"Grant-funded researchers can buy cloud resources for specific projects, but they typically can't guarantee repeat purchases or influence how their institutions spend on computing infrastructure more broadly. When vendors offer credits to support research groups, they're investing in what they hope will become profitable long-term partnerships. These expectations, however, often go unmet - research groups simply don't wield enough influence over institutional procurement to deliver the sustained business vendors anticipate."
"Cloud vendors' commercial models poorly serve scientists, forcing them to struggle for value amid tightening budgets, according to research. Modern science - from bioinformatics to astrophysics - depends heavily on sophisticated computer modeling, yet cloud providers' business-focused models clash with how scientific projects consume computing resources, argue Vanessa Sochat and Daniel Milroy, both post-doctoral researchers in computing at Lawrence Livermore National Laboratory."
Commercial cloud pricing favors persistent, business-style consumption with long-term discounts and preemptible instances, but scientific workloads are often short, sporadic, and require specialized, high-precision hardware. Preemptible instances create reliability risks for tightly coupled MPI-based simulations because a single instance failure can collapse an entire run. Businesses can continuously reinvest cloud spending to secure favorable terms, while many scientific projects are finite and lack institutional procurement leverage to negotiate similar discounts. Grant-funded teams can purchase resources for specific projects but cannot guarantee repeat purchases, so vendor credit incentives frequently fail to translate into sustained, mutually profitable partnerships.
Read at Theregister
Unable to calculate read time
Collection
[
|
...
]