3 questions you need to ask yourself about your HPC Strategy in challenging times

Posted: 4 May 2020

Challenging market conditions drive compute-intensive businesses to re-evaluate HPC strategy


The good news: smart colocation strategies increase resilience, reduce cost and ramp up output

In this time of uncertain economic climates and challenging conditions globally, businesses often find themselves reassessing their entire strategy from the ground up.

Compute-intensive organizations face a particular set of challenges. Their compute capacity takes on new significance in the light of a work-from-home workforce, an increasingly shaky global economy and a portfolio of workloads that must accomplish more for less, especially in the light of tightening budgets.

These three considerations each drive toward a single conclusion: The old ways of running your IT and HPC may not remain the smartest and most cost-effective going forward. It may be time for a from-the-ground-up reevaluation of your organization’s compute as well.

Have you located your compute-intensive infrastructure in the most efficient and sustainable locations?

Some 80% of Western European data center colocation has in recent years built up around four urban centers: Frankfurt, London, Amsterdam and Paris. These “FLAP” cities are essentially the face of European colocated computing. And because the FLAP cities are also key connectivity hubs for digital traffic, it makes perfect sense to host network-intensive workloads and content there.

However, compute-intensive workloads are another matter. HPC and big data applications, engineering and simulation, forecasting and modeling: These centerpieces of a compute-intensive business demand an infrastructure that’s designed directly for their needs.

Which means reconsidering the wisdom of compute-intensive colocation in expensive urban data centers, with electricity prices at a premium and grid reliability less-than-ideal and with cooling options limited by temperate latitude and less-than-sustainable resources available.

A crucial factor in determining where to site one’s compute-intensive workloads is the reliability and stability of the energy source itself. Good environmental stewardship is also good business in this case. No wonder, then, that Iceland, where 100% of electricity comes from renewable sources (primarily hydropower and geothermal), has been consistently rated among the safest and most reliable locations in the world for data centers.

Are you able to increase and decrease your compute and remain resilient to new incidents and challenges?

A large portion of compute-intensive loads today is still done in in-house data centers or in urban colocation facilities. Which locks that capacity in and keeps it restricted to five-year budgets and hardware. Yet with even optimistic OECD forecasts seeing global and European GDP shrinking at least into 2021 [1], the need for financial flexibility and cost-efficiency is all the more pronounced.

No doubt some capacity will remain in-house, while some might be in nearby colocated facilities or with public or private cloud services. However, a hybridized data center strategy — one that sources HPC and compute-intensive cycles to data centers optimized for them — taps into a synergy of competing efficiencies. Using multiple sources for the entire range of IT capacity adds layers of redundancy that no single source could achieve.

Are you able to take advantage of distributed resources with a modern approach to workloads, forecasting and your IT requirements?

Differentiating your HPC and other compute-intensive workloads from the rest of your compute cycles is the first step on a journey that can conclude with a substantial increase in HPC capacity for any given budget outlay. Colocating HPC and compute-intensive cycles in Nordic countries and Iceland, you’ll find, opens the gates to the very things FLAP data centers today struggle to deliver — low-cost power consumption, fast access to multiple Megawatts of power, superior sustainability, extremely abundant and inexpensive cooling and facilities that do not require a fortune to build, run and operate.

Today, as businesses are increasingly investing their time and resources in the most clever and efficient ways possible, it only makes sense to move HPC and other compute-intensive workloads to the IT setting where they are most cost-efficient — in locations where you get more capacity, more flexibility and world-class sustainability. Contact Advania Data Centers today to find out how you can deliver more compute for less while increasing your productivity and reducing your downtime.

[1] OECD Economic Outlook, Interim Report (March 2020) https://doi.org/10.1787/796989...

Want to discuss your IT strategy? Fill the form below. Right arrow