National Integrated Cyber Infrastructure SYSTEM
The National Integrated Cyber Infrastructure System (NICIS) promotes scientific and industrial development through the provision of high-performance computing capability, high-speed network capacity and a national research data infrastructure integrated hierarchically into globally connected systems and into local system systems, providing seamless access for the research and education communities of South Africa. It is a national initiative of the Department of Science and Innovation and implemented by the Council for Scientific and Industrial Research (CSIR).
The Centre for High Performance Computing (CHPC) is one of the three pillars of the NICIS. It provides massive parallel processing capabilities and services to researchers in industry and academia. The other main pillars are the South African National Research Network (SANReN), which provides high-speed connectivity and advanced networking services, as well as the Data Intensive Research Initiative of South Africa (DIRISA), which implements services that enable sound data management practices and support efficient data-driven scientific and engineering discoveries.
Advancing high-performance computing in South Africa: The CHPC
In 2016, the CHPC achieved a significant milestone in servicing its user base and in advancing high-performance computing in South Africa. To further improve the centre’s computational power, the CHPC introduced a peta-scale machine, with almost 33 000 cores. The machine was the fastest supercomputer in Africa and at launch, it was ranked 121 amongst the world TOP500 supercomputers. A peta-scale machine is a super computer with processing speeds capable of a thousand-trillion floating-point operations (FLOPS) per second. FLOPS are used in computing to compute extremely long numbers. The system, named LENGAU (Setswana name for Cheetah), owing to its speed of 1 000 Teraflops, is 15 times faster than the previous system of the centre. In 2018 the compute resources at the CHPC were expanded to include a Graphical Processor Unit (GPU) cluster consisting of 30 state of the art V100 GPUs.
The CHPC is providing leadership to South Africa and Africa on developing strong high-performance computing capabilities and supporting large-scale science projects. The Tier-2 CERN service provides 2 400 jobs per day to the ALICE (A Large Ion Collider Experiment). In support of the Square Kilometer Array (SKA) partner countries, the CHPC has provided computational clusters to Botswana, Madagascar, Namibia, Zambia, Ghana, Mozambique, Kenya and Mauritius. Locally, the University of Venda, Sol Plaatjie University, University of Fort Hare, the University of the Witwatersrand, Hartbeesthoek Radio Astronomy Observatory and Stellenbosch University received training and equipment for small-scale processing.
Human capital development is critical to the sustainability of cyber infrastructure and the CHPC provides high-end skills, relevant to the respective science and engineering domains of the research community. One of the CHPC’s flagship projects, the Student Cluster Competition, continues to produce world-class undergraduate students with high-performance computing skills. The CHPC has secured top accolades at an international level for years since 2013. The annual CHPC National Conference is also an important flagship project that contributes to human capital development and is a platform for sharing and feedback from the users of the infrastructure and the international and domestic stakeholders.
The industry programme at CHPC focuses on providing services such as stable and reliable high-performance computing systems, world-class consultancy in the areas of fluid-dynamics, materials science, finite-element analysis, discrete element modelling and general design of high-performance systems, optimised for specific clients. These services are currently used by South African industries in oil and gas, mining, engineering designs and weather forecasting.
Providing high-speed connectivity to research and education: SANReN
Government has invested in a dedicated national research and education network that is operated by the Tertiary Education and Research Network of South Africa with the CSIR-hosted South African Research Network (SANReN) in charge of its rollout and the incubation of services that run on the network. The aim of this investment is to improve the country’s cyberinfrastructure, specifically to provide high-speed connectivity to schools and higher research facilities.
Progress has been made in 2016 in growing and upgrading the network and ensuring connectivity for the higher education research community. SANReN has connected a further 11 higher education and research sites, including the Walter Sisulu University campuses in Queenstown and Whittlesea, as well as the South African National Space Agency Space Science site in Hermanus. This brings the total number of sites, connected to the network, to 210. The capacity of the links to various higher education and research sites were also upgraded, thereby growing the average bandwidth available per site from 2.8 to 3.5 Gbps.
In 2016, SANReN added the Mconf web conferencing tool that enables researchers to collaborate seamlessly. It uses web real-time communication, the latest technology in this area. The service has been particularly useful to the Human Heredity and Health in Africa (H3Africa) initiative. H3Africa scientists study genomic influences on disease across their continent, from differences in the progression of HIV in children, to developing new sequencing methods for the Ebola virus, to collecting more than 35 000 nasal swabs from children to show how concentrations of nose and throat micro-organisms may play a role in pneumonia.
During 2017 the South African National Research Network (SANReN) initiated the Cyber Security Games project which was aimed at establishing the hosting of an information security student competition with an emphasis on network security. This resulted in the first Cyber Security Challenge 2017 being hosted at the CHPC National Conference 2017. Building on the success and high interest of the Cyber Security Challenge 2017, SANReN hosted the Cyber Security Challenge 2018 at the CHPC National Conference 2018 held in Cape Town.
Coordinating data-intensive research: DIRISA
The Data Intensive Research Initiative of South Africa (DIRISA) promotes, enables and coordinates a data-intensive research ecosystem in support of national science and strategic priorities. This is accomplished through several key objectives, including the provision of a robust national data infrastructure and services, the promotion of sound data stewardship practices and the development of underpinning expertise in data management and data intensive research. DIRISA operates as an overarching national data organisation and as such, advocates research data sharing, coordinates publicly-funded initiatives and advises on a strategic agenda for data intensive research.
Towards attaining these objectives, DIRISA is expanding the national data infrastructure to accommodate increasing demands for research data storage and coordinates the establishment of a national e-Science postgraduate training and teaching platform and of a regional research data centre by a consortium of academic institutions. Data services developed by DIRISA include an Archive, Discovery and Analytical services that support more rapid data-based research innovation across all academic disciplines.
In supporting researchers to manage their research data, DIRISA has released an operational version of a research data management planning tool, called DMP Online. It is becoming standard practice for research funders to require such a plan as part of the research proposal as it supports improved data-driven research across all disciplines. As a collaborative effort by DIRISA and Universities South African (USAf), DMP Online will be demonstrated to local universities and research institutions to promote its adoption for use nationally.
In collaboration with the Department of Telecommunications and Postal Services, and following ministerial approval, DIRISA has established a Multi-Primary Administrator (MPA) in South Africa. This node forms part of a global network of MPAs that manage the allocation of unique and Persistent Identifiers (PIDs) for digital objects. The use of these PIDs greatly improves the management of digital objects, including data and valuable national assets.
DIRISA plans to host its first annual Student Datathon during the CHPC National Conference 2019. The Datathon will give students an opportunity to use open data to come up with creative and innovative solutions to assist in solving some of South Africa’s challenges. The DIRISA Student Datathon encourages the use of open data to promote open science and data science skills. Therefore, for this datathon, students will only use open data.