JUNE 30, 2010

CLOUD COMPUTING JOURNAL

A new frontier for life sciences and beyond

There are hundreds of life science labs in the U.S. using next-generation sequencing, bioinformatics, proteomics, and molecular modeling to identify the genes behind, and potential drug targets to cure, many diseases including diabetes, cancer and Alzheimer's disease.

With increasing data coming off of modern scientific instruments, the demand for compute power to analyze the data is increasing dramatically. Currently, life science researchers in bioinformatics, next-generation sequencing, and molecular modeling need to spend tens to hundreds of thousands of dollars to buy server clusters to run their scientific calculations.
High performance computing (HPC) has come a long way for life sciences. Twenty years ago, expensive parallel supercomputers were required to render proteins in three dimensions and run software that helped researchers understand their shapes. Now 3D rendering can be done on graphics cards in workstations, laptops and even phones.

It is important to note that there are two types of HPC. There's the sprinter type, where users try to run a highly parallel application, and then there's the marathon runner type of HPC, in which applications are pleasantly parallel. For sprinter applications, latency is of key importance and performance must be optimized at every level to get results. Currently these applications are best run on a single multi-core server in the cloud;
however, infrastructure from various providers may make this use case able to run on many servers. For the marathon applications, also called high throughput computing, many commodity servers can run jobs faster by taking advantage of the parallel nature of the work.

In either of these applications, compute clusters using many commodity servers have replaced expensive parallel supercomputers, but the data and problems being solved have grown to demand increased compute capacity. This leaves companies with large capital investments in fixed-size clusters that have all the traditional challenges of maximizing utilization, minimizing operational costs and shortening time-to-result for users.

» Read full article