With HPC the Future is Looking Grid

By Andy Morris, IBM Cognitive Infrastructure

July 22, 2019

Gone are the days when problems such as unraveling genetic sequences or searching for extra-terrestrial life were solved using only a single high-performance computing (HPC) resource located at one facility.

Today, HPC can involve thousands or even millions of individual compute nodes – including home PCs –located on different continents or with no certain location at all in the cloud – all marshalled to process hundreds of different applications at once.

This is grid computing.

The grid computing concept differs from parallel computing in supercomputers, which normally run highly connected applications, rather than independently functional nodes. Supercomputers operate over high-speed networks and are typically housed in one specialized data center. Grid computers, on the other hand, exchange little or no data and feed the project over Internet connections from geographically dispersed locations.

Cloud computing is another form of distributed computing that lies somewhere on the computing spectrum between grids and supercomputing. Cloud resources can be geographically distributed, but only to a few locations, as compared to thousands or millions feeding a grid project.

Grid computing is used to address projects such as genetics research, drug-candidate matching, even the search – unsuccessfully so far – for the tomb of Genghis Khan. The SETI@home project searches for extraterrestrial intelligence using millions of PCs running an application that analyzes segments of radio telescope data.1

As it turns out, the financial services industry offers a very good example of the power and advantages of grid computing. Capital markets never stand still, and investment banks must contend with increasingly complex risk environments and regulatory landscapes to succeed. To make sense of the many variables involved, firms rely on financial modeling applications that help them evaluate risk and price financial instruments while complying with industry guidelines.

[Also read how financial services companies are addressing the challenge of cluster sprawl.]

Excelian, a global consulting and software engineering firm, designs, engineers, operates, and optimizes technology and business solutions for companies in the financial services sector. “To support risk analysis and pricing, many of our clients rely on grid architectures, where commodity servers are linked together to provide huge performance at relatively low cost,” explains Karl Fallon, HPC Grid Management Consultant at Excelian.

In the past, homogenous environments were the norm. Today, more of Excelian’s clients are mixing components and operating systems within a single grid architecture. Excelian sought an effective management solution to ensure that it could keep cost and complexity under control for client environments increasing in size and diversity.

Clients don’t want to be locked into using one type of hardware or operating system,” says Fallon. “Instead, it’s important that we can build grid infrastructure tailored to their specific requirements, which can be adapted in the future. To ensure that we always meet the changing demands of our clients’ businesses, we wanted a solution that minimizes management effort for grid computing infrastructure, while driving great outcomes.”

The grid computing solution that Excelian and its clients have chosen for years is IBM Spectrum Symphony.

IBM Spectrum Symphony is an enterprise-class grid manager for distributed application services on a scalable, shared, heterogeneous grid. It can support up to 5,000 compute nodes, 128,000 cores, and 300 applications. It has the flexibility to adapt to changing priorities and can reallocate more than 1,000 compute engines per second to different workloads in accordance with defined sharing policies and application priorities. This translates into better application performance, better grid computing resource utilization, and an ability to respond quickly to business demands.

IBM Spectrum Symphony has become a crucial part of our daily toolset for many client environments,” notes Fallon. “It gives us the ability to ensure that grids just work, regardless of what is thrown at them. As a turnkey solution, Spectrum Symphony doesn’t take long to set up, so we can deploy it and start providing value to clients fast.”

For some clients, Excelian uses IBM Spectrum Symphony to manage hybrid cloud environments. The solution enables grids to burst into the cloud if demand exceeds the capabilities of on-premises infrastructure.

By reliably equipping decision makers with the information they need to make the right strategic choices, the IBM grid computing solution enables Excelian to give its clients an edge over their competitors. Karl Fallon states: “IBM Spectrum Symphony is extremely versatile, which allows us to consider evolving client environments in many different directions, including the cloud. It is the ideal foundation for where many of our financial services clients are today, and where they are headed in the future.”

[Now discover how industry leaders are making progress on the last mile challenge for HPC.]


References

[1] TechTarget: Grid Computing https://searchdatacenter.techtarget.com/definition/grid-computing

Return to Solution Channel Homepage
HPCwire