Rules: App & Data Services

This is the 4th installment of articles I am writing about App Services vs. Data Services. So far, we've come to realize the differences in these layers. In this article, we will see how the differences simplify design and implementation of these specific services.

Category	App Services	Data Services
server interface	HTTP, SOA	HTTP, SOA
implementation	stateless	stateless
client interface	REST JSON/ XML	POJO
addresses	business rules	data challenges
top concerns	extensibility, flexibility	scalability, performance, reliability
techniques	AOP, OOP	ADT, Collections, Algorithms
communicates to	http services	data repositories
testing	scenarios	data volume, repository accessibility
error mgmt	app errors	system errors
i18n, l10n	pass l10n header	choose resource string

An app service loads up pluggable business rules and logic, and then executes them. A java based implementation recipe might consist of the following:

• Spring libraries to take advantage of its AOP features
• Drools libraries to take advantage of pluggable business rules

A data service allows its clients to create different classes of data, and then manipulate them. A java based implementation recipe might consist of the following:

• Spring libraries to take advantage of its DI features
• Hibernate/ Mybatis libraries for their ORM layer (maybe other JPA for various other No SQL)

As infrastructure, we could evaluate:

• JMS provider (Sonic MQ, Rabbit MQ) for reliable async communications
• BOSH provider (ejabberd, punjab, lightstreamer) for streaming connectivity
• EVENT provider (mule esb) for EDA