DATAKAMER

Every data centre has two ledgers. One is visible: the racks and cables, the workloads and experiments, the discoveries that make headlines. The other is hidden: the power bills, the migration cycles, the egress fees, the hours staff spend moving petabytes from one generation of hardware to the next. That hidden ledger doesn’t appear in glossy presentations, but it quietly dictates what research institutions can and cannot do.

At Datakamer, the question of how to manage this ledger became a pressing concern. An IT leader from a world-renowned research centre described the constant challenge of balancing immediate scientific needs with infrastructure costs that are rising faster than budgets can adapt. Every choice, from media to cooling, has consequences, and often the most challenging part is not just enabling the science but sustaining it.

Vendors on the panel offered different ways to rebalance the books. SpectraLogic’s Matt Ninesling reminded the audience that tape, far from being obsolete, continues to survive for one reason: arithmetic. For prestigious Ivy League universities and world-leading cancer research centres, new generations of LTO continue to deliver the lowest cost per terabyte over the long haul. Automation has stripped away much of the friction. Tape may not be glamorous, but it is undeniably efficient.

Wasabi’s Drew Schlussel contrasted this with the cloud. For researchers, traditional pricing models turn storage into a guessing game; every retrieval and transfer becomes a line item, unpredictable and corrosive to budgets. Wasabi’s flat-rate, no-egress approach is an attempt to fix that. Predictability itself becomes a kind of innovation: not about the lowest unit price, but about ensuring that next month’s bill won’t derail the grant.

Looking further ahead, Steffen Hellmold of Cerabyte focused on longevity. Every migration incurs hidden costs: bandwidth consumed, checksums recomputed, and staff diverted from science to infrastructure. By extending media lifespans from years to centuries, ceramic-based storage aims to flatten that curve, reducing the cadence of refresh cycles that quietly drain money and attention. Longevity, Hellmold argued, is not just about permanence — it is about bending the economics of data into something sustainable.

And then there is compute itself. IQM’s John Bernart described how quantum systems, integrated with high-performance clusters, are beginning to shrink workloads that once consumed thousands of nodes for months. The gains are not only in speed but in power consumption. For institutions staring at multimillion-dollar energy bills, efficiency at the computational layer could be as transformative as any advance in storage.

What tied these perspectives together was not competition, but complementarity. Tape, predictable cloud models, long-lived media, and quantum efficiency all represent different strategies for balancing the same hidden ledger. None erases cost; each decides where to carry it — in dollars per terabyte, in predictability of billing, in reduced migration cycles, or in lower energy footprints.

For researchers, this is not an abstract debate. It is the boundary between what projects can proceed and which ones must be deferred. The hidden ledger of the data centre is, ultimately, a ledger of discovery itself. Institutions that learn to manage it creatively will be the ones able to push science further.

This essay draws on insights shared during the panel discussion “Improving the Total Cost of Ownership (TCO) of a Datacenter” at the inaugural Datakamer. Panellists included an IT leader from a world-renowned research centre, Matt Ninesling (SpectraLogic), Drew Schlussel (Wasabi Technologies), Steffen Hellmold (Cerabyte), and John Bernart (IQM).