High-Performance Computing 2019-20

HPC Yearbook 19/20

receive hardware and soſtware training, but they are not technology experts. Tere is no dedicated IT department on the vessels. For these reasons, Magseis Fairfield requires a storage solution with fast, simple operations and remote support. Survey teams also need a solution with an intuitive, graphical management interface, so non-technical personnel can perform maintenance. Panasas has intentionally designed its

systems to be as user-friendly as possible. In traditional HPOC environments, this helps to save on costs and maintenance, but in emerging markets or endpoint computing environments this can be the difference between a successful project or failure. Dale Brantly, director of systems

engineering for Panasas, reflected on the design principles that drive the development of new products. ‘It goes back to the beginning of the product design. When Panasas launched in 1999, the idea was to build an easy to use, easy to maintain appliance, so the product originated with that concept. ‘It is different than having a parallel file

system and then trying to make it easy to use – that is backwards from the design process. You need to have that thought that it is designed in such a way that it has as few replaceable parts as possible,’ Brantly

offshore projects has grown from 250TB per vessel to as much as 600TB. Tis data is critical for our clients, so we cannot afford failure of any type when it comes to storage.’

Rugged technology

Magseis Fairfield initially chose Panasas ActiveStor nearly 10 years ago to meet its shipboard data storage requirements for marine surveys. Te firm now runs several generations of Panasas storage technology throughout the enterprise. Together, these Panasas solutions provide 1,600TB of storage that supports a wide range of client projects. ‘Our storage solution must be robust

enough to survive the offshore environment,’ said Garcia, who is responsible for overseeing, managing, organising and facilitating the data for shipboard scientists, while ensuring data quality for clients. ‘Tis system is accessed 24 hours a day. We require flexibility to move equipment from one vessel to another, and one country to another. Te storage system must be able to tolerate this type of constant handling. We have considered other solutions, but none offer the full complement of technology and remote support Panasas does.’ Te geoscientists who use ActiveStor

Marine surveys are changing, covering wider geographic areas and acquiring much larger data volumes offshore

added. Panasas technology enables users to deploy and scale in hours, and user training requirements are limited. ‘Tanks to the Panasas architectural

design, it’s easy to move components, scale storage, test systems and keep things running effectively,’ says Garcia. ‘Our teams can perform maintenance and deploy new systems quickly. Te modularity makes the solution flexible enough to make any changes needed to support each client project.’

Reliable, consistent, and predictable performance

‘We want to make things easy to use, so you do not have to be a rocket scientist in a lab coat to use it. We provide an informative user interface, we have a support infrastructure

that helps to support that through, so people can understand error codes,’ said Brantly. ‘While I don’t think it is fair to say that anyone can do this – it does require some IT skills – but it doesn’t require a computer science graduate to get high performance from this system. Te system is designed to deliver high performance out of the box.’ A single offshore acquisition consists of

80 to 100 personnel, including navigators, geophysicists, marine mammal observers, and remotely operated vehicle pilots. Considering staffing a vessel with a full crew can cost firms more than $200,000 a day, an extended outage could cause a loss of hundreds of thousands, or even millions of dollars. A highly reliable storage solution reduces the risk of downtime. ‘Even when we bring on additional data,

Panasas still performs exactly as we would expect,’ says Garcia. ‘We can count on the solution to handle everything we can throw at it. Tat helps us meet our deadlines and ensure that clients get the information they need.’ Panasas’ 24-hour-a-day remote support

ensures that an expert is always available to geoscientists who need help troubleshooting storage issues. Equipment can be replaced when needed, but downtime is rare. ‘If our storage solution cannot meet client needs, our projects will miss the mark,’ says Garcia. ‘Part of the reason we’ve relied on Panasas for so long is that the solution and support they provide give us peace of mind.’ While uptime is important to many HPC

applications, the importance in this type of workflow is critical to the success of businesses like Magseis Fairfield. Without the ability to keep these systems working as intended, in harsh unforgiving environments for computing equipment, their business would not be successful. Removing the complexity and making

HPC as accessible as possible for users allows enterprises and academic or government institutions to reduce costs, which helps scientists and researchers focus on primary goals. Domain specialists or geophysicists should not have to worry about their computing infrastructure. Tis diminishes productivity and wastes valuable resources. Tortorello said: ‘We see this more oſten

in the smaller environments than the larger environments, such as a departmental lab or a genomics lab that is six guys doing their science. Tey don’t want to spend money to hire an IT guy, they want to do it themselves as a sideline. Tey do not want to spend more time with IT resources than they need to. ‘We aim to make HPC accessible to

non-IT experts, such as endpoint computing environments,’ added Tortorello.n


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32