How to Apply Queuing Theory to Capacity Planning

    November 15, 2017

    By Scott Adams

    Despite the negative impact of overprovisioning, many companies choose to buy more IT infrastructure than they need.


    Because they hope to safeguard themselves against the possibility of downtime. So, they boost their server capacity. And while this response to the need for capacity is certainly expensive, what these companies are dealing with isn’t a spending problem.

    They have a modeling problem. 

    Anticipating application response time is surprisingly tricky. And many IT leaders are fearful that workloads will spike and down their services. So, they overprovision their hardware to ensure good response times.

    This is because application response time is nonlinear. When servers reach unknown usage levels, it’s difficult to predict how they'll respond. It’s better safe—never reaching that threshold—than sorry.

    However, there is a surefire way to predict response times accurately. And it involves a healthy dosage of applied mathematics.

    That means queuing theory. It’s the only way to avoid the double trap of overspending on resources—but incurring outages because you can’t predict performance bottlenecks.

    How to Apply Queuing Theory to Capacity Planning

    Queuing theory sounds great. But how do you actually put theory into practice when it comes to queuing and capacity planning?

    There are two ways you can go about this. You can use analytic solvers to find problems you don’t even know about yet. Or you can use event simulation for what-if scenarios and problem-solve before you make a specific change to your environment.

    Both of these methods are effective for finding a problem before it becomes a problem. 

    1. Analytic Solvers

    Analytic solvers can help you find a problem in your environment before it becomes a problem. This is a must for using queuing theory on your day-to-day operations.

    This is a great method to use when you have thousands of servers and don’t understand what they can do. By using analytic solvers, you can figure out if there’s anything you need to be concerned about.

    You can cut to the chase and focus on the servers that need your attention—and let the ones that are doing just fine coast.

    2. Event Simulation

    Event simulation lets you figure out what might happen if you do x, y, or z.

    This is a great method to use if you have a specific business scenario in mind. For instance, use event simulation if you want to understand what will happen if you add more users. Or use it to find out what will happen if you experience a higher volume of business transactions.

    Event simulation can take longer to do than analytic solvers. It won’t be your best option if you want to test thousands of servers or scenarios, but it will work fine on a small scale. 

    What You Get Out of Capacity Planning with Queuing Theory

    Using queuing theory makes capacity planning faster, easier, and more accurate. And when you have effective capacity planning in place, the impact on your organization is huge.

    User Satisfaction

    At the end of the day, it isn’t just utilization that matters. It’s the impact on the users. And if a server is overutilized, it will have a negative impact on your users.

    Using capacity planning to proactively monitor utilization will have the best possible impact on your users—none at all. Their work won’t be disrupted. And they’ll be able to carry on like they haven’t before.

    Cost Management

    One of the most common responses to overutilization is overspending to guarantee uptime. But that’s not effective from a cost management standpoint.

    Proactive capacity planning with queuing theory helps you get a handle on IT costs—before they’re blown out of control. You’ll be able to purchase the right amount of hardware in just the right places.

    Bottleneck Detection

    Do you know where your potential bottlenecks are?

    When you use queuing theory in capacity planning, you can find a bottleneck before it becomes a bottleneck.

    Here’s how. By running thousands of queuing models automatically, you can find weaknesses in your hardware—which will alert you to processes that are likely to cause bottlenecks.

    Future Readiness

    Your organization’s business strategy is forward-looking. So, shouldn’t your IT strategy be forward-looking, too?

    When you use queuing theory in capacity planning, you’ll give your IT department a proactive edge.

    That’s because you’ll be able to provision effectively for the months and years ahead—even in advance of increased demands and service spikes. And that means you can determine the health and risk of your infrastructure today, tomorrow, and for years to come.

    The Easy Way to Apply Queuing Theory to Capacity Planning

    As the older generation of capacity planners retires, skills in queuing theory are becoming increasingly rare. And learning queuing theory is hard.

    If you spent all your time learning it, you could become an expert in 6 months to a year. But who has time for that?

    So, what’s the easiest way to apply queuing theory? Choosing software that can do it for you.

    TeamQuest capacity planning software uses a queuing network model to represent real systems. The software builds the model, solves it, and evaluates the results—automatically and with greater accuracy than you could on your own.

    And with TeamQuest, you don’t need to be the queuing theory expert. In fact, you don’t even need to be a capacity planner.

    Let the software take care of the modeling and predictive analysis for you. You can take care of the rest of your job—and transform your IT department into a proactive force to be reckoned with. 

    Learn more

    Category: capacity-planning