The Imperative of Measuring Workload Health to Improve Technology Efficiency

Did you know you could find out the yearly greenhouse gas emissions emitted by your car (on average)? The EPA has a formula for that. Simply put, it’s the gallons of gasoline consumed multiplied by the carbon dioxide per gallon of gasoline to determine the carbon dioxide emitted per vehicle per year. Why is this good to know?

It’s important for each of us as individuals to know how much we can contribute to reducing climate change by lowering our personal carbon footprint, in this case, by opting to take public transportation, switching to an electric or hybrid vehicle, working remotely, or other options to reduce our gas consumption. (The EPA has a list of formulas and calculators to estimate other greenhouse gas equivalencies, including electricity reductions, barrels of oils consumed, and even the propane gas you use to power up your barbecue.)

Unfortunately, no such formulas exist for the technology industry. Beyond the waste created by powering data storage facilities (which require enormous amounts of electricity for cooling), there are other significant impacts on your business’s technology environment that have largely gone unchecked. There is no way to calculate what a 1% reduction in inefficiency can do to impact the workload health of your system. 

Let’s Talk About Workload Health

When a business comes to my company, Fortified, to perform an assessment of their technology environment, they want to know how healthy their technology is and where they can make efficiency improvements. Through our proprietary software, they can see KPIs on performance, efficiency, capacity, and resource health, so they can then see progress, measure success, and plan next actions. Our clients can run our tool for themselves, watching their health scores go up, just as your fitness or sleep app score will go up if you’re exercising more or sleeping better. We’re doing the same thing for data; we’re creating a workload health score. 

We’re able to look at various workload health indicators across the system to see what’s trending up and down (which systems are healthy versus unhealthy). Inefficiency breeds waste, and no one likes waste. What if an application costs one thousand dollars to run, but only needs to cost one hundred dollars to get the same results?

We consider indicators from a database perspective, a server perspective, and an overall score. And everything that’s going on in a system will be keyed off a workload health score, which will indicate when to schedule processes because it can determine when the system is busy and when there is no friction. With a deeper understanding of the data, we can be smarter about when we schedule processes. We can solve friction (think of friction as a traffic jam in your server – how can we unclog the jam so that things run more smoothly?) With the right technology, businesses can have the ability to schedule processes just as we do with Microsoft Outlook when we schedule meetings. System, when are you free? Ok, I can schedule something here…

It may not be as far-reaching as the EPA’s equivalency calculator for saving the planet, but for one company looking to improve operational efficiencies, it can contribute significantly to the bottom line.

Where Credit is Due

Many organizations are working to create efficiency standards to run everything from construction projects to business operations, food production, electricity, and much more. Trusted third-party organizations like LEED, B Corp, and Green Business Bureau provide certification programs that can prove your company’s green cred. ENERGY STAR’s certification offers a path to a more efficient data center. Even the Department of Defense is in on this. From an article in Government Executive titled, “Lawmakers want to know how much bad software costs,” a committee aide is quoted as saying:

“Because the department and the military services often have what we consider underperforming, poorly performing software and IT, … service members are wasting an enormous amount of their time which is not spent … doing the things that we need them to do as a military because they’re literally staring, waiting at their computer for their computer to load, for their Email to load, for one system to talk to another. If we could quantify that … then we could have a number that we could take and illustrate that investing in things like software and IT actually will save the department money in terms of lost working hours.”

On the micro level, I recently heard Tim Brown, Co-Founder and Chief Innovation Officer of the sustainable shoe retailer, Allbirds, speak at an event. His company is all about measuring – and sharing – their carbon footprint. They print a Carbon Footprint label with the grand total as well as emissions by category on each package shipped. I can see the day in technology when we can similarly say, this application uses X amount of carbon over the year. 

Scoring workload health is just one piece of the puzzle in terms of saving on costs and improving efficiency today. But what about tomorrow? No one thinks about what the design decisions we make today will cost them 10 years. No one thinks about the costs of storing data, let alone the licensing costs for enabling more people to have access to that data as the business grows. And lastly, no one thinks about the inefficiency of storing data in different places throughout the organization in the form of Excel spreadsheets that are not streamlined for access by all.  

Changing the Efficiency Mindset

A popular tenet of the technology industry is, If it’s not broken, don’t fix it. In my experience, if it’s not causing an outage, I don’t know of a single application I’ve worked on in which the team has ever gone back and fixed something to make it more efficient. Why? Because they’re too busy putting out fires or developing new features. No one is focused on making his or her application better.

To increase efficiency, we must reduce the resources each application process or code uses, thereby lessening the burden on the server upon which everything is running. If you’re not using the capacity, you don’t need it, at least for now. We need to shift focus to the bigger, long-term capacity plays at hand if we want to make a real impact on efficiency. And we need to define a measurement standard for efficiency across the industry and over time. If your system is more efficient, it has more capacity, meaning, it uses fewer physical resources so the response times are faster and it can process more transactions per server, thereby improving performance. Healthy applications are more efficient, leading to healthier workloads. 

The Work That Still Lies Ahead

It’s my belief that over the next five years, developing efficient code and applications, not just functional code, is going to be one of the biggest shifts in how we think about application development. This will come down to changing our acceptance criteria to include functionality, efficiency, and cost. In an ideal world, there will also be a fourth criterion that is our environmental impact, i.e., carbon footprint savings.

The majority of applications remain online for 10 to 20 years to support a business. Even today in my business, we frequently see applications hitting their 15- and 20-year anniversaries without application owners going back to refactor and optimize them. And if they’re written inefficiently, that costs a lot of money in the long term and could negatively impact users. 

There is currently no transparency into the efficiency of the code development process, nor when it is in production. There is no forecasting out 10 years, in terms of capacity, performance, or costs. I’m not saying that every application has to be rewritten. But if we start allocating 5 to10 percent of development time in a year optimizing the top five percent inefficient codes, which, based on our averages, consume 40 to 80 percent of all resources in the application, we can start to demonstrate some real progress toward making our systems more efficient and improving workload health. 

The World Economic Forum reports that the pace of technological change is much faster now than it has been in the past – just think space travel, smartphones, generative AI, Siri and Alexa. As technology advances, more data will be generated and more code will be needed to run our operations. This translates to more costs for every organization that relies on technology. This only underscores the need for us to build workload health analytics into our daily operations now, so we don’t dig our inefficiency wells even deeper as technology advances.