DeepSense Combines HPC and AI to Bolster Canada’s Ocean Economy

By Tiffany Trader

November 16, 2018

We often hear scientists say that we know less than 10 percent of the life of the oceans. In September, IBM and a group of Canadian industry and government partners launched a public-private project intended to take on, from an economic perspective, our lack of oceanic insight.

DeepSense, a high-performance computing and analytics platform aimed at spurring growth in the “blue” ocean economy, including development of new food sources and offshore energy, predicted to hit $3 trillion by 2030*. Powered by IBM and Nvidia technology, DeepSense has both a commercial and scientific underpinning; more than 200 companies and 60 technology start-ups will leverage the platform to generate insights from vast ocean data for conservation research and economic advantage.

In a blog post (Sept 10), Ayman Antoun, president, IBM Canada Ltd., remarked on Canada’s unique position of having the world’s longest coastline. Antoun’s relayed that Canada is in an enviable position to capitalize on the projected growth of the multi-trillion-dollar ocean economy, but this can’t be done without data-driven insight propelled by new technologies, and it can’t be done by a single organization, he said.

In addition to IBM, collaborators include Dalhousie University, the Ocean Frontier Institute, the Centre for Ocean Ventures and Entrepreneurship (COVE), the Government of Canada, local industry and the Province of Nova Scotia.

“It is certainly a grand challenge putting emerging technology, infrastructure and services to work gleaning insights from vast amounts of ocean data,” observed Antoun. “But the DeepSense collaborative team is fully equipped to meet it. We are already making great strides to progress the important work of fueling the new regional ‘blue’ economy where ocean related activity generates $5 billion in revenue and produces 60,000 jobs in the Province of Nova Scotia alone.”

The DeepSense cluster is comprised of 40 IBM Power Systems nodes, 10 of them the Power8-based S822LC (“Minsky”) servers that are the precursors to the Power9 AC922 “Newell” boxes used in the world’s fastest supercomputer, the CORAL Summit machine at Oak Ridge National Lab. The Minsky servers are equipped with two Nvidia P100 (Pascal) GPUs and of course the tight NVLink coupling between the Power CPU and Nvidia GPU that is the hallmark of the Minsky and Newell platforms. Nodes are connected with 100 Gigabit Ethernet with signs pointing to Mellanox as supplier. Partners have the option of adding another 20-nodes of x86 infrastructure to the cluster, which is housed at Dalhousie University.

DeepSense cluster diagram
DeepSense cluster node specifications

IBM has also supplied 3 petabytes of IBM Spectrum Scale software-defined storage, as well as IBM Spectrum Coordinator and Spectrum LSF for resource and cluster management.

The GPU portion of the platform leverages IBM PowerAI software featuring existing precompiled major deep learning frameworks to support machine learning development.

The total cost for the project, including hardware, software and operation for five years, comes out to about $18 million Canadian (~$14 million USD). Nearly $6 million Canadian comes from the Canadian government’s Atlantic Canada Opportunities Agency (ACOA) with IBM Canada providing an in-kind contribution valued at $9.8 million. Dalhousie University and the Ocean Frontier Institute invested $2.1 million Canadian in the project.

“The DeepSense platform will build data analytics capacity in the ocean sector, attract investment opportunities, help generate wealth and create well-paying jobs,” noted ACOA.

One of the first use cases devised for DeepSense is the optimization of algae production as a sustainable food source. Using underwater wireless sensors, RealTime Aquaculture provides environmental monitoring for ocean farmers, delivering actionable intelligence gleaned from 11 million data points.

Other target areas for Canada’s “blue” economy include fisheries & aquaculture, seaport & logistics, security & defense, offshore energy, shipbuilding and even ocean datacenters.

* Source: Organization for Economic Co-operation and Development (OECD)

Feature image depicts the Centre for Ocean Ventures and Entrepreneurship (COVE).

Return to Solution Channel Homepage

IBM Resources

Follow @IBMSystems

IBM Systems on Facebook

Do NOT follow this link or you will be banned from the site!
Share This