Today's business applications rarely live in isolation. Users expect instant access to all the business functions an enterprise can offer, regardless of where the functionality resides. This requires disparate applications to be connected to a larger, integrated solution. This integration is usually achieved through the use of some form of middleware. Middleware provides the "plumbing" such as data transport, data transformation and routing.

They are lot of acronyms such as EAI, EBI, EII, EDI and ELI that have been used in the past several years with differing meanings offered as solutions to all our integration problems. In reality, they represent different areas of focus for the integration efforts that often overlap. Further, they tend to be focused on a particular tool vendor solution, service provider, business issue or direction of interest.

This article gives a brief outline of the various layers of integration and defines the related acronyms used in each layer of an enterprise. It also focuses on clarifying the most confusing and widely used application and data integration solutions by outlining their strengths and weaknesses. 

The enterprise integration layers can be broadly divided into four layers as shown in Figure 1. 

Figure 1: Enterprise integration layers 

Business Integration Layer

EBI (enterprise business integration): EBI is the restructuring of the enterprise from a business perspective. EBI is all about integrating processes, locations, product lines, infrastructure, intellectual property, markets, and production and delivery operations, catalogs and so on. These are all “pure” business things and only impact IT when there is a need for integration of infrastructure such as that used in the delivery of financial data. The techniques involve enterprise analysis and do not involve IT-based methodologies.

EPI (enterprise process integration): EPI refers to automating business processes both within an organization and across business customers, partners and supply chains. EPI also called BPI (business process integration) connects disparate technology systems to streamline the transfer of business information to and from various technology resources. For those departments using different technology tools with common data requirements and processes, BPI unites these systems, which ultimately improves efficiency and reduces staffing and system maintenance costs. BPI can be divided into internal (i.e., EAI) and external (i.e., BP2Ai, BP2BPi) classifications based on the integration with outside enterprises.  

B2B (business-to-business integration): B2B encompasses direct peer-to-peer exchange of messages between enterprises (without an intermediary) as well as interaction in marketplaces as sellers or buyers (where the marketplace is an intermediary). It also refers to all business activates of an enterprise that have to do with electronic message exchange between it an one or more of its trading partners.

Information Integration Layer

ECI (enterprise content integration): Content integration pulls together intellectual property from many sources and displays the results to interested parties. Main drivers for the integration are knowledge access, leverage, management for employees and customers, such as portal and knowledge management tools.

ELI (enterprise legacy integration): ELI is a software abstraction layer that resides between the Web-enabled software and the legacy layer or green-screen code such as RPG or COBOL. 

EII (enterprise information integration): EII provides real-time views across multiple sources of data within one or across multiple organizations. Optimized and transparent, this data access and transformation layer provides a single interface across all enterprise data.

Application Integration Layer

EAI (enterprise application integration): EAI integrates a set of enterprise computer applications. EAI is also defined as the ability to link transactions across multiple systems through existing applications like SAP and Oracle Financials. EAI is typically a push system that is not utilized for batch transformation. EAI is message-based, transaction-oriented, point-to-point (or point-to-hub) brokering and transformation for application-to-application integration. The traditional EAI approach is hub and spoke.

 

EAI comprises of message acceptance, transformation, translation, routing, message delivery and business process management. Commercial products providing EAI packaging solutions are listed below:

ESB (enterprise service bus): An ESB is a software infrastructure component which provides a service agnostic framework for providing multiple applications or "services" the ability to cooperate or interoperate. An ESB allows connectivity and standard service management utilities such as data persistence across applications as well as APIs for universal access. An ESB is also defined as a technology layer which makes communications necessary to support services-oriented interactions and combines messaging, transformation and content-based routing.

SOA (service-oriented architecture): SOA is a business-centric IT architectural approach that supports integrating business as linked, repeatable business tasks or services. An SOA is essentially a collection of services. These services communicate with each other. The communication can involve either simple data passing or it could involve two or more services coordinating some activity.  

BLI (business logic integration): BLI is the process of calling host systems through callable interfaces. Most of the newly developed IBM mainframe and midrange applications contain business logic that is separated from the presented screens. Bypassing these screens allows these interfaces to be called directly.

SLI (screen logic integration): SLI is the process of calling host systems through screen logic. This is where a sequence of screens are recorded, mapped to a function and then "played back" to the host in order to re-execute the function automatically.

Data Integration

There are three main techniques used for integrating data: data consolidations, data federation and data propagation.

Data Consolidation

A wide range of technologies available for implementing the data consolidation techniques is shown in Figure 2:

 

Figure 2

EDI (enterprise data integration): Enterprise data integration involves a framework of applications, tools, techniques, technologies and management services for providing a unified and consistent view of enterprise business data to business processes and business users.

ETL (extract, transform, load): ETL transfers batches of information from one system to another. Generally used to move large sets and perform data transforms in mid-stream and load into the target system(s). ETL is usually a pull system; however, some vendors are heading toward push/pull ETL. ETLis set-oriented point-in-time transformation for migration, consolidation.

Commercial products providing ETL solutions are listed below:

 

ELT (extract, load, transform): ELT occurs in the database after the loading cycle has been accomplished. ELT is generally a pull system with massive data sets.

Commercial products providing ELT solutions are listed below:

ETLT (extract, transform, load and transform): ETLT is a combination of in-stream and in-database transformations. ETLT is generally a pull system utilized mostly in a batch-oriented or small-batch (timed) fashion. ETLT comprise features from ETL and ETL; Commercial products providing ETLT solutions are:

ELR (extract, load in real time): ELR is the ability to connect to any system, extract data and load it into another system in real time with some transformation occurring within the RDBMS. ELR could be a push or a pull system for sending data through the pipes.

ELRD (extract, load in real-time with dynamic restructuring capabilities): ELRD automatically detects changes to structures and reacts to those changes automatically. ELRD will require a bit of engineering "magic" to develop.

 Data Federation

 

Figure 3

EII (enterprise information integration): EII provides real-time views across multiple sources of data within one or across multiple organizations. Optimized and transparent data access and transformation layers provide a single relations interface across all enterprise data.

Figure 4

 

Data Propagation

 

 Figure 5

EDR (enterprise data replication): EDR applications copy data from one location to another. These applications usually operate online and push data to the target location; i.e., they are event-driven. Updates to a source system may be propagated asynchronously or synchronously to the target system. Synchronous propagation requires that updates to both source and target systems occur in the same physical transaction.  

 

 Integration Landscape

Figure 6

 

 The following chart lists the potential criteria for choosing the right integration technique:

Recent articles by Narasimha Murthy

• 10 Cornerstones of Business Intelligence