6 days Ago
Charles Yeomans is the chairman and founder of Atombeam . getty Change is a constant in IT—change driven by new innovations, new use cases, new problems and new solutions. Over time, advancements from mainframe computers to the evolution of PCs to the rise of virtualization and the supremacy of the cloud—and even the move from desktops to mobile devices—fundamentally altered the very notion of enterprise IT.
Most CIOs are accustomed to this ever-changing reality, with many having cut their teeth when businesses looked to IT to oversee internal systems and a corporate, on-premises data center. With the hub-and-spoke model, responsibilities increased. Then, the cloud changed everything.
The cloud’s secret weapon transformed enterprise IT. Its magical ability to automate previously weighty tasks—from adding compute power to spinning up additional storage capacity—was a game-changer. Everything didn’t get easy, though.
The distributed organizations the cloud helped make possible called for inherently more complex IT infrastructure. And it didn’t take long for organizations to realize they still needed to keep some data under lock and key within their own walls. Organizations that previously looked to get out of the "data center business" altogether responded with hybrid approaches and single-tenet private clouds, even as many oversaw the building of new in-house data centers and the provisioning of the mission-critical systems within them.
But even as many IT leaders again looked inward, the edge of the network continued to expand. This brings us to today, when enterprise networks are more complex than ever before, and IT teams—even those armed with greater visibility and control—have even more to do than ever before. And that is before even taking into account two truly transformative computing trends: AI and the burgeoning Internet of Things.
With these technologies, we have entered a new era, one that will impact enterprise IT like never before. Enterprise IT teams today must oversee a highly distributed computing environment marked by unprecedented complexity. Not only is the edge of the network no longer bound by the constraints of a wired network, but the Internet of Everything is growing as satellite and mobile networks connect a dizzying array of devices, from autonomous vehicles and drones to appliances and sensors of every kind.
Machine-generated data is increasing exponentially, with IoT Analytics estimating the existence of 18.8 billion connected IoT devices globally. And that is not even factoring in what is arguably the most transformative computing trend of our time: the use of generative AI, which has upended entire industries while simultaneously forcing everyone to consider its disruptive potential in less than three years of public use.
Both trends pose immediate challenges to enterprise IT teams, who must account for, plan and address their significant impact on the most fundamental computing tasks, including: Heavy, data-intensive AI workloads are dramatically straining networks and using an unprecedented amount of power —a fact prompting hyperscalers to look to nuclear reactors to power their data centers. Notably, power consumption is particularly great in use cases that require low latency and near-real-time communications. AI workloads are also overwhelming the pipes, satellite and cellular networks that enterprise IT relies on.
In contrast, the machine-generated data shared from sensors and other small devices that comprise the IoT is typically lightweight but often burdens infrastructure with the continual ping of shared information. In both cases, networks—including the data centers within them, the pipes that feed them and the connections between machines—are increasingly bogged down. The data deluge that made companies like EMC some of the most profitable in history ultimately drove many enterprises to the cloud, in part for its elasticity.
Massive AI datasets upend that dynamic even in hyperscale environments, where data center operators are running out of capacity in their facilities and across their clouds. At the edge of the IoT, many endpoints— 98% of them , according to Palo Alto Networks—are unprotected because devices like low-power sensors lack the computational power and memory encryption algorithms required, leaving bad actors with an open door. In light of these factors, the responsibilities that have for decades defined IT’s estate and its core responsibilities—to provide sufficient compute power, networking speed and effectiveness, storage capacity and data protection—are increasingly untenable using traditional approaches and means.
For years, we have turned to more powerful hardware to provide and manage the IT infrastructure needed to handle the ever-growing amount of data we create. Those efforts will, by necessity, continue as we work to effectively address the demands associated with AI workloads and the dramatic growth of machine-generated data. In response, leading innovators are creating faster chips and more powerful processors , even as quantum computing emerges.
Simultaneously, a satellite network and data center building boom is underway, even as the aforementioned use of nuclear reactors reveals the very real concerns around how data centers will be powered. Time, however, may be the greatest challenge: In one projection for data growth, McKinsey sees demand potentially reaching 298 gigawatts by 2030 , a reality it notes will require “twice the data center capacity created since 2000 to be achieved in one quarter of the time.” Given this, one can effectively argue that no one approach or response will effectively address the very real reality that, with the growth of AI and machine-generated data, workloads have outpaced the infrastructure resources they require.
We are at an inflection point that requires us to accept that the need for faster data processing and storage is not going away. For that reason, it is imperative that take a step back and think not only of the additional infrastructure capacity we can create, but also what we can do differently—from bringing information closer to the edge to changing the very notion of how data is configured and how it can be moved, stored and protected. Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives.
Do I qualify?.
