Management at the Edge in Industry Verticals


Moin Moinuddin and Atanu Banerjee
Microsoft Corporation

February 2006

Applies to:
   Systems Management
   Management Architecture

Summary: Adapt rapidly to changing business conditions by bringing new products and services to market, adjusting internally to support business changes, and tightly controlling risks with Microsoft IT management products. (21 printed pages)


What Is Management?
Microsoft in Management
Management in Retail
Management in the Manufacturing Industry
Management at the Edge Using Today's Tools
Management at the Edge Using Tomorrow's Tools
Relevant Industry Standards


Enterprises today need to be agile and flexible enough to rapidly adapt to changing business conditions by bringing new products and services to market, adjusting internally to support business changes, and tightly controlling risks. Management is critical in achieving these goals. The IT department must be able to monitor systems and discover problems and resolve them before they affect the operations; however, the legacy infrastructure used for management is consuming the majority of the enterprise's IT budget in costly maintenance activities, leaving few resources to make the IT advances required. This paper provides the importance of management at the edge in the industry verticals and how Microsoft products can help in this area.


The vertical landscape is changing rapidly, be it retail, manufacturing, or finance. The edge, where the business interfaces with the customers, (store, manufacturing floor, banks) is becoming more and more complex with the addition of new tools, devices, and technologies. Vertical enterprises are depending on continuous, smoothly-functioning, and available systems at the edge at all times to meet the demands of their customers. This is putting tremendous pressure on the IT departments of the enterprises to ensure that the devices and systems at the edge function with minimal interruption. In addition, IT departments must seamlessly roll out new devices and systems to provide better customer experience.

As emerging technologies bring a generation of new smart devices to the edge, enterprises become more dependent on these devices, and their smooth and consistent operation becomes increasingly important. As the Web services proliferate into the enterprises, it will enhance the ability of an enterprise to remotely use devices at the edge. This will also lead to proliferation of servers at the edge. One could say the edge is becoming increasingly "thicker." So the edge is becoming critical to the operation of the enterprise.

This document covers the importance of the management in the industry verticals and some of the best practices and new developments. Specifically, we discuss management in two forums: retail and manufacturing.

What Is Management?

Systems Management refers to the process of configuring, updating (patch management), deploying, and monitoring the health of a device in real time from an external server. In a well-managed enterprise environment an application or an operator should be able to monitor the health of any device or application, maintain the health of a device, and take corrective action when needed. Since devices and systems at the edge of the enterprise and at headquarters both play a critical role in an enterprise's operations, keeping the devices and systems in service is a critical mission for the enterprise in general and the IT department in particular. Also, as devices play a critical role in enhancing the customer experience, keeping them in good health means keeping additional sales tools in front of the customer, or otherwise ensuring the smooth functioning of the manufacturing facility at the edge.

Systems Management refers broadly to configuration, deployment, monitoring, and maintenance of a device or application. Each of these categories is explained below in detail:

  • Configuration Management—the ability to automatically or remotely configure or update the configuration of devices, servers, systems, desktops, and so forth.
  • Software deployment—the ability to update software or firmware on a device or system. Updates can be to existing software or the installation of completely new software application.
  • Monitoring—the ability to monitor the health of a device in real time. Based on the defined thresholds, this refers to the process of monitoring the events generated, and responding to the alerts that were triggered as a result of a certain event. You can monitor hardware health, line of business applications (LOB), and so forth.
  • Asset tracking and inventory—for both hardware and software. Due to the increase in complexity of enterprise IT infrastructure, asset tracking is critical to not only tracking the inventory but also in making decisions around acquiring new devices and applications based on load and performance.

For devices and systems at the edge, to be managed well, these functional disciplines must work together at all times. As enterprises adopt emerging technologies, designing and implementing the right management solutions is critical.

Microsoft in Management

Microsoft has multiple tools to help customers manage their systems. Systems Management Server (SMS) helps large IT organizations manage client computers and servers in the corporate network, manage POS devices, manage other devices in the stores, manage sensors or RFID devices in a manufacturing plant, or manage ATMs at the branch of an enterprise bank. These management tasks include the distribution of software, the delivery of software updates and patches, asset management, and site configuration. The second important product offered by Microsoft in the management space is Microsoft Operations Manager (MOM), which is an event and performance monitoring tool for Windows servers. It enables the monitoring of events from systems and applications, provides event management capabilities such as alerting, and provides reporting and trend analysis.

MOM 2005 is designed to help users during the three key phases of dealing with server issues:

  • Identifying the server problem
  • Understanding the details of the problem
  • Resolving the problem

MOM provides the following key capabilities that help enterprises in managing their devices at the edge:

  • Alerts call attention to critical events that require administrator intervention. Alerts are created by rules that are the core of MOM monitoring capability, and specify the particular aspects of server health that must be monitored to ensure that the server is functioning properly.
  • State monitoring is an at-a-glance view of the state of devices and applications. It also details the health of each component of the application.
  • Views provide targeted drill-down details about device or application health to allow users to investigate device or application-specific health indicators, including performance data, collections of specific events/alerts, and groups of servers.
  • Tasks enable administrators to investigate and repair issues from the MOM Operations console. Tasks provide a key tool to facilitate rapid issue resolution, by allowing the Management Pack author to automate the common steps used to diagnose and resolve issues within the context of the affected server.
  • Reports allow historical data analysis to allow users to assess the long term trends in operational and application performance.

Downloads and Online Resources

Management in Retail

The retail landscape is changing rapidly. Only a few years back, retail stores used to be stand-alone unique environments. This is no longer the case. A retail store is now a complex IT environment in itself, made up of a myriad of devices such as PC-based POS terminals, handhelds, RFID readers, manger works stations, local hosts, local database servers, and so forth. This expanding array of devices at the edge is playing a critical role in enhancing the customer experience but is bringing its own set of challenges to the management. More importantly, it has become an extension of the bigger IT space that makes up the retailer's enterprise. A comprehensive IT management infrastructure in which all the resources can be managed, regardless of where or what those resources are, is key to getting a handle on the management at the edge. Also key to this is that new devices must be able to integrate with their existing management tools. Those new devices must therefore be immediately manageable without having to place new requirements on their management application vendors.

Combined with these reasons and some other reasons we will discuss, management at the edge is becoming critical.

The Coming Move to Processing at the Edge

Driven by falling PC prices and an increasing demand for more processing power, connectivity, and the ability to support many devices, PC-based point-of-service (POS) terminals are more often replacing the traditional stand-alone cash registers. PC-based POS terminals have been gradually absorbed into retail IT infrastructure. These devices can link into a range of back-end systems. Retailers have driven the evolution of POS hardware, but few regard POS devices as similar to other client computing devices elsewhere in their organizations. In this changing environment, retailers can significantly reduce IT management overhead by viewing POS devices, handhelds, and other devices as mission-critical PCs. By viewing POS devices as mission-critical PC devices and applying PC deployment best practices, most retailers can achieve significant reductions in IT management overhead and costs.

Device Proliferation and Enhanced Shopping Experience

Since the technology boom of the 1990s, retailers have been experiencing a shift in the technology landscape. This boom has brought with it new devices and new customer expectations, which are forcing retailers to use devices to enhance the overall shopping experience for the customers. Since devices play a critical role in the customer experience, their availability is business-critical.

Before the technology boom retailers had POS devices intended for use only at the end of the shopping experience. This is changing quickly, and now retailers are surrounded by devices that enable and support the entire shopping experience. Customers now are presented with technology at every step of the in-store experience. Retailers are using devices such as self-service kiosks/concession stands, cart-mounted wireless devices, credit card scanners and signature capture devices, biometric information readers, hand-held customer and attendant facing units, portable personal assistants for managers, informational Kiosks, automated signage, RFID devices, and many more, which play a critical role in enhancing the shopping experience by enabling the consumers to find the items they are looking for. This in turn increases the up-sell and cross-sell opportunities for the retailer.

While enhancing the shopping experience for the customer, these devices do bring a unique set of challenges. One of the most critical challenges is monitoring their availability around-the-clock. IT departments therefore need management tools to provide high availability service levels.


Figure 1. Device Proliferation

Customer Demand

Today's business climate is defined by constant change and increasing competitive pressures. As companies strive to maximize business results through growth and increasing profit margins, they face a more complex set of challenges than ever before. No longer can organizations expect to thrive by relying on static business models that worked in the past. Rather, competitive advantage is now determined by how effectively organizations can evolve to meet the ever-changing requirements of the marketplace. One of the key changes in the marketplace that retailers face is the demanding customer. Customers not only demand more responsiveness, but they are also more acutely aware of competitive alternatives. To differentiate themselves and keep the customers loyal, retailers are constantly looking to enhance the customer experience by providing information where and when they need it. IT plays a critical role in this.

Now more than ever, IT investments must increase customer impact and drive business value. They must do this while leveraging existing assets and without large up-front costs. And, just as retailers have evolved from stand-alone stores to connected enterprises, so too have the applications that run them. In recent years, applications have advanced from disparate islands of functionality to highly connected systems. These connected systems require availability of mission critical systems at all times.

Availability of Broadband and Wi-Fi

One aspect of this pressure on the IT infrastructure is better availability of broadband and Wi-Fi connectivity everywhere. The trend is such that most of the enterprise retail stores are connected to the enterprise with a broadband connection and have Wi-Fi connectivity for high bandwidth handheld devices. Even though the retailers are reaping the benefits of migrating to the broadband and Wi-Fi connectivity, they also realize the challenge of keeping them available at all times.

IT Complexity

An inoperative POS terminal directly results in the loss of a sale in many cases, as customers often don't have the time for store personnel to resolve the problem. In addition to the lost revenue on a sale, problems are compounded with dissatisfied customers and inefficient operations as store personnel attempt to resolve problems.

These devices, while enhancing the shopping experience for the customer, do bring a unique set of challenges:

  • Seamless and complete integration with the rest of the infrastructure, including not only devices located at the store but also within the enterprise.
  • Reliability and availability.
  • Monitoring the health and well-being of the devices.
  • Patch management, system upgrades, and overall management.

Retailers are facing rising operational costs in this distributed enterprise. This is becoming a critical issue, and the ongoing management and maintenance of various devices and client systems is a major cost contributor. Managing desktop deployments and the deployment of devices is putting pressure on operational costs. Systems management tools are designed to help customer service organizations reduce desktop and mobile support costs while increasing customer service levels and improving productivity.

Click here for larger image

Figure 2. IT-Complexity

Integration with the Enterprise

With greater functionality than ever before, POS devices are the front line of retail IT infrastructure. Integration with supply chain management, enterprise resource planning (ERP), and accounting applications is a goal for most retailers. Development continues, with many looking to integrate customer relationship management (CRM) and human resource (HR) applications. These trends are morphing the stand-alone POS device into a mission-critical customer facing terminal. These reasons put additional pressure on the IT departments to ensure the availability of these devices at all times. Retailers' revenues are directly linked to the availability and proper functioning of these devices.

In addition, the stores are requiring the better use of the local host, so many LOB applications such as inventory management, workforce management, and so forth are being deployed at the stores. Manager workstations are also experiencing an expanded role for communication and collaboration with the headquarters on pricing, promotions, training, and so on. So the retailer's IT department must be equipped to detect and resolve problems before they bring down the mission critical devices in the complex retail environment.

Although POS devices, the local host, and the manager workstation used to be based on proprietary operating systems, retailers are migrating to POS devices running a PC OS such as Microsoft Windows XP and Windows 2003. Many newer POS devices are based on Windows XP embedded, WEPOS, and similar systems. Local host and manager workstations are increasingly using Windows 2003 server.


Figure 3. Retail Application Landscape

Management at the Edge Using Today's Tools in Retail

Historically, integration of management of the store with the exsiting management solution has been lacking. However, flagship products of MOM and SMS provide the capability to combine the management of corporate systems with store or branch management. These products create a powerful management environment that centrally monitors the devices and systems at the edge and responds automatically to correct problems before they result in downtime or affect customer service. SMS further enables system administrators to deploy new applications, operating system updates, and data files to thousands of remote devices automatically and remotely. This allows for centralized easy packaging, distributing, and installing of software on an enterprise-wide scale. SMS can also enable enterprise-wide POS system configuration information to be readily available to the people who need it by allowing them to automatically scan each system for vital management information. This information is stored in a central database and the inventory component provides powerful facilities to report on this data.

MOM provides management solutions to monitor and report on a diverse set of devices such as POS or handheld devices, helping the retailers to fix problems before they even occur. The centralized MOM console is one of the best features of MOM. Here, a single view provides full visibility into the entire enterprise. This convenient console indicates any problems as soon as they occur. For instance, MOM will show in real time when a controller is starting to take a lot of errors. So, even if it is self-correcting the errors at that point, the retailer's IT department would know that they need to look at it before it fails. This saves many man hours of sniffing through the logs searching for problems after the fact.


Figure 4. The high level systems landscape for an enterprise retailer

One of the critical features of MOM is the Management Packs, which are used to monitor specific applications. Management Packs apply the discipline of monitoring to a specific application or technology. Each Management Pack includes the rules and rule criteria, tasks, views, and reports that are tailored to monitor the services provided by the application or technology. A Management Pack consists of the processing rule groups, and associated computer attributes, computer groups, processing rules, responses, scripts, public views, and a knowledge base for a monitored application or set of applications. Management Packs are stored in files that can then be imported into other installations. Management Packs are easy to create using the MOM Administrator console, unless the pack includes rules with scripts.

Management Packs are provided by the application vendors most of the time. For example, Microsoft provides Management Packs for Microsoft exchange server, SQL server, and so on. A Management Pack transfers knowledge from the product experts designing a Management Pack for a given application to the application users in the context of a problem on the user's server. Management Packs provide rich and detailed knowledge to the MOM users. This rich knowledge empowers them to be more productive, and can deal with issues in the initial levels of support without requiring escalation to more costly support.


Figure 5. MOM Management Packs

In the retail scenario, Management Packs can be used to monitor various key indicators; average check-out time, for example. An alert can be generated if the average check-out time goes over a certain threshold or an average credit card authorization time goes up over certain threshold, in which case perhaps the solution is to reduce traffic on the broadband connection from the store.

Management Packs provide many benefits, including rich alerts that will now contain additional information such as summarization of the issue, likely root causes, and steps for investigating and resolving the issue. Retailers should take advantage of the new capabilities in MOM 2005 for managing their complex environments.

Implementing MOM in Retail

As enterprise retailer environments grow, the complexity and challenge to monitor and maintain them increases. There are many ways to architect the MOM implementation to cope with the complex enterprise retail environment. This section describes a tiered approach. It is critical to model the enterprise environment so that the infrastructure is load-balanced to use the resources optimally.

Figure 6 shows a two-tier hierarchical MOM architecture. The architecture has two tier-2 Management Groups and one tier-1 Management Group. Within the two tier-2 Management Groups, one is dedicated to the corporate environment and one is dedicated to the retail stores environment. Based on the complexity of the retail enterprise there can be more tier-2 groups. For example there could be three tier-2 groups with one dedicated to the corporate environment and the rest of the two dedicated to the retail store environment, where one of them could for specialty retail stores and another one for general merchandising stores. All the servers in tier-2 will forward their alerts to the upper-tier for a consolidated alert view across the enterprise. The corporate environment maintains itself as a separate entity and does not forward alerts to the upper tier. To implement this type of multi-tier architecture Microsoft Connector Framework and MOM-to-MOM Product Connector may need to be utilized.


Figure 6. MOM Implementation in Retail

Each Management Group consists of a SQL Database instance and one or two Management Servers. Within the two Management Groups one is for the retail stores and the second one is dedicated to corporate environment. The MOM infrastructure also includes a reporting feature. The reporting function is implemented centrally instead of locally on each Management Group. This solution allows the storing all historic data centrally, and also allows report queries from one central data warehouse.

Implementation of SMS in Retail

Enterprise retailers have from hundreds to thousands of stores. Combine this with the fact that each store has myriads of devices, and management becomes a challenge. To simplify the management of large groups of devices, we recommend using SMS to structure the computers into groups of sites and using SMS domains. An SMS site is a set of one or more SMS domains that exist on the same local area network (LAN). An SMS domain is a set of servers and clients that are managed as a group within an SMS site.

Enterprise retailers can organize SMS sites in many ways; this paper describes the most common model in the following section. Organizing into SMS sites also helps to load balance the enterprise environment so that resources are used optimally.

In a hierarchical model you create parent and child sites. A parent site is a site (central or primary) that includes other sites beneath it in the hierarchy. A child site is a site (primary or secondary) that reports to a site above it in the hierarchy.

In enterprise retail, SMS sites can be organized based on criteria such as:

  • Business function (POS, Local Host, Database, and so forth).
  • Region of the store (state, city, and so forth).
  • Concepts (specialty, general retail, and so forth).

For example, figure 7 shows the hierarchical organization based on the type of business function. There is one primary site dedicated to the corporate desktops, another one for the stores. Based on the complexity of the retailer there can be other primary sites based on business type such as a primary site specialty retail stores, and another one for general merchandising, etc.


Figure 7. SMS Implementation in Retail

Secondary sites do not require databases, and rely on primary sites for storing the data. This model provides autonomy for each of the primary sites to apply different administrative requirements, distribution patterns, and escalation paths. For example, the event alerts from the specialty stores are routed to different personnel who are available 24 hours a day, compared to the corporate desktop event alerts, which may go to a different group to resolve the next morning. In case of a disaster, the impact is isolated to the secondary sites within the primary site, and this model scales well with the growth of the retailer.

Management in the Manufacturing Industry

Engineers tasked with overseeing operations at manufacturing facilities know that it is critical to monitor and manage their IT systems constantly. These systems could range from field devices such as sensors and controllers, to factory execution systems like a manufacturing execution system (MES) or a distributed control system (DCS), all the way up to LOB systems that are interfacing with the execution systems.

Manufacturing IT systems are already complex today, and such complexity is only likely to increase in the future. A new generation of "smart" field devices is coming, and this will spread technical innovations away from centralized IT data centers and closer to the production systems that have traditionally been the network edge for IT. Not only will there be more computing power available at the edge, there will be a lot more data generated there. Such systems might include wireless devices, RFID readers and tags, and even more exotic forms of sensor networks that are currently being researched. In addition, the increased proliferation of Web services will make it possible for these "smart" field devices to offer services that can be invoked remotely. Thus these devices will not just be clients, and there will be a proliferation of servers along the edge, as well.

The increasing use of complex technologies alongside production systems requires a higher level of system integration and process automation than before. This may be driven by the need for local efficiencies and cost reduction at the factory level, or it may be driven by corporate plant-to-enterprise (P2E) initiatives aimed at synchronizing the shop floor and the top floor (that is, real time synchronization between production and enterprise systems). All this has made attention to IT architecture a key success factor for manufacturers, and typically this will require the use of multiple technologies that must be accommodated by a single IT organization. This will require a strong focus on monitoring and managing these multiple technologies in a way to reduce complexity, cost, and downtime.

As a guideline to managing technologies in the process industry, the OPC Foundation defines three levels of management:

  1. Field management—Relating to "smart" field devices, such as sensors, actuators, or mobile devices.
  2. Process management—Typically control systems such as DCS (Distributed Control Systems) and SCADA (Supervisory Control and Data Acquisition).
  3. Business management—These relate to Line Of Business applications such as enterprise resource planning (ERP) and Supply Chain Management systems that drive production.

This is a convenient categorization that can be applied to manufacturing systems in general. This paper will discuss the management of systems that fall under the first two categories, but will not go into the management of business applications.

Systems Landscapes for Manufacturing

The systems landscape for a typical manufacturer is shown in Figure 8. Some of the important terms are explained here:

  • Enterprise LOB applications—corporate LOB applications such as ERP, SCM, and CRM.
  • Factory LOB applications—LOB applications at the level of a manufacturing facility.
  • MES—Manufacturing Execution System.
  • SCADA—Supervisory Control And Data Acquisition. A distributed instrumentation and control system.
  • HMI—Human Machine Interface. The user interface (typically graphical) for operating and monitoring the control system.
  • DCS—Distributed Control System. A networked system that allows remote monitoring and control of industrial equipment through connected:
    • Sensors—typically a measuring device for operational data.
    • Controllers—Programmable Logic Controllers (PLCs) are one example.
    • Actuators—Devices that transform electronic signals into motions.
    • Operator Terminals—these could be desktops distributed throughout the shop floor, or even mobile devices which can connect wirelessly to sources of information.


Figure 8. High level systems landscape for a typical manufacturer

The Coming Move to Processing at the Edge

Technology advances are enabling a new generation of field devices, and soon there will be a lot more computing power distributed close to the edge. These changes will be driven by the rapid adoption of RFID, the emergence of new and powerful mobile devices, and in the future by research currently underway on sensor networks. As more applications and devices are deployed at the edge, there will be a greater need to manage these devices and the data streams they generate. Thus the proliferation of edge devices can be expected to push the deployment of edge servers. In general, these servers will look like Figure 9.


Figure 9. Proliferation of new types of edge devices will push deployment of edge servers into the IT systems landscape

Take, for example, the case of RFID technology, which is making it possible to gather greater quantities of data at the edge. Microsoft is extending its platform to provide the services in Figure 9 for RFID—here the devices corresponding to the lowest layer in the stack will primarily be RFID readers. These readers will be generating events as they detect RFID tags, and then these events will get passed along to the edge server for processing. Within the edge server, an RFID service will deliver the device abstraction, device management, and event processing capabilities.

The device abstraction layer in the RFID service will ensure that devices from different vendors look the same as the device management and event processing layers, and also to the business applications being fed by the edge server. The device management layer will provide a single, consistent way to deploy, configure, and monitor RFID devices. The edge processing layer will make it possible to filter, transform, and aggregate events raised by RFID devices, to ensure that business applications see events in the proper context. Further complicating the matter is the fact that, as devices get smarter, some of the processing from the edge server will get offloaded to the devices (filtering, for example) and the capabilities of devices will vary across vendors. The device management layer of the edge server, however, would need to expose these varying capabilities through a consistent interface in order to manage devices collectively.

Another expected disruptive change is the broad industry support emerging for Web services. This will start a trend to introduce Web services at each of the different layers of the manufacturing system in Figure 9. For example, "smart" field devices (sensors, actuators, controllers) might expose Web services interfaces. These could be functional interfaces for setting and retrieving information, or could be management interfaces to configure and deploy the device. Client applications (an HMI or a management console, for example) could then connect directly to the device. Some standards bodies such as the OPC Foundation are releasing Web services specifications to ensure interoperability among vendors.

Need for Systems Management at the Edge

For certain deployments, the cost of having either edge servers or edge devices fail could be extremely high. This would be especially true if the control system were to fail, as in addition to monetary costs, there might be loss of lives. Even if lives were not lost, there could be considerable financial loss through equipment failure or product quality issues.

Consequently, the system needs to be able to monitor itself and raise alarms as needed, and the hardware needs to be rugged enough to withstand any harsh environments at the manufacturing site. Reliability and high availability of parts are critical, and this may be achieved through redundant hardware and communications channels.

Management at the Edge Using Today's Tools

MOM can also be used to monitor a collection of devices at the edge, and their corresponding edge server. For example, consider the case in which the devices are RFID readers, and there is an associated RFID service running on the edge server, corresponding to Figure 10. In this scenario, events can originate from any one of the devices, or from the RFID service itself. Events from RFID readers get passed along to the RFID service as well, and so the service becomes the conduit through which all events are surfaced. These events can be viewed with the RFID console, which is an MMC snap-in. However, these events could also be viewed through the MOM management console, which will show events across the entire landscape of deployed servers and devices.


Figure 10. Integrating an RFID service to MOM

Potential Benefits

Enterprises want to integrate their production systems with their business systems, both to achieve significant increases in productivity and to drive decision making from the factory floor. To do this, plant personnel should be able to monitor the health of their systems, and to be alerted to exceptions and potential problems. They should also be able to gain insight into the business metrics that will be affected by problems in production. This requires sophisticated yet easy to use management tools, which can scale across complex systems landscapes.

Also, as sensors and other devices become more sophisticated and pervasive, the potential to be able to monitor hard-to-reach manufacturing equipment increases, which gives manufacturers the potential to improve their industrial automation and to drive down costs. For example, the Air Force Research Laboratory was awarded a patent to monitor conduits such as electrical wires and hydraulic lines. This allows the detection of stresses to the conduit and alerting to potential leaks, while avoiding time consuming manual inspections. To make this even better, these smart devices at the edge could be connected into management technologies like MOM and SMS through edge servers, with unified management consoles and the ability to distribute reports.

Management at the Edge Using Tomorrow's Tools

The Dynamic Systems Initiative is a Microsoft-led industry effort to reduce complexity and greatly improve manageability in today's IT systems. Changes are being made to both development tools and management tools in order to be consistent with this initiative. The goal is to help customers reduce their spending on managing IT systems in the future. As the bulk of operational costs is a consequence of how the system was designed, a core tenet of this initiative is "design for operations." This is analogous to product manufacturing. David M. Anderson made a study of this in his book, Design for Manufacturability: Optimizing Cost, Quality, and Time-to-Market, in which he divided the process of new product introduction into five stages—concept, design, testing, process planning, and production. He showed that while only 8 percent of the cost of introducing a product had been incurred by the end of the design phase, 80 percent of the total cost had already been implicitly committed based on design decisions taken in the first two phases.

Using These Tools

The fundamental building blocks of DSI are the System Definition Model (SDM), which is the generic ability to model systems, and WS-Management, which is the generic ability to communicate with those systems.

To deal with heterogeneous IT environments, these building blocks use an open modeling framework, open protocols, and open interfaces.

The SDM is a formal model of a complete system, in a machine-readable format, with all the information required for deployment and operations. It is intended to capture information relating to system topology, application constraints, IT policies, installation directives, a health model, monitoring rules, service level agreements (SLAs), and reports. An SDM model should span the IT lifecycle to unify the operational policies of IT with the operational requirements of applications. It is relevant at both design-time—when an application is being created—and at run-time—when the application is being deployed and managed in your IT environment.

For example, as part of the process of building an SDM, a health model is developed for relevant services such as those exposed by devices or edge servers. This health model contains a state model, with states like Healthy, Stopped, Service Unavailable, and Service Partially Available, for example. The health model shows all possible transitions between these states, and the situations in which these transitions may occur. The health model is used in diagnosis; to do root-cause analyses of why a system is in a particular state; and to determine the steps needed to move to another, more desirable state. Once a health model for a set of production systems has been developed, instrumentation for monitoring can be aligned with this model.

A high level representation of how the DSI can be used for monitoring and troubleshooting is shown in Figure 11.


Figure 11. Using the DSI for monitoring and troubleshooting

Potential Benefits

The benefits from the DSI approach will come to all participants in the IT lifecycle—engineers, IT professionals and system administrators, applications developers, and business operations owners.

For example, engineers can monitor and manage operations across the entire breadth of a manufacturing facility from a single terminal. In fact, if this terminal was on a mobile device or a laptop, they could be doing this off-site, as well. Taking this one step further, an engineer could be monitoring meters, relays, pumps, RFID readers, and more, at multiple facilities from this device. When a problem occurs in some part of the system, the system definition model can be used for diagnosis and troubleshooting.

Relevant Industry Standards

There are many companies that are building systems for industrial automation. Although this situation provides customers with a lot of choices, it lays open the possibility of many different heterogeneous interfaces and proprietary protocols. This would make it difficult for customers to interconnect systems without a lot of custom "glue" code being developed. It is to fill this gap that standards bodies have evolved, to provide consistent interfaces and APIs for applications and devices to communicate and exchange data. This also improves productivity for the vendor community, as they only need to develop to a standard set of interfaces, and not to a collection of proprietary ones.

A very high level listing of relevant standards bodies in this space is provided as a starting point for further reading:

  • ISA-95—Enterprise/Control System Interface—Plant to Enterprise integration, Manufacturing Execution System Functions.
  • WBF B2MML—XML Schemas based on ISA-95—Business To Manufacturing Markup Language.
  • OPC—DCOM and XML interfaces for Industrial Communications and Automation.
  • MIMOSA—Asset Mgt and Maintenance Mgt Schema, Meta Data, and Interfaces.
  • Open O&M—Joint work by MIMOSA, OPC & ISA-95 to integrate operations and maintenance information.


About the authors

Moin Moinuddin is a member of the Industry Solutions Architecture team at Microsoft and is responsible for the retail solutions architecture.

Moin joined Microsoft from Lightbridge/Infospace where he worked in various roles for 6 years. At InfoSpace Moin was instrumental in designing and implementing one of the largest Payment Gateways for processing payment transactions from eCommerce merchants. Prior to joining InfoSpace, Moin worked as a developer, lead and the development manager at Progressive Software, Inc. designing and implementing Point-of-Sale solutions for the hospitality industry.

Moin received his M.S. in Computer Science from the University of North Carolina at Charlotte and holds bachelors from Osmania University, India. Moin lives in Bellevue with his wife and two kids.

Atanu Banerjee is a member of the Industry Solutions Architecture team at Microsoft, and is responsible for manufacturing solutions architecture.

Atanu joined Microsoft from i2 Technologies, where he worked in various roles for more than 7 years. He was chief architect for their supply and demand management product line, and prior to that served as development manager, product architect, team lead, and software developer for various teams. During that time he wrote a lot of code, designed new solutions, and worked with some large manufacturing customers.

Prior to joining i2, Atanu worked at Aspen Technologies in the advanced control systems group, designing and implementing model predictive control systems for the process industry.

Atanu received a Ph.D. from Georgia Tech in 1996, and holds a bachelors degree from IIT Delhi, India. He resides in Redmond, WA with his wife and son.