March 12, 2007
Penguin Computing
Pauline Nist
Senior Vice President, Product Development & Management,
How has your company adapted and responded to the changing economic climate in the past five years when so many others companies did not?
From the time Donald Becker, now our CTO, and Thomas Sterling invented the Beowulf Linux cluster architecture and started this new category, we have been growing and changing. Penguin Computing has been able to grow partly because the value proposition of Linux in general has become so much more apparent to technology users. And as one of the older pure Linux vendors out there, we have used this time to develop a real hands-on expertise in Linux clustering and servers. In addition, as more and more organizations have realized the value they can get from increasingly powerful processors, high performance computing (HPC) clusters have become a viable alternative to SMP machines. We are fortunate to be at that great intersection of market readiness and the right qualifications.
What is your vision for the future of the life sciences market over the next several years?
It is clear that the life sciences market is destined to be one of the largest users of HPC. The use of this computing technology promises revolutionary new therapies for diseases currently considered untreatable. As life sciences demands more and more computing power, Linux clustering can deliver it at lower cost.
What products and services does your company provide and what special capabilities do they offer the life sciences market?
The complexity of Linux clustering is a barrier to entry for many smaller organizations targeting key niche projects. Penguin Computing’s Scyld ClusterWare and our suite of pre-tested life sciences applications offers the leading solution to simplify the provisioning and administration of life sciences clusters, lifting this burden from researchers and enabling more rapid scientific progress without IT expertise. Penguin provides the support needed to make these complex configurations approachable for the average scientific customer.
At the same time, Scyld ClusterWare is an excellent solution for larger organizations that need to reduce time to results in their discovery process. It’s a fast deployment and an easy-to-manage tool that is ideal when there are a lot us non-computer scientists who need data analysis. It can scale to hundreds of nodes as needs grow.
To add even more value to these organizations, we offer Scyld TaskMaster a best-of-breed resource management and scheduling system. This software provides all of the sophisticated features a complex environment requires (such as fairshare, deadline scheduling, etc.) as well as a powerful management interface that permits changes to be tested in "simulation mode". This permits a cluster admin to know the effect of changes before making them to the running cluster. TaskMaster also provides a portal interface, permitting customers to submit jobs from their workstation, without logging on to the cluster.
Partnerships are an effective way to track life science advances and ensure that your company delivers timely products and services. Which life sciences companies or organizations have you partnered with or invested in and why?
We have worked closely with software providers like Schrödinger, but many of the applications that our customers utilize in this market are largely open source. As sife sciences matures we will not only partner as necessary, but also track emerging companies and open source developments that appear in this rapidly evolving market segment.
What are your most exciting products and initiatives in development, and how will they improve life science research?
Our new low-cost cluster compute node offers excellent price/performance, permitting our customers to build more clusters for less cost. Since the Bio-IT market is characterized by both large and small companies, the ability to satisfy ever increasing computing needs with extremely cost effective hardware, is key to insuring discovery can continue at all levels. We will continue to invest in software that provides ever more sophisticated management and virtualization tools for administering large Linux clusters.
Where do you see your company in five years?
Five years is an eternity in hi-tech. It is clear that hardware price performance will continue to improve through parallelization, very large scale integration (VLSI) and architectural improvements. As Linux matures, customers will expect more and better software tools, and automated management of utility-style, real and virtual computing resources.