Super computing was formerly the preserve of organisations with a need for complex modelling or complicated simulations. The advent of the Cloud has changed the situation radically with the emergence of the high computing market (HPC). Andrew Bull explains what this means for public sector organisations and how they can benefit from this new ‘on demand’ facility to tackle such problems as work load peaks.
Go back to the beginning of the supercomputing era in the 1960s and the technology was used primarily for specialist applications ranging from weather forecasting to complex telephone switching systems to nuclear weapon testing. Typically, individual applications required the investment of huge amounts of money, resources and skills.
The high-performance computing market grew up around these early implementations and remained narrowly-focused. It was dominated by organisations needing to carry out complex simulations and research from large supercomputing centres. The organisations involved included pharmaceutical companies testing new drugs, leading research groups and academic institutions as well as scientists looking for ways of predicting natural disasters.
Generally, only these kinds of organisations have been able to attract European funding for specific programmes of activities, or obtain access to large IT budgets for one-off projects, without being constricted by having to conform to a cost-focused business case. Often, they would buy complicated IT systems which would incur significant implementation and maintenance costs and need regular refreshes to accommodate new technology. The endeavours of these kinds of organisations would invariably require the application of complex algorithms to vast volumes of data in a bid to get results that are as near to real time as possible.
A New Model of Performance Computing
This situation has changed dramatically over recent years. The HPC approach originally evolved steadily over time to meet an ever increasing portfolio of needs and significant challenges against a backdrop of continuous performance development.
Today, the Cloud is changing the whole model, delivering new ways to produce performance computing from dynamic provisioning to automation and orchestration but also new ways to consume IT, including via as a service, on-demand and configurable models. We are seeing strong growth in cloud usage in the UK public sector among organisations with generally much more modest data requirements than the original pioneers. Research carried out by the Cloud Industry Forum in January 2011 indicated that the public sector was lagging well behind the private sector in terms of cloud adoption. The most recent research indicates a dramatic increase in public sector cloud adoption moving them slightly ahead of the private sector (62 per cent and 61 per cent respectively).
This growing usage puts many public sector organisations in a good position to take advantage of the new cloud-based HPC model. In the past, few such organisations have had the money to justify the capital investment required to install the latest supercomputers. Today, they no longer have to. Instead, they can get access to all that processing power in the Cloud. Rather than aspiring to be the smartest buyers of HPC, they increasingly want to be the smartest users of it.
Cloud is helping to democratise the usage of HPC across all sectors of the economy by driving new opex based economic models; lower entry barriers; lower costs through shared pooled resources and the ability to share workloads within communities and industry sectors.
Cloud HPC, sometimes dubbed “HPC-on-Demand” has the potential to appeal to any organisation with fluctuating power needs and limited budgets. Usually, it involves the solutions provider investing in infrastructure that gives prospects the opportunity to buy access to that computing resource rather than make an upfront investment in a complex IT hardware implementation.
Much More than the Public Cloud
Unlike commercial clouds, HPC-on-Demand services address the requirements of HPC customers by providing on-demand remote compute facilities access with a pre-installed and configured environment where independent software vendor applications and open source codes are installed and available.
HPC-on-Demand services are usually targeted at users with complex modelling and simulation needs. Public cloud resources designed in a more one-size-fits-all fashion cannot match the requirements of these kinds of users.
In contrast, the key to the success of any HPC-on-Demand service is flexibility. The customer can choose everything from the operating system to the task scheduler and the size of the storage capacity used. They should also have the flexibility to choose from a range of different services whether their preference is for Infrastructure as a Service Platform as a Service or Software as a Service.
The new HPC-on-Demand service is likely to attract attention from organisations that already have their own computing resources but are looking for additional processing power to deal with workload peaks. It should also appeal to smaller departments and offices that have basic computing resources and would like to seize the opportunity to outsource hardware and technical support so they can concentrate on their core business.
One of its greatest benefits however, especially in today’s continuing economic downturn, is the potential it offers to kick-start economic recovery and contribute to the success of UK PLC by helping the long-beleaguered manufacturing and engineering industry sectors achieve operational efficiency.
HPC-on-Demand can drive ‘time to insight’ – the time taken between the presentation of the problem and reaching an understanding of how to solve it. By shortening this cycle, manufacturers have the opportunity to cut the design phase, add time to the build space and reduce time to market. HPC gives manufacturers the opportunity to drive through innovation, understand more complex issues, achieve more accurate and predictable outcomes and make product development faster and more efficient. It is a principal most commonly applied in the private sector but the concept can apply equally well to public sector organisations. And all of this has the potential to translate to renewed economic recovery by positively impacting on GDP.
HPC-on-Demand, in particular, addresses the major difficulty that many organisations have in managing their requirement for IT systems: their inability to accurately anticipate their future workload. In most cases, there will be no predictable pattern of need. If an organisation implements HPC in a traditional way, they often run the risk of either having to overprovision or being limited in the design work they can do because they do not have sufficient equipment in place.
By meeting peak requirements on demand, organisations no longer need worry about the complexities of running their own environment or the negative impact this might have on profitability. Instead they can tap into the available computing capability across the web as and when needed.
Freed from the shackles of technology, organisations now have complete freedom to innovate; with HPC-on-Demand, they have neither need for physical infrastructures nor maintenance, while still having access to high-performance tools for innovation. HPC-on-Demand can bring a raft of benefits to private and public sector organisations and the broader UK economy. The earliest developers of supercomputers back in the 1960s, who were primarily focused on pushing the technical boundaries of what could be achieved, would be amazed and delighted to see the sheer breadth of business applications that high-performance computing can deliver today.
Andrew Carr is CEO at Bull UK & Ireland.