DBR came as an antithesis to the OPT concept and it came from the author of OPT development-Dr. Goldratt himself. Instead of ultra-sophistication in trying to solve a complicated net of links between the processing steps and resources, of which several might have limited capacity (bottlenecks), a vastly simplified concept emerged: in any chain, there is one link that is the weakest. That link determines the strength of the whole chain; thus, detailed planning of that specific link should be the kernel of the overall production plan. The name given to the core planning-scheduling the one bottleneck ensuring its smooth and effective utilization-was the drum. The resulting understanding was that the bottleneck is the only resource whose efficiency really counts. However, planning the bottleneck does not ensure that the plan would be executed as is. Murphy, the symbol for everything that might go wrong, could mess things up and the bottleneck might face a situation where it has to stop processing because parts are missing. Instead of sophisticated synchronization of all resources, the concept of providing a buffer to protect the bottleneck from being starved had emerged. This buffer is not made of stock-it is a time buffer. The idea was to release the materials for the bottleneck exactly a time-buffer length before the bottleneck is supposed to begin work on the job, giving all the required resources enough time to let the parts reach the bottleneck before the scheduled time. This concept of the buffer as time-supporting the timely arrival of the parts rather than parts sitting in front of the bottleneck-was a key in understanding the paradigm shift that lies in the change from great sophistication to simplicity. It consists of understanding that buffers are necessary to deal with uncertainty, and that in order to protect a schedule, which is built of time-based instructions, we need to use time as protection. The time buffer meant that even when Murphy messes things up, the expectation is that the parts will reach the bottleneck on time in the vast majority of cases. Of course, specifying long enough time to cover for Murphy meant that in most cases the parts would arrive too early to the bottleneck and simply sit there. So, it looks like a buffer of inventory, but actually the real protector against starvation of the constraint is the time provided for parts to go through the route to the bottleneck.

The term "bottleneck" was the key term in the OPT days, and even when the DBR methodology was developed together with the famous book The Goal (Goldratt and Cox, 1984), the terminology was based on bottlenecks. It is always important and enlightening to have a historical perspective of the development of such major managerial approaches as TOC. At that time, in 1984, the much more generic term constraint was not yet coined.

OPT registered trademark of Scheduling Technologies Group Limited, Hounslow, U.K.

The important insight, partially acknowledged in the OPT2 days, but becoming clearer later, is: As complex as the production shop floor may be, the performance of the shop as a whole is impacted by a single work center, which determines both the response time and the maximum potential output of the floor.

Is there really only one capacity constraint (called CCR-capacity constrained resource), or could there be two? Well, technically, it is possible to have two, but assuming we speak about interactive resources (one feeds the other) being driven to their limits, then the performance of the shop is doomed to be unstable and even erratic because of the statistical fluctuation that inevitably occurs between dependent resources.

This chapter is not focused on DBR, but on S-DBR and the transition of the understanding that paved the way from DBR to S-DBR. We just stated one transition from OPT to DBR and the main way is still ahead of us. Before we proceed, let's fully understand three different aspects of the TOC approach. Each is material in understanding the development from DBR to S-DBR and the internal logic of S-DBR.

Three Views on Operations Planning and Execution

The basic TOC philosophy was first expressed by the Five Focusing Steps (5FS), which already explain the logic of the TOC production planning and its related BM control. The second viewpoint recognizes the difference of defining the rules behind planning in a world with a significant amount of uncertainty (planning with uncertainty) versus planning to optimize in a deterministic world. At the time of the execution whatever is dictated in planning lays out the objectives and the resulting actions. But then, there is a need to define the rules for the decision-making required to deal with the impact of "Murphy" in executing the plan. It is fascinating to realize that defining the rules for planning and execution lead to the S-DBR planning rules and the role of BM in leading the execution decision-making.

The third viewpoint looks at the achievements of Henry Ford and Taiichi Ohno and their focus on flow as the central objective of operations. It seems that even that viewpoint fully supports the TOC methodology for production planning and execution. The three different aspects bring together a better understanding of the methods and how to match them to different environments.

The Five-Focusing Steps (5FS)

The concept of the 5FS3 was developed in 1985 just as internal knowledge transfer within Goldratt's company, Creative Output signaled the emergence of the comprehensive managerial approach of TOC. It is the first time the term constraint had replaced the older concept of a bottleneck.

The importance of the 5FS (Goldratt, 1990b, 7) is that they define the rules for a "well-behaving organization." The first three steps define the state of the short-term: 1. Identify the system constraints.

2. Decide how to exploit the system's constraint.

3. Subordinate everything else to the above decision.

The longer-term steps give an umbrella for developing the scheme for growth coupled with stability: 1. Elevate the system's constraint.

2. Go back to Step 1. Warning: Beware of inertia.

For a better understanding of the DBR methodology and the transition to S-DBR, only the first three steps have a direct impact. Beginning in 1985, the three steps were extensively used to explain DBR thinking-even though the DBR rules preceded the 5FS. The first three steps are prominent in explaining the shift to S-DBR.

The Critical Distinction between Planning and Execution

The Appropriate Rules of Planning

The role of planning is to synchronize the system in a way that would enable achieving its objectives. Many times, the planning affects objectives by identifying what is realistic to achieve and what is not. Planning is viewed as the higher-level decision making, while the execution is viewed as just having to follow the planning.

There are two main difficulties for any kind of planning. One is the internal complexity in synchronizing many different variables. The other is dealing with uncertainty. The main problem in dealing with uncertainty is that planning decisions are made ahead of time and most decisions are converted to specific actions. This time difference between planning and execution allows Murphy to mess things up to a degree that the planning cannot be executed as is. The situation where in the execution phase it seems impossible, or not worthwhile, to follow the planning not only causes problems in achieving the system objectives but also generates tension between the planners and the people in charge of the execution.

Viewing the DBR methodology versus OPT might shed light on the way TOC treats the planning rules. Later, we will look at the resulting insights regarding the impact of TOC on decision rules in execution.

OPT4 was all about planning. It planned all the perceived bottlenecks in detail under finite-capacity and then went on planning the rest of the shop floor where all the non-bottlenecks were scheduled under the infinite-capacity assumption in a similar way to MRP. The hidden assumption was that there was no need to make any significant decisions at the execution phase-just follow the schedule. If Murphy messed things up, then running OPT again was the reasonable option.

DBR is a planning algorithm that is much less detailed than OPT. Only one constraint is scheduled in detail.5 All the rest of the resources are not given any schedule.6 However, the material release was scheduled in detail with the notion that the schedule for the material release meant: Do not release before!

Our current understanding is that having good planning means that, in most cases, the plan is eventually executed without changes and it draws good performance from the shop floor as a whole. Any instruction that is included in the plan but is not absolutely necessary to be taken at the time of the planning endangers the sustainability of the whole plan. The rules for what should be included in good planning are: 1. Any instruction where any deviation might disrupt achieving the objectives.

2. Such instructions must be protected from Murphy. Buffers have to be included in the plan protecting the ability to carry out the instructions.

3. Nothing else should be included in the planning.

The DBR methodology has clearly defined the critical points in the product structure that must be planned carefully. Three major control points are: 1. The due dates for all orders after careful validation that these dates are quite safe.

2. The detailed schedule of the CCR.

3. The schedule for the material release.

The criticality of the first control point is self-evident-we should not commit to dates we cannot meet. The second one is simply the essence of Step 2 in the 5FS-exploiting the system constraint. The criticality of the third control point is not self-evident. We usually see in many environments, particularly in manufacturing, that there is a lot of work-in-progress (WIP) that just sits and waits for resources. The immediate cause is the release of work too early to the shop because the first resources are available. The assumption is that the earlier work is released and starts, the higher the probability of finishing the work on time. However, once the first resources finish processing those orders they simply join the queue for the subsequent resources. The damage of having too much work without a clear and rigorous priority mechanism is enormous. While the resources upstream might be looking for work, some of the other resources might be flooded with work. When this happens, the resource under pressure tries hard to optimize its own efficiency, often at the expense of orders that are truly urgent. Actually, in most cases, the operators do not have any idea what is urgent and what is not. Many manufacturing orders are comprised of large batches. When the manufacturing orders are large, they often contain urgent customer orders and much less urgent stock orders all in one manufacturing order. Thus, too many manufacturing orders have a certain quantity that is very urgent to customers and some other quantity that is not. The loaded resource cannot do all of the manufacturing orders at the same time. Therefore, while a work center is working on a large manufacturing order with several urgent customer orders buried within it, other manufacturing orders that may also have urgent customer orders wait their turn.

The rope in the DBR planning methodology is the mechanism to ensure release of only orders that are soon required by the detailed schedules of the CCR and shipping buffers. This mechanism also forces the minimization of batching. The rope ensures against work that is not truly required being released to the shop.

The Implications for the Execution Phase

We have shown the general idea behind "minimum planning." Let's now describe the execution decision-making rules. When planning is not detailed, much more is left to the execution phase.

Including buffers in the planning has a special meaning for the people in the execution phase. The local objective in execution is to be able to execute the critical planning instructions. The state of the buffer is an excellent signal of whether things are going according to plan.

BM is the control mechanism on the progress of executing the plan. First, let's introduce my definition of the term control: Control is a proactive mechanism to handle uncertainty by monitoring information that points to a threatening situation and taking corrective actions accordingly.

The definition makes it clear that any control system is targeted to identify the actual emergence of a known threat, and it clearly belongs to the execution phase. It must have the most current and accurate information that the execution people need to carry out their jobs.

The need in the execution phase is to validate that everything is ready on time for the next critical directive based on our planning. The obvious possible threat is either being late to the CCR and thus starving the constraint or being late to the due date of the whole manufacturing order and thus making the customer order late. These two areas are protected by time buffers according to the DBR methodology.

Let's define the state of the buffer as the percentage of the time the buffer has already used (the time that has passed since the start of the time buffer). When the state of the buffer is less than 33 percent, we call the state (region or zone) green. When it is between 33 and 67 percent, it is the yellow state, and above 67 percent it is the red state.

A status of red means less than one-third of the original buffer remains and thus it is now priority one to flow the order to its destination (either the CCR or shipping).

Thus, the decision rules for the execution phase are based on the status of the buffers. BM imposes one clear set of priorities and does not tolerate any others. Thus the buffer status of any order can be checked and according to the resulting priorities every resource is able to decide what to do next. Following the BM priorities yields the highest probability that we will ship everything on time and utilize the CCR to its planned exploitation level.

The viewpoint of minimum planning requires the extra emphasis on a priority scheme for the execution phase. It fully supports the move from the excessive planning in OPT to the leaner planning of DBR, but with the addition of BM as the execution aid to achieve the objective of reliable due date performance coupled with good exploitation of the CCR. Later in this chapter, it will be shown how this view also supports S-DBR.

Concentrating on the Flow

The third viewpoint on operations comes from Goldratt's (2009) article, "Standing on the Shoulders of Giants." The flow concepts are attributed to both Henry Ford and Dr. Taiichi Ohno and highlight the TOC approach to planning and execution in manufacturing.

Goldratt7 (2009, 3) has verbalized the four concepts that lie behind the work and achievements of both Ford and Ohno: 1. Improving flow (or, equivalently, lead time) is a primary objective of operations.

2. This primary objective should be translated into a practical mechanism that guides the operation when not to produce (prevents overproduction).

3. Local efficiencies must be abolished.

4. A focusing process to balance flow must be in place.

Certainly, the four concepts for flow link the Lean concept with TOC and particularly DBR (actually, it is more attuned to S-DBR as will be explained later in this chapter). We certainly do like to get faster flow throughout the shop floor. Moreover, the rope is just another tool to prevent any overproduction. Of course, the main point is how to distinguish between overproduction and what should be produced.

The fourth concept is interesting as it can be interpreted as applying both to the immediate state as well as to longer periods. For the immediate situation, it fully supports the idea of giving higher priority to orders that seem "almost late," thus enabling faster flow for the urgent orders. We still need to develop a more global view on how to focus the efforts on improving the flow in the longer-term period of time.

Challenging the Traditional DBR Methodology

When DBR first appeared in 1984 (The Goal), it signalled a departure from the very detailed production planning process like OPT as well as a contrast to MRP. It was much later that we learned that when planning is minimal then execution gets more responsibility and it needs better guidelines for decisions. BM was mentioned first in Goldratt and Fox (1986), The Race. It took additional time for Goldratt himself and other researchers to define clearly the linkages of BM to DBR and additional time for practitioners to understand it fully. BM is a necessary condition for DBR to work effectively.

The three viewpoints poses various questions regarding the central role of an internalcapacity constraint. Let's first summarize the claims and then inquire into each of them more deeply.

1. From the 5FS perspective, there are some questions that are at the core of the challenge: a. Is the proper strategic constraint an internal resource? Should the capacity of an internal resource be the constraint of the whole organization?

b. Suppose we do have a real capacity constraint in the shop, isn't the market demand a constraint as well? If so, do we have interactive constraints and how do we handle them? In other words, how do we exploit both market and capacity constraints?

2. From the minimum planning perspective, the critical question is whether the detailed schedule of the capacity constraint is truly necessary? What would be the damage if the sequence of the constraint would not be followed as is? Do we always lose capacity in such a case?

3. From the flow perspective, the challenge is the emphasis on the CCR buffer because from the overall flow it looks like a disruption to the flow. The trigger to the flow is certainly a customer order. Do we need to create an artificial time delay at the CCR? Is it something that improves the flow or is it a blockage of the flow?

What Should the Strategic Constraint Be?

A worthy capacity constraint for being the strategic constraint is a resource whose capacity is very difficult to elevate. The difficulty might be that it is very expensive, but it could also be that enlarging the capacity is a large project because the ramifications are very substantial for all the functions within the company. Think about a basic steel company where a huge furnace is the most obvious capacity constraint. To build another furnace is a multimillion investment and it takes several years. Then, as building another furnace adds much more than a mere 2 to 3 percent to capacity (in many cases it doubles the capacity), many additional workers are required, not just for the furnace, but also because the new furnace induces elevating most of the other equipment as well.

This is probably an extreme case where the difficulty in elevation is very clear. Even in such a case, a clever CEO might find alternatives to bypass the limitation of the existing furnace by buying basic steel from other manufacturers who do not have a good market for the capacity they have. However, even in the case where the market potential is far larger than the limited capacity of the furnace, the market demand could still be an active constraint because gaining more market demand would improve the performance of the company by allowing producing and selling more of the highly profitable products instead of the less lucrative ones.

Two characteristics of the market demand make it the major practical constraint in the vast majority of cases: 1. Clients do not like to be subordinated to an internal constraint of a supplier of products or services. In most cases, the clients have an alternative supplier, and if that one offers better service, then the clients might choose to move to that supplier. Once this is done, then the company no longer has a capacity constraint.

2. When the potential is far larger than the internal capacity, then the organization can find ways to increase its throughput even without elevating the internal constraint. One obvious way is by increasing the price. Another is to concentrate on the more profitable niches of the market demand.

However, if the market demand is the system constraint, how can an internal resource be a capacity constraint?

The claim is that it is perfectly possible to have both the market demand and the limited capacity of a specific resource as interactive constraints.

Having both the market and one CCR is quite a good match. The necessary assumption is that proper exploitation of both constraints could leave just enough protective capacity on the CCR to ensure that whatever commitment is given to the market will be met. In this manner, the market constraint is given the higher consideration without neglecting the capacity limitation of the CCR.

This is actually the way to handle any situation of interactive constraints: Decide which one is the major or primary one and ensure that the other or secondary constraint would be somewhat less loaded (take less commitments on that one).

Note that a constraint can be defined as anything that cannot subordinate to another constraint and thus cannot be ignored. This definition fits the reality of most CCRs. The market is still the major or primary constraint, but the capacity limitation of the CCR or secondary one cannot be ignored; thus, fewer commitments from the market should be placed on the secondary constraint.

If the market demand is the major constraint and we see the need to have some protective capacity provided on the CCR, then there is no need to have a special CCR buffer that would protect the sequence of the production orders on the CCR. We still need to carefully monitor the load on the CCR, but not necessarily to schedule the CCR in detail. Of course, we'll expand on this point later when we describe the S-DBR process.