This week in New Orleans we were able to secure time with some leaders in the HPC cloud space, both in terms of vendors and research institutions, to get their take on how a movement like cloud is shaping the future of HPC and IT in general.
One such luminary we were able to find a few moments to speak with was Songnian Zhou, co-founder and CEO of Platform Computing. If you have about ten minutes to spare, take a listen to what he has to say—our conversation is not about Platform but rather about maturation stages for IT in general, taking into account developments in the grid to cloud movement and beyond.
In Zhou’s view after 18 years in the industry, HPC as a field can now contribute more to the world economy due to the enhanced access that cloud computing provides. While he sees that grids and cloud computing as variations on the same idea, coming as they did from the first moment the concept of clusters was strung together and distributed computing—he sees the missing piece that separates grids and clouds as the management layer; the interface for the user.
Certainly, since this is Platform Computing Zhou has an investment in touting the management layer as the critical missing component that will allow for a new era of IT via cloud computing—but he does readily admit that this is still an area that requires a significant amount of innovation still.
Many can argue that the performance and data movement issues might outweigh management of distributed or remote resources, but between these three topics there is a common theme that suggests the triad of barriers for true HPC (focus on the “P”) are not insurmountable. Further innovation is required since we are still in the infancy of clouds for HPC.
Depending on whom you ask, there are a few different answers to the question of what separates cloud computing from what came before it. Some in HPC believe that the only difference lies in some slick marketing, while others, like Zhou, feel that the key element lies in ease of management of complex resources via an automated, abstracted layer of software that brings the vision of grid computing into a new era of development and usability.
As Songnian Zhou states in this interview, cloud are “the biggest invention in computing models in the last 30 years” and that even after you wash away the hype cloud is an architecture, one that delivers applications as services, where users consume them as services because “services are the essence of cloud—and every industry that becomes mature becomes a service.”
In this way, Zhou sees that we are still in the maturation stages as the integration of these disparate services come together in innovative ways. In this sense he feels “this is now the renaissance of IT.”
By the way, earlier this summer we featured an in-depth interview with Mr. Zhou. This quickly became one of the more popular articles on the site, in part because Zhou is well-known in high-performance computing circles and also because the company’s history is quite interesting. While coming straight from academia into company status is nothing original or new, Platform has managed to remain a private company since the beginning, which we discuss in the transcribed interview here, which you can check out if you want a little more background. Again, it’s free from product talk and provides a great mini-background on some key movements in the history of distributed computing.