Equivalent to Why Enterprise Service

¶ … equivalent to Why Enterprise Service Bus Architectures in Manufacturing Matter Now the many business drivers that are influencing and growing the use of Enterprise Service Bus (ESB) architectures throughout global manufacturing are also making it critical for manufacturers to take into account the need for greater synchronization of supply chains across the many sourcing, supply chain, manufacturing, fulfillment and service centers. Over and above the need for manufacturers to synchronize these disparate, legacy and often non-integrated systems with ERP, supply chain, distributed order management, and often, global CRM systems is the more fundamental challenge of turning all these systems into competitive advantages in the marketplace.

The key business drivers in manufacturing leading to greater adoption of SOA as a competitive strategy in conjunction with ESB architectures include the following. First, there is the exponential rise in mergers and acquisitions across many of sectors of global manufacturing. The heightened levels of mergers and acquisitions has in turn lead to the development of distributed governance models, and specifically for those companies operating in the U.S., an adherence to compliance more than ever before due to the Sarbanes-Oxley Act (SOX) (2002) and many other forms of compliance legislation. Indian outsourcing companies including Infosys, HCL, and others claim that during their fiscal 2005 financial year, one of our every three dollars earned was the direct result of compliance spending by global manufacturers who had significant presence in the U.S. market and traded securities on American-based exchanges. Clearly the business drivers that are enabling the growth of ESB adoption, development, and use also favor the adoption of BPEL-based workflows in response to the demanding compliance requirements for manufacturers who choose to participate in U.S. capital markets. In reality SOX has done more for the adoption of SOA and with it, ESB frameworks and the resulting gains in process re-definition and efficiencies. SOX, for many global manufacturers participating in U.S. capital markets, have become the impetus to completely re-define their core business processes. The logic of many global manufacturers is that while they are incurring the costs of attaining compliance to government regulations and standards for financial disclosure, they can at the same time completely re-define their it strategies to better align with the needs of their business. In this regard many manufacturers today look to compliance as the impetus for change, and the primary reason to re-architecting how information technologies and services support making their organizations more agile and flexible. As Abrahams (2005) mentions, compliance is also the impetus for overcoming swivel-chair integration, which has continually plagued companies who have not specifically addressed their manual integration challenges yet. Many manufacturers are in this category.

The key it drivers that are supporting the growth of ESBs throughout manufacturing include the need for route requests more efficiently throughout the many ERP, order management, supply chain, logistics, CRM, and service applications that manufacturers in many cases have allowed to become too decentralized and siloed in performance. This key it driver of ESB adoption has to do with the integration of multiple ESBs located throughout a global manufacturing enterprise. XML-based messaging and schema also contribute to the real-time integration between ESB hubs, allowing for multiple state engines specific to the dominant processes in each location to be synchronized and also predicted in their overall performance. One critical measure of this overall level of synchronization of systems is the attainment of the Perfect Order, a key performance indicator manufacturers who have mass customization strategies rely on for measuring performance gains in processes over time. Additional drivers from the architectural standpoint for ESB adoption and growth in manufacturing include resolution in many manufacturers as to if they should electronically enable their direct and indirect channels for real-time integration, and if so, which business models are most amendable to the change. From an it perspective, the development of real-time integration connections throughout either an indirect or direct channel is challenging, costly, and very time-consuming. The flipside of the argument in must manufacturers today however is that it is a critical competitive advantage and sets the stage for deploying more Web Services-based applications outside the firewall into the broader indirect channels. The growth of ESBs, their integration points, and the need for allowing for loosely coupled ESBs to exist in the broader SOA strategy of deploying services and using the many diverse databases, all originally created to serve the channels and the sales forces, is being explored by many manufacturers, the foremost being Dell Computer, Intel, and HP in the high tech discrete manufacturing industry, and Proctor & Gamble who has created an impressive set of private trading exchanges specifically using the strategy of gradually increasing the number of real-time integration points over and above the manually-based ones that had pervaded their channels prior to the pilots. Table 1 provides a comparison of the strategies manufacturers are using to accomplish real-time integration throughout their channels, and also defines those tasks by channel strategy that manufacturers today are leaving in batch mode relative to enabling through ESB architectures to become more private trading exchange-like in their ability to quickly sense and respond to market demands. In those areas of real-time integration, manufacturers mentioned previously are piloting Web Services and relying on BPEL-based workflows to create greater efficiencies and a strong focus on measuring business benefits as a result.

There are the exacerbating factors of ensuring order state engines across ESBs can function correctly and can communicate reliably when any channel-based strategies is considered in the context of SOA. Many manufacturers are using XML-based linking to provide real-time analytics visibility throughout their demand channels, ensuring that forecasts arriving at manufacturing centers in fact can be build specifically to the requirements of the businesses ordering them. These scenarios as they relate to real-time vs. batch integration strategies in Web Services are dominated by B2B companies who measure, monitor, and modify their strategies in the context of channel velocity.

Table 1: Planning Web Services for Channels: Real-Time vs. Batch Integration

Channel Strategy or Business Model

Channel and Business Model Characteristics

Batch or Real-time Integration Web Service Structure?

Direct Sales Model

Sales force calls directly on customers; High margin and high ticket products and services sold directly to customers; cost per sale is high; response time on sales cycles is high; price elasticity of products is low.

Batch-oriented tasks: Lead management and escalation; structured content and catalogs; content for sales configurators; training materials; competitive analysis; pre-sales support and training; post-sales support and service training.

Real-time tasks: Pricing for quotes; order management and supply chain links for Available-to-Promise (ATP) and additional supply chain information that customers in specific industries expect; quoting and pricing approval process.

Indirect Sales Channel - Single Tier

Manufacturer's sales force calls on distributors and value-added resellers only; product information is sporadically provided and pricing is updated periodically.

Batch-oriented tasks: Product information including new product introductions; changes to products' positioning and end-of-lifing a product; lead management and escalation; definition of pricing strategies.

Real-time tasks: Quoting responses; Special Pricing Requests; Lead escalation on products targeted for specific promotions; exception-based transactions.

Indirect Sales Channels - Multiple Tier

Manufacturers' sales force calls on both distributors and resellers; supports the first tier of distribution with leads, pricing, training and product knowledge to second tier of dealers. Lag time between manufacturers and second tier resellers is often perceived as lack of responsiveness.

Batch-oriented tasks: Comparable to Single-Tier Indirect Channels including new product introductions; changes to product positioning and end-of lifing products; multi-tier channels also include more coordination of batch processes for sales promotion programs; greater requirement for batch-oriented warranty reimbursements to 2nd tier channel participants; this channel strategy requires the greatest level of support for channel partners and their dealers and resellers.

Real-time tasks: Support for Special Pricing Requests; coordination of training and benefits internally throughout multiple layers of distribution channels; extensive support for exception-based transactions; heavily focused on automating Special Pricing Requests and providing real-time self-service tools to multiple layers of the channel.

OEM Sales

Manufacturer sells directly to Original Equipment Manufacturers (OEM)s who many times have products tailored to their specific needs.

Batch-oriented tasks: Mftrs rely on direct sales forces to sell into OEM accounts; heavily batch-oriented for presales, sales and support.

Real-time tasks: Intensive real-time support on design and engineering validation; pricing is often done in real-time; supply chain integration necessary for tracking customized builds for OEM accounts.

The ongoing debate of whether to bring real-time Web Services integration into indirect channel partners, and with it, a fundamental re-ordering of process using BPEL as the modeling framework for re-defining in-channel processes, all interlinked via XML and inter-ESB communication, is today focusing on the velocity of transactions that manufacturers can accomplish first, and cost reductions second. The re-thinking of channel partners as being on their own systems, often in batch-mode and very constrained in terms of performance and responsiveness both to other channel partners and customers, is pervasive. The impediments these "islands" of batch integration are causing is forcing manufacturers to further focus on how to propagate key state engines throughout their global distribution channels for greater effectiveness and timeliness of response throughout global distribution and selling networks.

As has been mentioned throughout this thesis, the entire aspect of mass customization as it relates to the development of a stable quote-to-order process throughout manufacturing is critical. In this specific area is where many manufacturers face the dilemma of being entirely project-based in their manufacturing and business strategy approach or move more towards functional manufacturing with the exception being mass customization and a more fluid, agile, quote-to-order process. Figure 1 shows this dilemma graphically.

Source: LWC Research

Figure 1: Manufacturer's dilemma regarding mass customization and channel implications of integration

Underscoring the synchronization of ESB hubs and accompanying use of state engines to unify channels and ensure complex processes including order-to-cash, inquiry-to-order, inquiry-to-cash, and many others. The span of integration that manufacturers are achieving with ESB platform integration in support of these core business processes is also illustrated in Figure 2, an example of a Brokered ESB Pattern Model, from Redbooks (2005).

Figure 2: An example of a Brokered ESB Pattern Model

Source: IBM Corporation 2005

What is significant about the Brokered ESB Pattern Model is the applicability it has specifically to distributed order management and distributed process functions throughout the many indirect and direct channel relationships manufacturers rely on for staying aligned with demand and selling. The focus of the Brokered ESB pattern is ideally suited for many manufacturers in that its architecture aligns perfectly with the approach manufacturers use for managing their many channel relationships in that this model separates integration logic and business roles from the ESBs themselves, according to Redbooks (2005). This architectural delineation of the Brokered ESB Pattern Model is congruent with the structure of the constraint engines discussed elsewhere in this thesis, specifically in the differentiation of rules, constraints and logic on one layer of the model and data on another. Constraint-based configuration engines and optimization engines including those from Fair-Isaac with their Blaze technology further support the concept of relying on a distributed approach to separating logic and data into a more agile architecture than one where logic is directly tied to data.

Implications of Visualization Grids on ESB in Manufacturing

Workload, platform, and information visualization strategies as they relate to the integration of ESB hubs across global manufacturing operations are critical for the growth of real-time integration with channel partners and the growth of exchanges. Figure 3 provides an illustration from IBM in their Redbook (2006) that shows a Grid Access Composite Runtime Pattern.

Figure 3: Grid Access Composite Runtime pattern

Source: IBM Corporation 2006

Notice from the Grid Access Composite Runtime pattern that the use of multiple approaches to workload, platform, and information visualization separates logic structures from data stores, and in the process also accomplishes integration across multiple ESB synchronization points and also allows architecturally for the development of a series of state engines that serve each specific strategy area. An example of this type of usage within visualization for the synchronization of order state engines would be the deployment of global order management systems across a series of manufacturing centers located in geographically diverse and distributed locations. The focus within Redbook (2006) specifically on the coordination of workload, platform, and information visualization is closely tied to the coordination of ESB architectural interlinking and the ability of manufacturers to significantly increase their performance on key performance criteria over time.

IBM's extensive work specifically on WebSphere but more globally on the issues of providing analytics as part of a strategic SOA strategy are well defined in many books published by IBM including IBM Workplace for Business Strategy Execution (2006) which details the structure of the framework of IBM's approach to integrating analytics into BPEL-based in addition to BPEL4WS, and area that IBM continues to provide thought leadership in as is evidenced with several major developments in the definition of standards of the integration of Web Services and BPEL-based process workflows. IBM has also published a roadmap specifically in this area that is found on their website, IBM Developer Roadmap (2006).

The major focus of much research in BPEL4WS specifically and BPEL in general is in the quantifying of business value over time of modifying processes permanently and with key performance indicators attached to the change in performance. Elsewhere in this thesis there are many examples of key performance indicators that quantify the impact of re-engineering, more precisely aligning, and making more efficient critical customer-facing processes including quote-to-order, inquiry-to-order, and inquiry-to-cash. The quantification of the gains found in these processes have been quantified and in Figure 4, which shows the approaches IBM is using in the context of their WebSphere Series of applications to provide business process modeling tools that also deliver the financial performance of the process over time. Figure 4 provides this specific workflows area, showing how the integration of financial reporting and business process redefinition and its quantified result. The quick read-outs on the percentage completed against each task in an objective is also shown. IBM has specifically designed this approach to align with the needs of business analysts vs. programmers. This type of programming environment and its pervasive use by business analysts is a primary reason for IBM acquiring FileNet in 2006, a company that has continued to expand out its expertise in the areas of Business Process Management and its associated modeling and design applications.

Figure 4: Combining BPEL4WP and Key Performance Indicator Financial Performance

Source: IBM Corporation 2006

The need for analytics and the quantification of performance relative goals is at its essence one of the major if not the most major driver there is for manufacturers to embrace SOA strategies, and with them, ESB architectures and the growth of BPEL4WS workflows. The collaborative effects of all these technologies must be measured in many manufacturing companies, as they have specifically been planned for, deployed, and monitored to have a major impact on the ability of the company to compete and grow globally. Accountability over SOA strategies is core to the development of any lasting strategy, and both ESB and BPEL4WS, in addition to XML integration, are all critical components. All must be focused on measurable and quantifiable objectives however to earn a strong Return on Investment (ROI). This new era of accountability was exacerbated during the last recession, when many manufacturers failed to accurately measure the performance of their internal strategies on external performance. As a result of these factors and the meteoric growth of XML as an integration standard that connected previously isolated and often legacy systems with each other, dashboards have become commonplace throughout many manufacturing companies. it's important to keep in mind that despite the fact that many manufacturers still operate using Microsoft Excel as their primary means of communicating results. Using the structure of ESB in conjunction with SOA architectures including WebSphere, dashboards are starting to increasingly be used throughout manufacturers' divisions. Figure 5 shows an example of one manufacturer's dashboard. The programming and XML integration necessary along with the BPEL4WP-support process workflows are all governed by and aligned with the key business objectives the manufacturer is looking to accomplish. The need for real-time integration is clear from the read-outs within the dashboard as well, given the need to quickly correct direction in the context of demand sensing and overall market feedback.

Figure 5: An Example of a Manufacturing Dashboard

Source: IBM Corporation 2006

The underlying analytical framework that IBM's SOA approach relies on for delivering analytics is defined in the document IBM Workplace for Business Strategy Execution (2006), specifically including a hierarchical model that shows the strategic role of analytics, collaboration and enterprise application integration in the context of Portlet Factories delivering apps for use in WebSphere Portal Server, culminating in the Dashboard Framework and its presentation layer, the IBM Workplace dashboard. Figure 6 shows an example of this specific hierarchical model.

Figure 6: IBM WebSphere's Hierarchical Model

Source: IBM Corporation 2006

ESB as the Catalyst for attaining the Perfect Order

There are more than enough measures of performance to include in any scorecard, as can be seen in the previous section of this thesis. Manufacturers relying on SOA often over-measure every aspect of their performance, especially when it relates to measuring the accuracy and efficiency of their order management and fulfillment requirements. This is especially true in the context of the Perfect Order, which is a supply chain key performance indicator (KPI) that specifically measures the synchronization of and use of the many supply chain systems in addition to order management, fulfillment, pricing, and production. The Perfect Order as a metric is also being actively used by those manufacturers who have been early adopters of SOA-based strategies in conjunction with product customization and build-to-order approaches of managing and responding to demand.

Table 2 lists the most common measures of performance that LWC Research (2005) has seen manufacturers use to attain their objectives of increased business process performance in the context of their mass customization selling and production strategies.

These are the most common problem areas that serve as obstacles for manufacturers looking to attain the Perfect Order.

The adoption of ESB infrastructures and their propagation throughout an enterprise specifically to provide for loosely coupled ESB framework is what's working for manufacturers including General Electric, Proctor & Gamble, and others who have production locations scattered geographically across a wide range of locations.

Identifying the interdependencies of the metrics being used to track progress to the goal of delivering the Perfect Order requires by default an extensive SOA infrastructure in place in addition to a robust and comprehensive analytics layer to accumulate, analyze, and graphically report the key performance indicators in a structure manufacturing managers can actively use. The concept of best practices in using scorecards to measure progress towards the Perfect Order is just becoming visible in many industries, manufacturing included.

Righting the wrongs in broken quoting systems is the first place to look when fixing incorrect orders is where many manufacturers are finding BPEL-based approaches to increasing process performance paying off quickly and with financially significant results. The role of BPEL4WS is also being quantified in the context of the performance of the Perfect Order. When the quoting process for example is de-constructed with the purpose of streamlining it using BPEL, the disconnected parts of quoting systems become very apparent and the first steps are taken in constructing a strategy for making fulfillment of the Perfect Order possible.

Using the foundation of SOA and with it, a re-definition of the quoting process yields more accurate quotes, and alleviates products that are built to what was believed to be the customers' requirements - the more complex a product and the higher the level of complexity required to customize it, the greater the potential for disconnects. That's why building a strong quoting and order capture system needs to be at the heart of the Perfect Order.

Defining supply-chain performance measures including Available-to-Promise (ATP) requires integration and sets a strong foundation for delivering Perfect Orders - While many manufacturers feel that delivering ATP with their quotes isn't critically important, the integration required to deliver this metric to customers and therefore set expectations as to when their product will ship requires a depth of integration that pays off when demand needs to be synchronized between sales channels and the supply chain.

Realizing that getting to the Perfect Order requires the balancing of supply chain costs relative to the level of responsiveness you want to deliver to your customers, channels and partners relative to the costs of managing the supply chain - the most effective manufacturers that are delivering the Perfect Order are ones that have discovered the equilibrium between order costs and the continual investment needed for their supply-chain operations.

Synchronizing order capture, quote-to-order, and product configuration strategies across all channels is a must-do if the Perfect Order is ever going to be attained, and this level of integration specifically requires the use of an SOA architecture and strategy to attain the real-time integration necessary. The choice of creating either a real-time or batch-oriented integration throughout both indirect and direct channels is a dependence that manufacturers attaining a high percentage of their orders with high order accuracy and on time delivery date have considered and using BPEL-based integration along with XML interlinking, created real-time links throughout their entire demand chains. From the research available on this topic, it is clear that real-time integration is a must-have for attaining the Perfect Order.

Table 2: Key Performance Indicators (KPIs) Relative to the Perfect Order

Measure of Performance

What it Measures

Perfect Order

An order that is complete, accurate, on time, and in perfect condition

Demand Forecast Accuracy (DFA)

The difference between forecasted and actual demand

Quote-to-Cash Cycle Time

The time between when a quote is accepted by a prospect to when their first invoice is paid

Cash-to-Cash Cycle Time

The length of time between when a company spends cash to buy raw materials to the time cash flows back into the company from its customers. Includes the following metrics:

Ship to Customer Delivery - Time taken from shipment of finished goods to delivery at customer's address

Raw Materials Receipt to Payment - Time from receipt of raw materials to payment; also called Days Payable Outstanding (DPO)

Days Sales Outstanding (DSO) - Measurement of the average collection period from invoicing to cash receipt.

Supply Chain Transparency and Performance

Integrating customer-facing systems with supply chain, ERP, fulfillment and service systems gives manufacturers the ability to track their progress toward a Perfect Order. These metrics include:

Available-to-Promise (ATP) - Defines for both standard and customized products when the shipment will occur.

Capable-to-Promise (CTP) - Defines stock levels relative to demand to the purchase order level.

Order Visibility - Provides both at the individual order level and an aggregate view new order activity and its implications on the supply chain, existing production schedules and fulfillment.

Supply Chain Management Cost

SCM cost includes the following components: Direct purchasing operating cost

Manufacturing operating cost

Transportation cost

Warehouse/distribution center operating cost

Inventory holding cost

Customer service operating cost

Source: LWC Research

Enterprise Service Bus Case Studies

Praxair Air Products and Chemicals - Relying specifically on the quote-to-order manufacturing workflow for the development of demand-driven selling strategies that would serve to turn the companies' extensive manufacturing expertise into a competitive advantage in the market, Praxair created an initiative to first create Web Services aimed at the quote-to-order process, with further plans to complete a Field-to-Production model transformation using an SOA architecture. Inherent in this specific strategy is a strong focus on bringing an analytics layer into the total platform to measure the performance of the processes over time.

The company embarked on an SOA strategy to make their sales order entry, order management, engineering design, content management, and links to supply chain and manufacturing systems transparent. Once this was specifically achieved in a pilot program where Web Services and an analytics layer was used for the creation of a dashboard and specific functional measures of performance. In conjunction with this development, the company started working through the concepts of a broader SOA that would allow for BPEL4WS workflow analysis and functionality measurement, including a strong focus on the development of workflow analysis and quantification of results.