Is Technical Debt Making Your Digital Infrastructure Brittle?

Technical debt is becoming a major digital party pooper. While every company understands its definition, it remains a significant part of IT balance sheets—around 40%, according to McKinsey.

Companies must become honest about how technical debt is formed in order to tame it. It also highlights the challenge of maintaining a long-term view of digital strategy vs. short-term fixes that band-aids a problem.

“Technical debt often arises as a result of organizations making temporary fixes that become permanent, not updating solutions that become outdated, favoring fast technology delivery over long-term benefits, or implementing one-off solutions to meet business priorities,” explains Tejaswini Tilak, vice president for marketing in APAC at Digital Realty.

For example, technical debt occurs when an organization that applies fixes to maintain legacy systems or reduce OPEX on older architectures no longer sees expected beneficial outcomes.

The problem with technical debt is that it creates issues beyond just TCO or ROI. It can reduce operational efficiency, bring down employee morale, and even adversely impact the environment.

Tilak also sees technical debt hampering business transformation efforts and holding back the company's competitiveness in the long run. With new technologies, such as generative AI, evolving quickly, technical debt can be a bane for those ensuring their digital cores keep up with the infrastructure demands of these technologies.

Rooting out technical debt

Digital Realty is at the frontlines of the war against technical debt. It is constantly looking to use its vast network of colocation facilities to help companies tame their technical debt.

"Generally, adopting a colocation hybrid IT strategy that allows for data interaction at the edge, with access to a dense ecosystem of partners, customers and providers, offers companies the best path to enhance network flexibility and potential, and to solve for scale, speed to market, security and costs,” says Tilak.

AI offers a good example. It has a voracious appetite for computing, network, and storage resources. The growth in requirements is growing unabated as more companies jump on the AI bandwagon for various reasons. More importantly, they are not doing it alone but building or becoming part of AI ecosystems.

Tilak sees this as where Digital Realty comes in with its state-of-the-art tools "to build optimized AI architectures and the ability to digitally engineer deployments in virtual environments."

“Our data center designs contain modularity and large capacity blocks to support legacy and high-density AI deployments, all interconnected with ServiceFabric™, a global, purpose-built network fabric,” she explains.

This modular design philosophy allowed Digital Realty, for example, to deploy liquid cooling in its data centers to ease power usage and heat emission. The design approach will enable them to deploy the most efficient technology that fits a specific customer's need, mixing and matching across even a single floor on a global scale while meeting rising sustainability demands.

“For example, our portfolio in SG is fully BCA Greenmark Platinum certified and was recently recognized by the Singapore Environment Council for our water-saving innovation and achievement at our SIN10 facilities. We also recently completed our first green building certificate in South Korea in 2022, adding to our cumulative 12+ million square feet of certifications throughout the portfolio. We also achieved NABERS ratings of 4 stars or higher for three Australian data centers," says Tilak.

Modular innovation allows the company to veer toward renewable energy supplies and cooling technologies. It reached 1GW of new renewals under contract in 2022, a 9.8% YOY increase. 126 of its data centers today are matched with 100% renewable electricity, while its Australia multitenant sites will transition to renewable energy sources and will be matched with 100% renewable energy sources by January 2026.

It's one reason why companies like SURFsara, and Graphcore are choosing the company's data center platform, PlatformDIGITAL®, for groundbreaking AI training sets and applications, says Tilak.

The data dilemma

Another key reason many companies choose Digital Realty is its approach to data. That's because AI success often hinges on data.

The company heavily promoted the concept of data gravity way before AI exploded into the public consciousness and business leaders' top demands.

Data gravity's concept is intuitive and practical. As applications continue to use larger data sets, moving these data sets around becomes more difficult. This rising inertia becomes compounded as more applications and data are added to a specific location. It creates a major problem for AI applications, like smart manufacturing or autonomous vehicles, which require the data to be close to where it is processing.

Tilak points to Digital Realty’s efforts to host many of the world’s largest data lakes, including its Santa Clara location.

“(Santa Clara) is fast becoming the hub for autonomous driving, with many autonomous vehicle software companies deploying their autonomous cars for pilots so that the data created within their autonomous vehicles’ computer systems can access these large data lakes to enable the autonomous vehicles to continually learn and improve. As more complex models emerge, they will rapidly ingest more and more data, which will continue to expand these data lakes in our facilities over time,” says Tilak.

She notes that Digital Realty's focus has always been on data and managing data gravity challenges "to help ensure our customers can efficiently store, analyze, and extract value from their data by providing the meeting place where companies, technologies and data come together." "We do this by delivering the right infrastructure and expertise in the right locations our customers require around the world," she adds.

Fears of insecurity

Data and infrastructure efficiency are not the only contributors to technical debt. Security and compliance fears can multiply it by several folds.

For example, financial companies often need their data to reside in specific markets and be stored in resilient architectures to meet data sovereignty regulations. "As a result, changes to the data are deferred to maintain compliance," Tilak observes.

Shifting to a data-centric approach that looks at the crucial placement of data infrastructure in the right places can address technical debt.

“To create an architecture that puts the data at its core, and where public, private, and shared datasets are connected, integrated, and secured, we need to not only create a data-first environment that promotes integration, security, and collaboration but also be increasingly aware of the geography around digital hubs and population centers," says Tilak.

AI adds additional complexity. There is a greater need for data centers that can meet specific requirements, such as high-speed processing capabilities and extensive storage capacity. The data centers must also be scalable, have low-latency networking, robust security measures, high availability and reliability, and energy efficiency to accommodate the demands of AI.

“Where only last year a data center operator may have been able to plan on an average of 10-kilowatt power draw per rack of customer equipment, the need for increasingly large blocks of 25, 50 or even 100-kilowatt racks at different places across that same data center facility is here and will only continue to grow,” says Tilak.

Digital Realty recently announced the availability of high-density colocation services across PlatformDigital, its global data center platform. It supports workloads of up to 70 kW per rack, utilizes innovative air-assisted liquid cooling (AALC) technologies, and is configurable to flexibly and sustainably scale according to customers’ consumption models.

Digital Realty has around 400 MW of data center capacity coming online globally over the next 18 months. "To our knowledge, this is unmatched across the industry and will position us well for winning new AI deals. We also announced our first NVIDIA DGX H100-certified status at our latest data center facility, KIX13 Osaka, and are working on getting more facilities worldwide similarly certified," she explains.

Igniting the technical debt debate

Digital Realty thinks technical debt cannot be solved by a single company acting alone. It requires an ecosystem of providers and customers with the grit to tackle it upfront. Regulations that continually impact technical debt and data gravity are also evolving, and providers need to be mindful of this. 

“From data localization based on enterprise concentration in the Americas to data privacy and regulations in EMEA and data sovereignty and residency rules in the Asia Pacific, data laws and regulations have a big impact on the effects of data gravity and how enterprises adjust,” says Tilak.

To address technical debt holistically, it first needs a conversation. It is what Tilak is looking to ignite at the upcoming sixth Chief Digital and Data Officer Hong Kong Summit. Besides talking about the latest findings from the company’s Data Gravity Index™ 2.0 study, which correlates the proliferation of data globally with Gross Domestic Product (GDP), she is looking to get to the root causes of technical debt.

It's a conversation that all data-driven companies should have now rather than later.

Winston Thomas is the editor-in-chief of CDOTrends and DigitalWorkforceTrends. He’s a singularity believer, a blockchain enthusiast, and believes we already live in a metaverse. You can reach him at [email protected].

Image credit: iStockphoto/tiero