserv, Inc. or its affiliates.

SOA with Message BUS

Use Messaging to transfer packets of data frequently, immediately,

reliably, and asynchronously, using customizable formats.

23

2012 Fiserv, Inc. or its affiliates.

SOA using ESB

Limitation of MOM: Message Oriented Middleware

(MOM) provides the ability to connect applications in a loosely

coupled, asynchronous fashion. However, MOM by itself requires
low-level coding in an application. Using a traditional MOM, along
with custom coding techniques, can get you a long way toward a
distributed integration solution. However, without a higher level of
abstraction of the routing logic, this approach also suffers from
having integration logic hard-wired and intertwined with the
application logic. Depending on the MOM being used, even the
distributed characteristic might be limited because some traditional
MOM infrastructure is not capable of spanning physical network
boundaries very well.

24

2012 Fiserv, Inc. or its affiliates.

SOA using ESB

Why we need ESB: Finally, in an ESB, services

can be configured rather than coded. Process flow and service

invocations can transparently span the entire distributed bus. An
ESB provides a highly distributed integration environment that
spans well beyond the reach of hub-and-spoke architectures, and a
clear separation of business logic and integration logic such as
routing and data transformation. An ESB architecture forms an
interconnected grid of messaging hubs and integration services,
with the intelligence and functionality of the integration network
distributed throughout.

25

2012 Fiserv, Inc. or its affiliates.

Capabilities and Features of an ESB

Message Transformation
Protocol Transformation
Message Routing
Security
Transaction
Orchestration
Registry

Transform Provider message format to

Consumer message format and vice versa
(XML to ASCII, XML to XML, etc.)
Consumer Services requests use a transport
protocol (e.g. JMS) different from the
protocol (e.g. HTTP) of the Provider
Content based routing of Consumer
messages to appropriate Provider & Vice
Versa. Support for rules based routing
Support for WS-Security, SAML, encryption,
Authentication and Authorization
Support for WS-AtomicTransaction,
WS-Coordination, WS-BusinessActivity
Support for Orchestration & Choreography

26
26

2012 Fiserv, Inc. or its affiliates.

Typical ESB Architecture

Extended
Management
Services
Base
Services

WS-*

Enterprise
Service

Enterprise
Service

Enterprise
Service

Monitoring
XML
Services

Deployment

JCA

JMS

Messaging Backbone
27
27

2012 Fiserv, Inc. or its affiliates.

ESB Definition

28

2012 Fiserv, Inc. or its affiliates.

Without ESB SOA is difficult to manage.

29

2012 Fiserv, Inc. or its affiliates.

Issues with SOA using ESB

In a SOA environment each service must perform identification, authentication, and

authorization every time. Establishing a security context can be compute intensive

relative to the rest of the work the service may perform.In SOA environment share
state either passing large XML documents between steps (which can be computer
and I/O intensive) or writing intermediate state to a database which adds the latency
of disk I/O and created data management challenges if a process is abandoned before
it is finished.
Another difficulty may be encountered with long-running business processes, where

managing state across multiple Web Service calls presents challenges in dealing with
session time-outs. XML is often the preferred data standard in SOA, mainly due to
its elegance from an application interoperability standpoint. Unfortunately the
flexibility comes with associated performance inefficiencies in storing, marshalling/
unmarshalling and manipulating XML documents.

30

2012 Fiserv, Inc. or its affiliates.

Relevance of EDF (ELASTIC DATA FABRIC)

Using MOM in combination with EDF

There are two especially important ways that help to reduce application latency is by

reducing network and disk-based communication, and the other is by staging the
data in an object format that works better with applications.
A distributed data grid, also called a distributed data cache, operates between the

database and the in-memory of an application and provides a temporary repository

for data, enhancing performance by improving access and eliminating bottlenecks.
The best way to describe a distributed cache is something that acts across one or

more nodes. that they can add and remove nodes while running.
if you start with four servers and add four more, these platforms will rebalance the

data fairly evenly across the nodes and if any node goes down you are not down
because they replicate the data. "So elastic caching also provides fault tolerance and
high availability at a fraction of the cost of what it would take just to re-architect a
database,"

31

2012 Fiserv, Inc. or its affiliates.

Relevance of EDF (ELASTIC DATA FABRIC)

To achieve the real promise of Service Oriented Architectures, IT organizations need to move
from a centralized data management system, characterized by back-office databases and
warehouses, to a more distributed middle-tier data fabric that provides a scalable way to access
and distribute data across multiple services. Simply put, data must be accurate, timely, highly
available, and instantly accessible to any business process.

From a scalability viewpoint, services need ways to keep large datasets operationally ready for
access by distributed business processes. SOA infrastructures must scale transparently and
handle incremental additions of new data sources as well as additional volume of current
sources
Data Virtualization:
High Performance Data Access
High Availability
Low Latency Data Distribution
SLAs.

32

2012 Fiserv, Inc. or its affiliates.

Relevance of EDF (ELASTIC DATA FABRIC)

Business Application Data:by which operational data can be
held in memory for low latency access and increased system performance and throughput. The subset of
data held in memory is swapped intelligently based on client request patterns. In certain cases,
perational data volumes may stretch the physical limits of RAM available on a machine. In such cases
data can either be partitioned across nodes or moved to disk (with indexes to the disk storage areas) to
accommodate surging data loads.
Security Context:
SOA environments perform security checks and look up a users authorization with each invocation of a
service. This means that a service invocation results in a check with a policy server or other security
mechanism. These mechanisms inevitably look in a database to find this information. For globally
distributed organizations this often means a trip across a WAN as well. All of this work adds latency and
significantly slows down fast services. For composite processes with N number of steps this work is
performed N times. If the security context is cached in GemFire it can be seamlessly shared across
all of the steps in the workflow and the services that compose that process avoiding unnecessary delays
that impact QoS and

33

2012 Fiserv, Inc. or its affiliates.

Relevance of EDF (ELASTIC DATA FABRIC)

34

2012 Fiserv, Inc. or its affiliates.