HOW CLOUD HPC IS REDUCING TIME TO INSIGHTS IN PRECLINICAL RESEARCH
share:
P u b l i s h e d by
by F i e r c e C u s to m
P u b l i s h i n g
How Cloud
HPC is reducing
time to insights
in preclinical
research
P
reclinical research is an unpredictable, fast-
moving field. New insights quickly open up
or shut down avenues of research, forcing
teams to rethink their experiments. Computing sys-
tems have historically been less adaptable, but that
is changing with high performance computing (HPC)
in the cloud. Using on-demand, pay-as-you-go cloud
HPC, preclinical teams are achieving research goals
faster than ever.
HPC refers to the aggregation of computing pow-
er. By having multiple processors working in parallel,
HPC delivers more power than a desktop computer.
Access to such power has underpinned advances in
genomics, molecular modeling, computational chem-
istry and other areas of life science research. HPC
enables researchers to manage the 50 terabytes of
data Illumina's HiSeq X Ten can generate each week,
and equips Big Pharma firms to virtually screen mil-
lions of compounds.
The downside is HPC clusters are expensive to
acquire, configure and maintain. Procuring technol-
ogy is a long, costly and complex process. It takes
months and many millions of dollars. Configuring pro-
cessors to effectively run researchers' computations
takes more time and expertise. Once installed, tech-
nical problems and computing-capacity bottlenecks
lead to downtime and distractions, slowing research.
These characteristics undermine on-premise HPC in
preclinical research.
"On-premise HPC isn't as agile as our researchers
want to be. Because research collaborations can
start up spontaneously, the result can be demand for
IT that wasn't forecast," said Lance Smith, associate
director of IT at Celgene.
The IT sector responded to these challenges with
cloud HPC. This paper discusses the benefits of
deploying preclinical workloads in the cloud.
Cutting time to science
Cloud HPC shifts the burden of acquiring, config-
uring and maintaining computing equipment from
preclinical researchers to service providers. Cloud
providers such as Amazon Web Services (AWS)
employ compilers, schedulers, distributed network
file systems and other tools familiar to users of on-
premise HPC clusters.
S P O N S O R E D BY