Current enterprise resource planning systems (ERPs) are firmly rooted in the requirements, resources and capabilities of large manufacturing concerns. But according to The Manufacturing Institute’s report, "The Facts about Modern Manufacturing" (eighth edition), about 90 percent of the manufacturers in the United States employ less than 100 people. Smaller manufacturers have adopted ERP to varying degrees but have always struggled because the systems were not optimized for smaller operations. Traditional ERP assumes a central control point for operations and a supply chain that is controlled by the firm running the ERP system. Such a design assumes a centralized relational database, a hierarchically designed communications network supporting the ERP system, and a rigidly defined set of application program interfaces (APIs) that convey production information as requested by the central ERP. It is, in most cases, a hub-and-spoke arrangement with the ERP at the hub. Smaller manufacturers, however, live in a very different environment. Their operations are cooperatives with peers and clients. The driving force is often cash flow, not necessarily production optimization. They do not often dictate standards and often have to accept data structures from suppliers or customers. In other words, there are lots of spokes with no real hubs.
Data Storage & Sharing Driving Shifts in ERP
Recent developments are driving the shape of a new kind of ERP. Systems are getting smarter. Multi-core processors are now being used for process controllers. Computer power that was once reserved for network servers is now embedded in shop floor equipment. Advances in battery life and display technologies mean that managers and workers can carry workstations with them throughout the workday. Storage prices continue to plunge. The street price for one terabyte of storage is below $50. Solid state drives (SSDs), which are highly resistant to vibration and dust, and can be used in intelligent shop floor equipment, are less than $100 for 32 gigabyte drives. Storage can now be liberally distributed throughout the workspace. All of that data must be shared to be useful. Over the last 10 years, the extensible markup language (XML) has become the de facto standard for exchanging information. XML-based systems do not need rigorously defined software interfaces. They can be used in both publish (push) and request (pull) modes. XML has the additional advantage of being able to store all kinds of information and file sizes that can comprise megabytes of information. Moving all of that information around takes bandwidth. Fortunately, bandwidth is easy to come by now. Wireless networks supporting 450 Megabits are common. Fiber-optic based local area networks are even faster and are immune to the electromagnet noise caused by shop equipment.
Moving information between organizations is also easier. All but the most rural locations in the United State have landline broadband available and satellite technology can bring high speed connections to any place on earth, for a price.
The data and the storage come together in repositories. Historically, data repositories are kept in relational database management systems (RDBMSs). RDBMSs were designed to maximize efficiency of computer resources when storage was highly centralized, processors were slow, and communication lines were expensive. The database is typically stored on a single logical drive (which may be comprised of multiple physical drives) attached to a single logical system. RDBMS data is segmented into tables that are cross referenced by relationships. Data is converted to tabular form when stored only to be retrieved and reconstituted when required. Next generation-databases do not have the same constraints. A single logical database can be spread over multiple systems located miles or even continents apart.
Convergence of Technology Creating Unified ERP
Another important difference is the introduction of dynamic modeling. Traditional ERP depends on set reorder points for stock. The reorder points are adjusted by production demands and projections. While the inputs vary with demand, the actual model is static. Advanced ERP systems consider order leadtimes and may even have a list of substitute parts and alternate suppliers but the actual heuristics used to determine quantities and order times do not change. In next-generation ERP, the models are dynamic and the system can pick which model best suits the current situation, or can chose among models if several closely agree but there are one or two outlayers.
The final piece in this whole ERP puzzle is the advent of cloud computing. Not only does that allow manufacturers, suppliers and clients to have very powerful systems available at subscription prices instead of leasing or purchasing, it can also provide a neutral site to process ERP transactions.
The convergence of all of these developments and technologies lead to a very different vision of ERP where the participants are unified, not hierarchical. This changes the nature of ERP both within the enterprise and in the firm’s supply chain. Within a firm, the ERP is not a single system, but a number of different systems that are highly distributed but densely interconnected. For example, the warehouse system handles receiving and storage. But unlike current systems, which show inventory bin location, count and lead time, a next-generation system will show all inventory, including items that are in transit from suppliers, being transferred between work locations, and reserved and staged for production. When an order is received, a complex process of negotiations and prioritization begins. The order processing system checks the inventory system (which in turn queries the warehouse system, the production system and the purchasing system) to determine which components are immediately available and which must be bought or made. Next the production is scheduled. The system first looks at the process to create the product. Current capacity is examined, including the priority of jobs that are already in process and which could conceivably be diverted, postponed, accelerated or cancelled. It is at this point that an outsourcing decision is made. The system will determine which jobs, existing or new, should be outsourced to support cash flow or maintain delivery schedules. The system looks at options like expanding capacity by adding staff or shifts. Production is constantly monitored and reevaluated. Each shop machine reports its use and capacity in real time. Jobs are dynamically rescheduled based on projected resources, including staff, equipment and materials. The difference between current systems and the new systems is that production decisions are constantly revisited and not checked only when new orders are received. For example, a change in the price or availability of certain ingredients or parts could re-balance a build-or-buy decision. A change in the price of diesel fuel could shift the balance between a closer vendor with a higher cost and a further vendor with a lower cost. Operating and depreciation costs are additional considerations, as are overtime and holiday pay. Any of these costs can be considered in current systems; but next generation ERP will take all of these into account automatically. Cost information is updated constantly throughout the day, not just once a month or once a quarter. The most profitable jobs are prioritized, and promised delivery dates and previous client performance are also key factors.
Firms As Suppliers & Buyers in Next-Gen ERP
Relationships with other partners in the supply chain are just as dynamic. While relationships will always be an important characteristic in business, the new forum will be the reverse auction. In a reverse auction, the buyer describes the required goods and the prospective sellers bid on price and terms. Although not new, reverse auctions are gaining favor for procurement with online sites working as facilitators. The Department of Homeland Security was buying scanner maintenance parts on the Government Accounting Office’s FedBid reverse auction site more than four years ago. This month, The Wall Street Journal reported that large firms such as GlaxoSmithKline and Toyota were using reverse auctions to constrain costs of legal services. In tomorrow’s ERP environment, most firms will be both suppliers and buyers, sometimes at the same time. Firms will be buyers of raw materials and subcomponents, and suppliers of finished goods and sub-assemblies to other firms. The key difference from current systems is that most of the routine buying and selling will take place with limited or no human intervention. The systems will be pre-programmed with limits for prices and terms. Then, they will negotiate with each other, seeking human guidance and conformation only when necessary. Price will not be the only consideration. Delivery times, location, past performance, and corporate and national politics may all be factors. Favored firms may be given special consideration but would still have to compete in the reverse auction. Firms may be both suppliers and buyers in the same transaction when the ERP system is making build-or-buy calculations for subcomponents. Of course, supplier safeguards must be in place to prevent order cascading. If a firm receives queries about the ability to provide 200 widgets within six weeks from seven new and different prospects, it might not mean that widgets are now more fashionable and widget production should increase. It most likely means that all six firms are bidding on the same contract. If management decides to use more traditional procurement methods, the next generation ERP systems will still provide an advantage. For example, the database system can make all of the distributed data from all of the distributed systems, even those from suppliers and customers (with appropriate restrictions and security) available in one top-level view. Managers could make decisions based not only on vendors supply and customer’s demand, but on vendor’s projections and customers expectations.
Thumbnail image courtesy of Leszek Kozlowski.