Throughput Quotes

People don't need enormous cars; they need admiration and respect. They don't need a constant stream of new clothes; they need to feel that others consider them to be attractive, and they need excitement and variety and beauty. People don't need electronic entertainment; they need something interesting to occupy their minds and emotions. And so forth. Trying to fill real but nonmaterial needs-for identity, community, self-esteem, challenge, love, joy-with material things is to set up an unquenchable appetite for false solutions to never-satisfied longings. A society that allows itself to admit and articulate its nonmaterial human needs, and to find nonmaterial ways to satisfy them, world require much lower material and energy throughputs and would provide much higher levels of human fulfillment.

The entire bottleneck concept is not geared to decrease operating expense, it’s focused on increasing throughput.

Alex, the goal is not to reduce operational expense by itself. The goal is not to improve one measurement in isolation. The goal is to reduce operational expense and reduce inventory while simultaneously increasing throughput,” says Jonah.

They’re measurements which express the goal of making money perfectly well, but which also permit you to develop operational rules for running your plant,” he says. “There are three of them. Their names are throughput, inventory and operational expense.

Converting a classic batch-and-queue production system to continuous flow with effective pull by the customer will double labor productivity all the way through the system (for direct, managerial, and technical workers, from raw materials to delivered product) while cutting production throughput times by 90 percent and reducing inventories in the system by 90 percent as well.

The world would be a different place if instead of competing to have the highest per capita GNP, nations competed to have the highest per capita stocks of wealth with the lowest throughput, or the lowest infant mortality, or the greatest political freedom, or the cleanest environment, or the smallest gap between the rich and the poor.

Such an AI might also be able to produce a detailed blueprint for how to bootstrap from existing technology (such as biotechnology and protein engineering) to the constructor capabilities needed for high-throughput atomically precise manufacturing that would allow inexpensive fabrication of a much wider range of nanomechanical structures.

The great majority of “consumption” (throughput) does not involve individual product users at all. For example, the average rate at which people produce waste, mentioned above by [Jared] Diamond, is calculated by dividing the total population into the total waste. But since 99 percent of all solid waste in the United States today comes from industrial processes, eliminating all household waste would have little effect on per capita waste. Diamond’s “average rate” is meaningless.

You need slack to enable continuous improvement. You need to balance demand against throughput and limit the quantity of work-in-progress to enable slack.

we have developed the following simple rules of thumb: Converting a classic batch-and-queue production system to continuous flow with effective pull by the customer will double labor productivity all the way through the system (for direct, managerial, and technical workers, from raw materials to delivered product) while cutting production throughput times by 90 percent and reducing inventories in the system by 90 percent as well.

Part of what makes real-time scheduling so complex and interesting is that it is fundamentally a negotiation between two principles that aren’t fully compatible. These two principles are called responsiveness and throughput: how quickly you can respond to things, and how much you can get done overall. Anyone who’s ever worked in an office environment can readily appreciate the tension between these two metrics. It’s part of the reason there are people whose job it is to answer the phone: they are responsive so that others may have throughput.

Improve performance through process improvements introduced with minimal resistance. Deliver with high quality. Deliver a predictable lead time by controlling the quantity of work-in-progress. Give team members a better life through an improved work/life balance. Provide slack in the system by balancing demand against throughput. Provide a simple prioritization mechanism that delays commitment and keeps options open. Provide a transparent scheme for seeing improvement opportunities, thereby enabling change to a more collaborative culture that encourages continuous improvement. Strive for a process that enables predictable results, business agility, good governance, and the development of what the Software Engineering Institute calls a high-maturity organization.

In order to find and eliminate a Constraint, Goldratt proposes the “Five Focusing Steps,” a method you can use to improve the Throughput of any System: 1. Identification: examining the system to find the limiting factor. If your automotive assembly line is constantly waiting on engines in order to proceed, engines are your Constraint. 2. Exploitation: ensuring that the resources related to the Constraint aren’t wasted. If the employees responsible for making engines are also building windshields, or stop building engines during lunchtime, exploiting the Constraint would be having the engine employees spend 100 percent of their available time and energy producing engines, and having them work in shifts so breaks can be taken without slowing down production. 3. Subordination: redesigning the entire system to support the Constraint. Let’s assume you’ve done everything you can to get the most out of the engine production system, but you’re still behind. Subordination would be rearranging the factory so everything needed to build the engine is close at hand, instead of requiring certain materials to come from the other end of the factory. Other subsystems may have to move or lose resources, but that’s not a huge deal, since they’re not the Constraint. 4. Elevation: permanently increasing the capacity of the Constraint. In the case of the factory, elevation would be buying another engine-making machine and hiring more workers to operate it. Elevation is very effective, but it’s expensive—you don’t want to spend millions on more equipment if you don’t have to. That’s why Exploitation and Subordination come first: you can often alleviate a Constraint quickly, without resorting to spending more money. 5. Reevaluation: after making a change, reevaluating the system to see where the Constraint is located. Inertia is your enemy: don’t assume engines will always be the Constraint: once you make a few Changes, the limiting factor might become windshields. In that case, it doesn’t make sense to continue focusing on increasing engine production—the system won’t improve until windshields become the focus of improvement. The “Five Focusing Steps” are very similar to Iteration Velocity—the more quickly you move through this process and the more cycles you complete, the more your system’s Throughput will improve.

Individuals should be given every incentive possible to work as a team. If the team’s throughput is increased by my helping someone else, that’s what I should do. Team velocity matters; individual velocity doesn’t.

How can we ensure that paying attention to security doesn’t reduce development throughput?

I say assertively, “Give me three lists. One that requires Brent work, one that increases Brent’s throughput, and the last one is everything else. Identify the top projects on each list. Don’t spend too much time ordering them—I don’t want us spending days arguing. The most important list is the second one. We need to keep Brent’s capacity up by reducing the amount of unplanned work that hits him.

Capacity: How much stuff will fit Throughput: How much stuff will flow They are not synonymous. All

Operational expense,” he says. “Operational expense is all the money the system spends in order to turn inventory into throughput.

By speeding up flow through the technology value stream, we reduce the lead time required to fulfill internal or customer requests, especially the time required to deploy code into the production environment. By doing this, we increase the quality of work as well as our throughput and boost our ability to innovate and out-experiment the competition. The resulting practices include continuous build, integration, test, and deployment processes, creating environments on demand, limiting work in process (WIP), and building systems and organizations that are safe to change.

Throughput is the rate at which a system achieves its desired goal.

If you don’t know your Throughput, make it a priority to find out—measuring Throughput is the first step toward improving it.

Throughput es el dinero que entra. Inventario es el dinero que está actualmente dentro del sistema. Y gasto operativo es el dinero que tenemos que pagar para que suceda el throughput. Un indicador para el dinero que entra, uno para el que está atrapado dentro y otro para el que sale.

The GNP lumps together goods and bads. (If there are more car accidents and medical bills and repair bills, the GNP goes up.) It counts only marketed goods and services. (If all parents hired people to bring up their children, the GNP would go up.) It does not reflect distributional equity. (An expensive second home for a rich family makes the GNP go up more than an inexpensive basic home for a poor family.) It measures effort rather than achievement, gross production and consumption rather than efficiency. New light bulbs that give the same light with one-eighth the electricity and that last ten times as long make the GNP go down. GNP is a measure of throughput—flows of stuff made and purchased in a year—rather than capital stocks, the houses and cars and computers and stereos that are the source of real wealth and real pleasure.

When properly designed, multithreaded programs can improve throughput by utilizing available processor resources more effectively. Using multiple threads can also help achieve better throughput on singleprocessor systems. If a program is single-threaded, the processor remains idle while it waits for a synchronous I/O operation to complete. In a multithreaded program, another thread can still run while the first thread is waiting for the I/O to complete, allowing the application to still make progress during the blocking I/O. (This is like reading the newspaper while waiting for the water to boil, rather than waiting for the water to boil before starting to read.) 1.2.2.

Throughput is a flowbased system. It measures success by the amount of quality work flowing from READY to DONE over time, not just the volume of work we can cram into our schedule.

Capacity: How much stuff will fit Throughput: How much stuff will flow They are not synonymous.

If you define the goal of a society as GNP, that society will do its best to produce GNP. It will not produce welfare, equity, justice, or efficiency unless you define a goal and regularly measure and report the state of welfare, equity, justice, or efficiency. The world would be a different place if instead of competing to have the highest per capita GNP, nations competed to have the highest per capita stocks of wealth with the lowest throughput, or the lowest infant mortality, or the greatest political freedom, or the cleanest environment, or the smallest gap between the rich and the poor.

Like that glass, a freeway ranges between 0 and 100% capacity; it can be anywhere from totally empty to completely filled with vehicles. But unlike that glass, the freeway doesn’t optimize for capacity, it optimizes for throughput. Capacity is a spatial relationship, while throughput is a flow relationship.

This approach seems completely backward to traditional managers who have been told all their lives that competitive advantage in manufacture is obtained from automating, linking, and speeding up massive machinery to increase throughput and remove direct labor.

the use of statistical process control tools to evaluate variation, correlate root cause, forecast capacity, and anticipate throughput barriers. By measuring incidence of preventable venous

Let's not forget that in the throughput world the linkages are as important as the links. Which means that if we decided to do something in one link, we have to examine the ramifications on the other links.

The concept of moving the data to the processors was an outdated process. If the data is in petabytes or terabytes, data transfer throughout the cluster would be a time-consuming process. Instead, it would be logical to move processing towards the data. That way, it would resolve the latency issue and also ensure high throughput.

Henry Ford achieved the highest throughput per worker of any car manufacturing company of his time.

Throughput is the number of transactions a system can process in a given timespan. It

Now the roads just feed into a parking system -- not a lot, not a ramp, but a system -- and lose their identity. Getting through the intersection involves tracing paths through the parking system, many braided filaments of direction like the Ho Chi Minh trail. CSV-5 has better throughput, but Cal.12 has better pavement. That is typical -- Fairlanes roads emphasize getting you there, for Type A drivers, and Cruiseways emphasize the enjoyment of the ride, for Type B drivers.

20% on detailed planning (Their poor throughput and high lead times were misattributed to faulty estimation, and so, hoping to get a better answer, they were asked to estimate the work in greater detail.)

contain into new, durable configurations, which can handle huge energy flows without disintegrating. This, we will see, is the characteristic pattern of all such thresholds. New configurations emerge quite suddenly as once independent entities are drawn into new and more ordered patterns, held together by an increasing throughput of free energy (see chapter 4).

No,” he says. “Through sales— not production. If you produce something, but don’t sell it, it’s not throughput. Got it?

The results were astonishing. No one station had noticeably poor uptime or yield, but each had some small problems. Neither uptime nor yield was at 100 percent for any station. This mapping and measurement of the process showed that if each station’s yield and uptime were just 1 percent better than current performance, they would raise the effective capacity of the whole process by more than one-third. (Remember the 1 percent improvement is compounded 38—19 × 2—times.) In other words, the way to get higher and more flexible capacity was to control the existing process more rigorously. This would reduce the cycle time—increase throughput—for existing product volume so that there would be time to run additional products. It would also help reduce the cost of variety, because the faster the cycle time, the sooner a new product variant could be set up and run through the process and shipped to a customer.

Although Sanger Sequencing is still used, it is now increasingly being replaced by newer technologies that are developing at an astounding pace. These technologies, collectively referred to as next-generation or high-throughput sequencing, allow DNA to be sequenced much more quickly and cheaply. The Human Genome Project, which used Sanger sequencing, took ten years to sequence and cost 3 billion US dollars. Using high-throughput sequencing, the entire human genome can now be sequenced in a few days at a cost of 3,000 US dollars. These costs are continuing to fall, making it more feasible to sequence whole genomes.

The six steps in the recipe are Focus on Quality Reduce Work-in-Progress Deliver Often Balance Demand against Throughput Prioritize Attack Sources of Variability to Improve Predictability

There was also Dfinity, Stellar, Tron, NEO, Steem, Loom, Waves, Tezos, and others, vying to become the dominating decentralized applications platform, or at least the preferred platform for a specific use case. Ethereum developers and fans would scoff at most of them, pointing out how they compromise on security, decentralization, or both to increase throughput.

Our throughput society depends on insecurity, gluttony, and planned obsolescence. It’s how we keep the lights on.

Well, because if we don’t have enough capacity, we’re cheating ourselves out of potential throughput. And if we have more than enough capacity, we’re wasting money. We’re missing an opportunity to reduce operational expense.

Throughput is the money coming in. Inventory is the money currently inside the system. And operational expense is the money we have to pay out to make throughput happen. One measurement for the incoming money, one for the money still stuck inside, and one for the money going out.

Through sales—not production. If you produce something, but don’t sell it, it’s not throughput.

Increase throughput while simultaneously reducing both inventory and operating expense.

CONTINUALLY IDENTIFY AND ELEVATE OUR CONSTRAINTS To reduce lead times and increase throughput, we need to continually identify our system’s constraints and improve its work capacity. In Beyond the Goal, Dr. Goldratt states, “In any value stream, there is always a direction of flow, and there is always one and only one constraint; any improvement not made at that constraint is an illusion.

The Recipe for Success presents guidelines for a new manager adopting an existing team. Following the recipe enables quick improvement with low levels of team resistance. I want to acknowledge here the direct influence of Donald Reinertsen, who gave me the first two and the last steps in the recipe, and the indirect influence of Eli Goldratt, whose writings on the Theory of Constraints and its Five Focusing Steps greatly influenced steps four and five. The six steps in the recipe are Focus on Quality Reduce Work-in-Progress Deliver Often Balance Demand against Throughput Prioritize Attack Sources of Variability to Improve Predictability

The doctor asks the patient some form of the following: “So, what is wrong?” (or, in my case, my doc always asks “So, what are your concerns?”). The doctor listens for an average of 9 seconds, then intervenes with a prognosis. The amount of time the doctor is willing to listen before intervening has gone down over time, presumably as health insurers have pressured doctors to increase throughput and as they have greatly increased the amount of paperwork required of doctors. In other words, it is in the name of efficiency. The efficiency fairies are at work in the doctor’s office to eliminate all that wasteful time spent in creating a doctor-patient relationship.

But the important thing is that we, in our plant, have switched to regard throughput as the most important measurement. Improvement for us is not so much to reduce costs but to increase throughput.” “You are right,” Stacey agrees. “The entire bottleneck concept is not geared to decrease operating expense, it’s focused on increasing throughput.” “What you are telling us,” I say slowly, trying to digest it, “is that we have switched the scale of importance.” “That’s precisely what it is,” Lou says. “In the past, cost was the most important, throughput was second, and inventory was a remote third.” Smiling at me he adds, “To the extent that we regarded it as assets. Our new scale is different. Throughput is most important, then inventory—due to its impact on throughput and only then, at the tail, comes operating expenses.

For most work centers every such switch necessitates spending time to do the required setup. Since the containers, by design, called for a relatively small number of parts the production batches that they dictated were, many times, ridiculously smaller relative to the setup required. Initially for many work centers the time required for setups was more than the time required for production, resulting in a significant drop in throughput. It is no wonder that Ohno faced enormous resistance—so much so that Ohno wrote that his system was referred to as the ‘abominable Ohno system’ from the late 1940’s to the early 1960’s.8 Ohno (and his superiors) certainly had an extraordinary determination and vision to continue to push for the implementation of a system, that for any person who looked at it from a local perspective, as most shop personnel must have, simply didn’t make sense.

The rocks and water analogy of Lean is useful for understanding how this is done. The water level corresponds to the inventory level, while the rocks are the problems disturbing the flow. There are many rocks at the bottom of the river and it takes time and effort to remove them. The question is which rocks are important to remove. The answer is given by reducing the water level; those rocks which emerge above the water are the ones that should be removed. At the initiation of the Kanban system, to achieve reasonable throughput, Ohno had to start with many containers each holding a non-negligible quantity of a particular part. Gradually, Ohno reduced the number of containers and then the quantities in each container. If the flow was not noticeably disturbed, then the reduction of the number of containers and quantities per container continued. When the flow was disturbed the Five Why’s method was used to pinpoint the root cause. It had to be fixed before the quantities could be further reduced. It took time but the end result was a remarkable improvement in productivity.

But, there is another effect that stems from restricting the accumulation of inventory. It makes it very visible to spot the real problems that jeopardize the flow—when one work center in a line stops producing for more than a short while, soon the whole line stops. Ford took advantage of the resulting clear visibility to better balance the flow by addressing and eliminating the apparent stoppages.4 The end result of abolishing local efficiencies and balancing the flow is a substantial increase in throughput. Henry Ford achieved the highest throughput per worker of any car manufacturing company of his time.

What was the reason that they gave us for not introducing the new models? Can you recall?” “Yes,” he says slowly. “They are convinced that introducing the new models will force them to declare all the old ones they’re holding in stock as obsolete. That would cause a major blow to the bottom line.” “So, we continue to offer the old stuff rather than the new. We continue to lose market share, but it’s better than to bite the bullet of write-offs. Do you understand now the impact it has on throughput?

What is obliterating the picture is that the end result of focusing on flow and ignoring local cost considerations is a much lower cost per unit. Exactly like the end result of abolishing local efficiencies resulting in much higher efficiency of the labor force. If it looks strange, it is because managers have not yet internalized the conceptual difference in guiding operations to concentrate on improving throughput rather than concentrating on reducing costs. One of the ramifications of concentrating on cost reduction is that almost all initiatives to foster a process of on-going improvement quickly reach a point of diminishing returns and as a result many deteriorate to lip service.

Performance Tactics on the Road Tactics are generic design principles. To exercise this point, think about the design of the systems of roads and highways where you live. Traffic engineers employ a bunch of design “tricks” to optimize the performance of these complex systems, where performance has a number of measures, such as throughput (how many cars per hour get from the suburbs to the football stadium), average-case latency (how long it takes, on average, to get from your house to downtown), and worst-case latency (how long does it take an emergency vehicle to get you to the hospital). What are these tricks? None other than our good old buddies, tactics. Let’s consider some examples: • Manage event rate. Lights on highway entrance ramps let cars onto the highway only at set intervals, and cars must wait (queue) on the ramp for their turn. • Prioritize events. Ambulances and police, with their lights and sirens going, have higher priority than ordinary citizens; some highways have high-occupancy vehicle (HOV) lanes, giving priority to vehicles with two or more occupants. • Maintain multiple copies. Add traffic lanes to existing roads, or build parallel routes. In addition, there are some tricks that users of the system can employ: • Increase resources. Buy a Ferrari, for example. All other things being equal, the fastest car with a competent driver on an open road will get you to your destination more quickly. • Increase efficiency. Find a new route that is quicker and/or shorter than your current route. • Reduce computational overhead. You can drive closer to the car in front of you, or you can load more people into the same vehicle (that is, carpooling). What is the point of this discussion? To paraphrase Gertrude Stein: performance is performance is performance. Engineers have been analyzing and optimizing systems for centuries, trying to improve their performance, and they have been employing the same design strategies to do so. So you should feel some comfort in knowing that when you try to improve the performance of your computer-based system, you are applying tactics that have been thoroughly “road tested.” —RK

On a slightly lower level of abstraction, we can propose the following theorem: time-space appropriation plus time-space compression equals high risk of zoonotic pandemics. Capital grows by dilating its material throughput. The more biophysical resources that can be processed into commodities and sold, the greater the profits; the greater the profits, the more resources can be acquired and so on. Capital takes hold of land where the resources sprout - a law of a tendency with few countervailing forces that can be read off from aggregate data: in the year 1700, 95 percent of the planet's ice-free land was either wild or modified and used so lightly as to be categorised as 'semi-natural.' By 2000, the proportions has been reversed.

In any system of work, the theoretical ideal is single-piece flow, which maximizes throughput and minimizes variance. You get there by continually reducing batch sizes.

Quorum leases [Mor14] are a recently developed distributed consensus performance optimization aimed at reducing latency and increasing throughput for read operations.

If you consider the total time from the moment the material comes into the plant to the minute it goes out the door as part of a finished product, you can divide that time into four elements. One of them is setup, the time the part spends waiting for a resource, while the resource is preparing itself to work on the part. Another is process time, which is the amount of time the part spends being modified into a new, more valuable form. A third element is queue time, which is the time the part spends in line for a resource while the resource is busy working on something else ahead of it. The fourth element is wait time, which is the time the part waits, not for a resource, but for another part so they can be assembled together. As Jonah pointed out last night, setup and process are a small portion of the total elapsed time for any part. But queue and wait often consume large amounts of time—in fact, the majority of the elapsed total that the part spends inside the plant. For parts that are going through bottlenecks, queue is the dominant portion. The part is stuck in front of the bottleneck for a long time. For parts that are only going through non-bottlenecks, wait is dominant, because they are waiting in front of assembly for parts that are coming from the bottlenecks. Which means that in each case, the bottlenecks are what dictate this elapsed time. Which, in turn, means the bottlenecks dictate inventory as well as throughput.

What happened was that even as throughput increased, we continued loading the plant with inventory as if we expected to keep all our workers fully activated. This increased the load dumped upon the milling machines and pushed them beyond their capacity. The first-priority, red-tagged parts were processed, but the green-tagged parts piled up. So not only did we get excess inventory at the NCX-10 and at heat-treat, but due to the volume of bottleneck parts, we clogged the flow at another work center and prevented non-bottleneck parts from reaching assembly.

Stacey points out immediately that in no case does Y ever determine throughput for the system. Whenever it’s possible to activate Y above the level of X, doing so results only in excess inventory, not in greater throughput. “Yes, and if we follow that thought to a logical conclusion,” says Jonah, “we can form a simple rule which will be true in every case: the level of utilization of a non-bottleneck is not determined by its own potential, but by some other constraint in the system.

With the bottlenecks more productive now, our throughput has gone up and our backlog is declining. But making the bottlenecks more productive has put more demand on the other work centers. If the demand on another work center has gone above one hundred percent, then we’ve created a new bottleneck.

Each of us is like an operation which has to be performed to produce a product in the plant; each of us is one of a set of dependent events. Does it matter what order we’re in? Well, somebody has to be first and somebody else has to be last. So we have dependent events no matter if we switch the order of the boys. I’m the last operation. Only after I have walked the trail is the product “sold,” so to speak. And that would have to be our throughput—not the rate at which Ron walks the trail, but the rate at which I do. What about the amount of trail between Ron and me? It has to be inventory. Ron is consuming raw materials, so the trail the rest of us are walking is inventory until it passes behind me. And what is operational expense? It’s whatever lets us turn inventory into throughput, which in our case would be the energy the boys need to walk. I can’t really quantify that for the model, except that I know when I’m getting tired. If the distance between Ron and me is expanding, it can only mean that inventory is increasing. Throughput is my rate of walking. Which is influenced by the fluctuating rates of the others. Hmmm. So as the slower than average fluctuations accumulate, they work their way back to me. Which means I have to slow down. Which means that, relative to the growth of inventory, throughput for the entire system goes down. And operational expense? I’m not sure. For UniCo, whenever inventory goes up, carrying costs on the inventory go up as well. Carrying costs are a part of operational expense, so that measurement also must be going up. In terms of the hike, operational expense is increasing any time we hurry to catch up, because we expend more energy than we otherwise would. Inventory is going up. Throughput is going down. And operational expense is probably increasing. Is that what’s happening in my plant? Yes, I think it is.

there is more than one way to express the goal. Do you understand? The goal stays the same, but we can state it in different ways, ways which mean the same thing as those two words, ‘making money.’ ” “Okay,” I answer, “so I can say the goal is to increase net profit, while simultaneously increasing both ROI and cash flow, and that’s the equivalent of saying the goal is to make money.” “Exactly,” he says. “One expression is the equivalent of the other. But as you have discovered, those conventional measurements you use to express the goal do not lend themselves very well to the daily operations of the manufacturing organization. In fact, that’s why I developed a different set of measurements.” “What kind of measurements are those?” I ask. “They’re measurements which express the goal of making money perfectly well, but which also permit you to develop operational rules for running your plant,” he says. “There are three of them. Their names are throughput, inventory and operational expense.

Throughput,” he says, “is the rate at which the system generates money through sales.

If you produce something, but don’t sell it, it’s not throughput

Money for knowledge has us stumped for a while. Then we decide it depends, quite simply, upon what the knowledge is used for. If it’s knowledge, say, which gives us a new manufacturing process, something that helps turn inventory into throughput, then the knowledge is operational expense. If we intend to sell the knowledge, as in the case of a patent or a technology license, then it’s inventory. But if the knowledge pertains to a product which UniCo itself will build, it’s like a machine—an investment to make money which will depreciate in value as time goes on. And, again, the investment that can be sold is inventory; the depreciation is operational expense.

Interesting, isn’t it, that each one of those definitions contains the word money,” he says. “Throughput is the money coming in. Inventory is the money currently inside the system. And operational expense is the money we have to pay out to make throughput happen. One measurement for the incoming money, one for the money still stuck inside, and one for the money going out.

Bottlenecks temporarily limit throughput. Maybe your plant is proof of that. But they have little impact upon inventory.” “It’s completely the opposite, Hilton,” I say. “Bottlenecks govern both throughput and inventory. And I’ll tell you what my plant really has shown: it’s proved our performance measurements are wrong.

That’s how Jonah knew. He was using the measurements in the crude form of simple questions to see if his hunch about the robots was correct: did we sell any more products (i.e., did our throughput go up?); did we lay off anybody (did our operational expense go down?); and the last, exactly what he said: did our inventories go down?

So the way to express the goal is this? Increase throughput while simultaneously reducing both inventory and operating expense.

in situations with high message throughput, where each message is fast to process and where message ordering is important, the log-based approach works very well.

Shortly after acquisition, and with accelerated research and development, the second-generation Genome Analyzer was launched, delivering reads up to seventy-five base pairs and a throughput of 2.5 billion bases of data per day. This rocketed Illumina to front and center of the global sequencing market. In 2010, an engineering advance allowed clusters to be grown on both surfaces of the glass slide (flow cell), something that dramatically increased the sequence output. In fact, this approach, labeled HiSeq, made possible the simultaneous sequencing of the genomes from five people (or, if sequencing just the genes, one hundred people). This dramatic boost in capacity led to massive uptake of Illumina’s machines in the United States, in China, and beyond.

A puzzle that remains in the field of allostery in this high-throughput era is that we have had very limited tools that allow us to answer the general question of which proteins in the proteome are allosteric and who their binding partners are. Despite Monod's characterization of the allostery phenomenon as the second secret of life, because of this important knowledge gap, as a field we are often flying blind because of our ignorance of how the key molecular players in the signaling pathways have their activity modified by other chemical agents, and because of our ignorance of the identity of those chemical agents themselves. To that end, the emergence of mass spectrometry has provided an exciting opportunity to query not only the posttranslational modifications suffered by a given signaling molecule but also, because of recent innovations, when signaling molecules have bound a given small molecule. We see that by lysing cells in the absence and in the presence of some small-molecule allosteric effector candidate, some proteins will bind that small molecule and, as a result, be resistant to limited proteolysis by proteinase K. This means that when the proteins are denatured and trypsin digested, the pattern of cuts in the polypeptide chain will be different for any protein that was bound to the candidate small molecule. Approaches such as this hold the promise of systematic identification of the allosterome for any organism and will be a critical part of our resolution of the puzzles of how the macromolecules of the cell are controlled by a battery of small molecules.

Work In Progress (the number of items that we are working on at any given time), Cycle Time (how long it takes each of those items to get through our process), and Throughput (how many of those items complete per unit of time).

Big data systems, such as data processing pipelines, tend to care about throughput and end-to-end latency. In other words: How much data is being processed? How long does it take the data to progress from ingestion to completion? (Some pipelines may also have targets for latency on individual processing stages.)

My word processor has spell-check capability, which lets me add words that didn’t originally come in its comprehensive dictionary. It’s interesting to see what words I had to add when writing this book: feedback, throughput, overshoot, self-organization, sustainability.

Throughput is always limited by a constraint in the system—a bottleneck. Optimizing performance of any nonbottleneck part of the system will not increase throughput.