Bancolombia plans an investment of COP 1.6 trillion (~USD 432 million) in IT; this is a staggering figure. To put this in perspective, that budget could theoretically purchase some 450,000 top-of-the-line smartphones or nearly ten 1980s-era Cray X-MP supercomputers (see Cray X-MPs at current USD.)  But as any systems architect knows, throwing hardware at a problem won’t fix a bottleneck governed by Amdahl’s Law.

Why do I specifically mention these brand and device models?  First, I have been Bancolombia’s client for more than 40 years; Initially starting with a checking account at BIC and a savings account at CONAVI.  Second, I used the Cray X-MP as my benchmark to represent the 40-year evolution of computing, even though my first interaction with a supercomputer was the Cray Y-MP, 37 years ago, which had been online for just a few weeks at the Pittsburgh Supercomputing Center (PSC).  Third, I use the Samsung Exynos 2600 processor as current benchmark since it was recently released on the Samsung Galaxy S26 smartphones line.  Here are some comparisons to illustrate what has happened over the last 40 years:

By comparing these two extremes, we can visualize the sheer scale of the bank’s investment and why uninterrupted operations are no longer just a matter of raw computing power, but of architectural efficiency.  Therefore, I am contrasting Amdahl’s Law predictions against what the architectural solution must be.

Reportedly, on a typical peak day and hour Bancolombia currently processes up to 2,000 TPS. Hence, I am initially assuming that 2,000 Exynos 2600 processors could be put together in one or several computing clusters, making 20,000 processing elements (PE).  Further, the acquisition cost (consumer price, USD954 apiece) of the entry-level (12 GB RAM, 256 GB storage) S26 is what an organization would pay to set up a cluster, including hardware, software, and communications elements and licenses^[2].  Assuming a 60% parallelization ratio (p = 0.60), thus the combined capacity of those 2,000 such processors, working as a robust cluster, would be 12,000 transactions per second (TPS).  Then, with its planned investment over the next year, the bank will theoretically be able to meet the peak demand of TPS, with extra capacity to process unexpected peaks.  Why theoretically?  Because of Amdahl’s Law:

Amdahl’s Law Formula

Where:

Notice, though, that the maximum speedup solely depends on s.  Using the assumptions above, and plugging in, the calculation would be:

Namely, the fraction S = 1/s is what really matters; the serial fraction is the bottleneck.  Nonetheless, this speedup would be enough to process 2,000 * 2.5 = 5,000 TPS, which would provide a good margin to take care of the current peak demand of TPS.  It is necessary to state that OLTP systems often scale via throughput[^3].  Also, a conclusion from Amdahl’s Law is that to increase the speedup, processes (or rather, the algorithms running with them) must be optimized, increasing their parallelizable portion.  Let us suppose that ideally, p is increased from 60% to 99%, then this is Amdahl’s Law prediction:

This result clearly proves that even with 20,000 cores, speedup is the fraction of 1/s or 1/0.01.  Therefore, speedup coming from extra computing capacity is limited; but increasing computing capacity may improve reliability, availability and serviceability (RAS)^[4], provided new architecture is used, such as private or hybrid cloud computing and redundant data centers.  Numerically, availability refers to the number of ‘nines’, representing the percentage of time that a computer system is running.  The ideal 5-nines means 99.999% uptime.  This article from Splunk Blogs describes the concept.  Unanticipated outages, which reduce uptime, may have many causes, including events completely out of control of an organization.  The cited Splunk blog lists some of them.

Let us revisit the recent Bancolombia outage.  Initially, it was a scheduled stoppage, so it did not qualify as an outage affecting the availability index of the OLTP system.  Once human error occurred, it transitioned into an unanticipated outage.  Reportedly, the system took approximately 4 days to return to full operation.  According to the table in Splunk’s article, this would place Bancolombia OLTP system around 99% availability (2-nines).

How to increase availability ratios?  Paraphrasing realtors, “redundancy, redundancy, redundancy”.  This article by IBM shows that to increase a ‘nine’ means to reduce downtime by 10, i.e., it is an exponential progression with r = 0.1; so is the cost, with an r > 1.  According to this other IBM’s article, in a case like Bancolombia’s, a starting point to increase “nines” would be to determine the worst acceptable scenario of downtime.  In my opinion, Bancolombia needs at least 5-nines (99.999%) availability.  To get it, the bank would need redundant online servers (web, database, apps, storage, OS, etc.), redundant locations, redundant power supplies, redundant low latency communication elements and networks, all kept online so that at an outage of the production systems, the back-up systems are ready to take over.

Let us use again the example above with p = 60% and s = 40%.  Furthermore, suppose that Bancolombia purposes to install a capacity of 5,000 TPS, to have room for growth.  Since the attainable speedup is dependent on s, the necessary number of processing elements (PE) would be just around 600 per OLTP server-site; or

If 99.999% availability is the goal, then the OLTP servers must be redundant, i.e., an OLTP server-site could, itself, process the peak demand.  In other words, and by Amdahl’s Law, 20,000 cores would be equivalent to just 60 Exynos 2600 processors since adding more PEs after 600 would not increase speedup.  Moreover, adding more hardware cannot overcome poorly optimized, non-parallel software.  Hence, the total OLTP system must be at least 2 redundant online clusters, which have redundant everything.  Thus, the cost of acquisition and installation of 120 Exynos 2600, set up as two clusters of 60 Exynos 2600, or 600 PE each, would be USD114,480, which is just 0.02% of the total project budget –see footnote 1–.

Takeaway

1. The increase in performance is limited by Amdahl’s Law, which establishes that speedup is limited by the serial fraction of a process (S = 1/s).  Extra computing capacity has limited effect in speedup; it may improve RAS (reliability, availability, and serviceability), provided efficient architecture is set up.
2. Most of the project budget will go to improving RAS, not to improving performance.  Indeed, the assumptions made indicate that computing hardware would be just ~USD 114,480 (0.02% of budget).  In other words, most of the COP 1.6 trillion is an investment in RAS. The bank is not paying for faster calculations; it is paying for the system to stay alive when things go wrong; this design feature sometimes is referred to as ‘fault-tolerance’ and ‘resiliency’.
3. The partof the budget destined for Infrastructure & Operations* is by far biggest: ~USD 432,317,952 (99.98% of budget).

* Includes: architectural components like data centers, low-latency fiber, power redundancy, cooling, licenses, and specialized engineering staff, in general, RAS improvement.

The conclusion is that computing power by itself is a small fraction of the total cost of the project.  Most of the investment will go to operation costs and premises preparation and redundancies.  I wish Bancolombia success in their endeavor.

[1] While the Cray brand name existed, its machines were mostly SIMD machines, appropriately called vector processors.  Conversely, current processors, like the Exynos 2600 and Nvidia’s Grace and Vera CPUs, have multiple NPUs and GPUs, so computing clusters made up of current processors work as MIMD machines.

[2] There may be extra costs involved; for instance, use of server hardware, such as, AMD EPYC or Intel Xeon processors, ECC memories and NVMe storage arrays.

[3] Handling many independent serial tasks concurrently.  It is worth noting that RDBMSs are the heart of OLTP—See OLTP throughput definition—. Particularly at peak times, an OLTP must process many transactions against the database.  RDMSs have mechanisms to manage concurrency (e.g., many users inquiring balances or withdrawing and transferring money at the same time), such as row-level locking, thus processing those transactions in parallel.

[4] Reliability refers to a computer system operating well, as intended; availability, it operates reliably whenever it is necessary; and serviceability the computer system can undergo maintenance or updates, without affecting ‘R’ nor ‘A’.

In today’s fast-paced workplace, training can no longer afford to be an afterthought or a static checklist item.  For learning and development (L&D) professionals, learning design isn’t just about structuring content, it’s about creating experiences that drive performance, engagement and talent retention.  As organizations embrace learning management systems (LMSs) to deploy and measure training, the effectiveness of that training hinges not just on the platform, but on the design behind the learning.

Why Learning Design Is the Missing Link

While LMSs provide the infrastructure to scale training and report outcomes, they don’t guarantee impact.  Impact depends on whether the learning resonates with users, supports business goals and results in behavior change.  This is where learning design becomes a strategic differentiator.  L&D teams must align learning experiences with the flow of work, design content for various modalities and build learning pathways that guide employees from awareness to mastery.  This requires a shift from content-centric thinking to learner-centric design grounded in real-world application.

5 Learning Design Strategies That Drive Results

• Design for Data: Build in Measurable Moments

Conclusion: From Implementation to Impact

Learning design is not just a supporting role; it’s the foundation that determines whether your LMS investment drives real change.  By intentionally designing learning experiences that are goal-aligned, modular, engaging, data-driven and personalized, L&D professionals can transform training from a compliance exercise into a strategic force.

As we look beyond LMS implementation, we must prioritize design as the engine of learning effectiveness.  When thoughtfully executed, great design doesn’t just inform —it inspires, activates and delivers business impact.

Note: This article was written with the assistance of OpenAI’s ChatGPT.

When I first arrived in the US, I had a charming encounter that perfectly illustrates one of the most critical aspects of any creation: testing. My roommate, Alexandre, a fantastic French cook, and I hosted a dinner for some friends, including Fuyuko, a Japanese girl new to Mexican food. Alexandre whipped up a delicious meal. Not wanting to be left out, and despite my lack of culinary skills, I offered to make something simple for Fuyuko: guacamole.

I followed a recipe meticulously, proud to present her with a small bowl of my creation. Days later, I ran into her on campus and asked about the dip. She started teasing me, calling it a “dip of glass.” Glass? I was baffled, picturing shards in my guacamole. Then she pointed to the lawn nearby. “Ahh, grass!” We burst out laughing, but my curiosity—and embarrassment—got the better of me. Back home, I tried my concoction. Sure enough, it tasted exactly like grass.

Had I tested “my creation” before serving it, I would’ve avoided that hilarious confusion and my red face. From that day on, I stuck to being a helper in the kitchen, never again offering to cook for anyone.

From Kitchen to Code: The Universal Need for Testing

My guacamole mishap isn’t just a funny story; it’s a perfect metaphor for the need for testing in any field. Chefs taste-test their food, not just at the end but throughout the cooking process. They even get independent opinions from peers. It doesn’t matter if it’s a new recipe or their signature dish they’ve made a thousand times; they always taste.

This crucial step isn’t unique to the kitchen. Think about cars, home appliances, or mobile phones. Every one of these devices undergoes rigorous testing before it hits the market, especially new models. Car manufacturers, with all their vast experience, meticulously test new designs. Electronic giants follow the same routine.

So, why should software production be any different? It shouldn’t be!

Software Development: An Art, a Science, and a Necessity for Testing

Software development has its unique blend of art and science. It’s an art, much like cuisine, requiring creativity and intuition. But it’s also a science, governed by formal principles, technology, and standards. However, no matter how knowledgeable or experienced a software development team is, they must test their system before it’s published or released to users.

This is where Software Quality Assurance (SQA) comes in. SQA encompasses two broad activities: verification (static testing) and validation (dynamic testing). Software engineering textbooks dedicate significant sections to SQA, detailing testing categories, techniques, and strategies. Roger Pressman’s “Software Engineering: A Practitioner’s Approach,” for instance, devotes entire chapters to quality management and even introduces testing concepts early in its modeling sections. This highlights just how integral testing is to the entire software development lifecycle.

For our discussion, let’s use Pressman’s definition of quality software: “An effective software process applied in a manner that creates a useful product that provides measurable value for those who produce it and those who use it.” So far, we’ve established the need for testing, SQA, validation, and verification. But let’s get to the real question that matters to a CFO: what about the cost?

The Alarming Cost of Late Software Errors: A Financial Perspective

Finding and fixing errors is a natural part of software development. But the cost of those fixes dramatically escalates the later an error is discovered. Imagine a hypothetical software project lasting 100 weeks, with different percentages of effort for each stage: 20% for requirements, 25% for design, 20% for coding, 20% for testing, and 15% for maintenance.

Let’s assume one error is found and fixed each week. Based on Pressman’s 2009 figures (which would be significantly higher in today’s dollars due to inflation and increased complexity!), here’s a financial breakdown. To make this comparable to a capital investment, we introduce discount rates, accounting for the increasing risk (or internal rate of return – IRR) as a project progresses. Higher risk in later stages means that future costs, if unaddressed, are more impactful on your bottom line today.

The Staggering Financial Truth

Look at that last column: Total Discounted Cost. It clearly shows how quickly the cost of finding and fixing the same number of errors spirals out of control as the project progresses. An error caught in the maintenance phase (weeks 86-100) costs nearly 80 times more than an error caught during requirements gathering (weeks 1-20)!

And remember, these are 2009 figures. If we adjust for inflation and current developer salaries, these numbers would be exponentially higher today. Imagine if 30 errors were found during the maintenance stage instead of 15. Using the same formula, the cost would jump from $198,070 to over $370,000 for just that stage!

This isn’t just about avoiding embarrassment like my guacamole. It’s about protecting your organization’s budget and preventing massive financial drains.

A Message to the CFO

So, Ms./Mr. CFO: when your IT developers request funding for training in SQA techniques to catch errors early, or for hiring independent validators/verifiers, or for investing in software tools that automate and assist with these crucial tasks… please don’t see it as an expense.

Instead, compare the relatively modest investment they’re asking for against the potential Net Present Value (NPV) indicated by these error-fixing figures. I’m confident the numbers will overwhelmingly favor proactive testing, ensuring quality from the start, and ultimately saving your organization enormous sums down the line. It’s not just “playing with toys”; it’s a strategic investment in financial health.

In future posts, we’ll dive deeper, moving from software testing to the broader landscape of software development, IT support, and how it all aligns with business strategy. Stay tuned!

Note: This article was proofread and improved with assistance of Google’s Gemini.

We understand that your team is your main asset. High-performance organizations implement high-performance learning and development strategies to drive communication and talent development to improve performance, reduce risks, and increase revenue. From on-boarding through end-of-service, your aim is to develop and nurture a high-performance team equipped with the knowledge and resources to implement and lead strategic efforts.

Optimizing technology drives business. Leveraging technology to improve processes, gather business intelligence, inform process improvements, and drive new business initiatives is today’s reality. Many organization miss the mark when it comes to optimizing business efficiency through technology. Granted, newer technology initiatives may be costly to an organization. Nonetheless, your organization needs to understand the full potential and limitations of your current systems and how these may impact strategic business goals and objectives. With knowledge at hand, you will be able to make informed decisions regarding technology as a business driver.

Develop a business model that is as dynamic as your organization. A business model describes how your organizations creates, delivers, and adds values to customers, and provides a high-level view of how your organization operates. Within the business model you establish your value proposition, develop your financial structure, communicate business goals and objectives, and outline the best practice strategies you use to meet desired business outcomes.

Acureach is your strategic partner. Acureach supports business owners, entrepreneurs, and organizations who aim to achieve strategic alignment in business. Acureach consultants are learning leaders who help you and your team get the most value-add from technology, business best practices, and learning and development strategies to ensure that your processes are aligned to your core business needs. You have the vision on where you want to take your business. We bring our business acumen and expertise to help you get there. Your success is our success. We look forward to working with you. Contact us today!