Designing better quality products, shorten time to market, reducing product failure early in design, and increasing return on investment can be achieved by making high-performance computing (HPC) simulations part of the early phase of the product life cycle. Today engineers and scientists have access to three major computing tools for their simulations: workstations on their desk, in-house computing servers, and HPC cloud resources. All three options come with benefits and challenges. We will examine these benefits and challenges for manufacturers and scientists using HPC in the Cloud which complement in-house workstations and servers. We will also present UberCloud’s approach to take HPC to the masses – a highly sophisticated claim which sounded completely unrealistic when we started with UberCloud three years ago. But look and see how close we got in the meantime; today, it’s indeed better, faster, cheaper, with cloud computing.
Your Choice: Workstation – Server – Clouds
Why then are many engineers running simulations still just on their workstations, although – according to the US Council of Competitiveness  – 57% are regularly dissatisfied with performance limitations of their workstations? The main reason is that the alternatives – in-house HPC servers and public HPC clouds – are still coming with several challenges, at least until recently.
Alternative one, buying a powerful HPC server, comes with high “Total Cost of Ownership” (TCO) demonstrated by IDC in 2007 : in addition to server cost, the expenses for staffing, training, software, downtime, and maintenance easily sum up over three years to the ten-fold of the original server cost. And, in addition, there are often long and painful internal procurement and approval processes. Finally, for many, “Return on Investment” (ROI) is not easy to describe, although it is expected to be huge according to a recent IDC study on ROI in HPC .
HPC Cloud – What’s In It For Me
The second alternative for SMBs to experience the benefits from HPC is recently offered by cloud computing.
For engineers and scientists: HPC in the Cloud enables engineers and scientists to continue using their workstation for daily design and development, and to submit larger, complex, and time-consuming jobs or many jobs in parallel to the cloud. Major benefits of the HPC Cloud solution are on-demand access to ‘infinite’ resources especially for SMBs which cannot afford to buy larger more expensive HPC serversPC server, pay per use or short-term subscription, reduced capital expenditure, greater business agility, and dynamically scaling resources up and down as needed.
For independent software vendors (ISVs): For ISVs the Cloud is a new delivery platform, and it is an entirely new route to market, and possibly a channel to new customers, that will provide its software more flexibility while also requiring a change in strategies related to pricing, reseller relationships, licensing and revenue streams. The Cloud enables ISVs to instantly respond to their customers’ needs for more and short-term software licenses when peak loads occur on the customers systems and business needs to react quickly. New software container technology enables ISVs to install, evaluate, and test their software once and then it runs anywhere, without complex re-installation, on any resource. As already demonstrated in the enterprise IT world the economies of scale that come with the cloud combined with the innovation and agility of fully SaaS (Software as a Service) ISVs has led to a world of primarily SaaS-based applications which now influences the technical software community too.
For cloud resource providers: That’s a tough one. In case you are not Amazon, Google, or Microsoft you might face the danger of getting squeezed between the big guys and their price battles. The only way to escape is to add value on top of your cloud infrastructure resources. A common way today is to closely work with ISVs to bundling software and hardware and make it accessible ‘all in one’, together with an excellent customer service and open information policy. However, a drawback for the customer can then occur because it is often a challenge to find out pricing for the bundle, a situation which might scare away many customers.
For consultants and freelancers: This group is very lucky! In a sudden consultants have a completely new business opportunity by adding software and computing resources in the cloud to their existing consulting services without paying upfront, and at the end of the project they just charge it to their customer’s account.
The HPC Cloud Market
Global spending on Infrastructure as a Service (IaaS) cloud computing is expected to reach almost US$16.5 billion in 2015, an increase of 32.8 percent from 2014, with a compound annual growth rate (CAGR) from 2014 to 2019 forecast at 29.1 percent, according to Gartner’s latest forecast . Other forecasts  look at Software as a Service (SaaS) and find that the SaaS market is expected to grow steadily into the future, estimating that by 2015, SaaS revenue will have climbed as high as 27.94 billion U.S. dollars up from 13.46 billion in 2011.
However, the HPC in the Cloud market is still far behind the enterprise SaaS market, for all the good reasons we are mentioning in this article. The only market data currently available for HPC Cloud are from an IDC research involving 905 HPC sites of various sizes in government, academia and industry , where 23.5% of these sites used cloud computing of some type in 2013, up substantially from 13.8% in 2011. Of the HPC sites that use cloud computing, half use private clouds and the other half uses public clouds – our guess: in an early stage. Early adopters include government, manufacturers, bio-life sciences, oil and gas, digital content creation, financial services and other high-end analytics, online consumer services, and health care firms.
One important market segment and potential cloud opportunity is not considered in the IDC study, the digital manufacturing market of engineers and scientists using mainly workstations. According to a study of the Council of Competitiveness , about 97% of the companies perform virtual prototyping or large-scale data modeling just on their desktop computers, making them more likely prospects for HPC beyond their desktops. 57% of these companies said they have problems that they can’t solve with their existing desktop computers, and thus they have a real need for more computing power. According to Jon Peddie Research , workstations currently are a $7 Bio revenue market, worldwide, with about 4 million units shipped last year. That’s about 20 million users using mainly workstations and PCs for their daily design and development simulations, making them potential candidates for HPC Clouds.
More around this topic...
© HPC Today 2020 - All rights reserved.
Thank you for reading HPC Today.