In this special guest feature, Kristin Hansen, Chief Marketing Officer at Bright Computing writes that HPC doesn’t need to be held back by commonly held misconceptions.
As high performance computing continues to work its way into the enterprise tech mainstream, a variety of myths and misconceptions conspire to slow the technology’s growth. Companies looking to take advantage of HPC to grow their business need to tell the difference between the real challenges facing high performance computing adoption and misinterpretations with little basis in reality.
So is there any truth to some of the more popular HPC myths? With the industry analyst group IDC predicting the HPC market to grow to near $15 billion over the next three years, it’s important to have an accurate picture of the HPC landscape.
Is the cost of ultrascale systems prohibitive?
Many CIOs feel the high cost of HPC systems makes purchasing them an unrealistic expenditure for their company. While it’s true some systems cost upwards of tens of millions of dollars, a company can acquire a cluster for less than one million dollars – the price of a typical large scale enterprise computing system.
The data analysis driving today’s competitive business world thrives on the power provided by ultrascale computing. It is important for a firm to perform the necessary due diligence to determine the true costs of an appropriate HPC system rather than assuming price puts this setup out of reach.
No one else is adopting HPC, right?
Some enterprises feel high performance computing is still in its early, experimental stages and that practical applications are still years away. The presumption is that if government investment is still vital to the HPC market, it shows the private sector is not yet interested in spending meaningful dollars on the technology. However, what works for one company may not work for another. It is important to note that when it comes to strategies, one size does not fit all.
As noted earlier, a study by IDC reported that the HPC business sector continues to grow at a robust pace. They predict an annual growth rate of 7.4 percent through 2018. Applications combining data analysis with high performance computing (HPDA) are a prime factor driving this expansion. IDC predicts the market for HPDA servers and storage to grow to $2 billion next year.
Important high tech firms like PayPal and Google are currently leveraging HPC. PayPal manages its IT infrastructure with HPC technology, while Google is currently designing high performance computing architectures for their next-generation systems. There’s no denying that HPC is currently relevant to the modern enterprise.
Aren’t HPC systems too complex to manage?
Another misconception about high performance computing concerns the relative complexity of the technology compared to mainstream IT. Some enterprises feel the additional investment spent in training IT staff to manage an HPC system makes the overall cost too prohibitive. They think the technology still belongs in the supercomputer lab at a university.
Many HPC systems use standard operating systems, most notably Linux, which increases the familiarity for most network administrators. HPC manufacturers have also taken steps in recent years to improve the usability of their systems. In fact, more business users now directly access HPC servers and clusters to run high-end data analysis applications.
As the high performance computing market continues to mature while migrating from the academic lab to the corporate data center, enterprises can’t afford to ignore this vital technology. An organization’s competitive advantage might very well depend on dispelling these few HPC myths.