Nowadays, the flow of information is happening at an incredible pace and on a huge scale. When we look at the example of a long-distance video call between a person located in New York and in Warsaw – at first glance it does not seem amazing. Is it really so? In a straight line, these cities are separated by a distance of 6854 kilometers. So how do video and audio reach the callers without a noticeable delay?
Data transmission – scale and current challenges
The problem of data transmission concerns both the speed and capacity of the links used (Bandwith). This poses quite a challenge for data transmission, since on a classic fiber-optic link the Internet uses the TCP/IP protocol which actively exhausts only about 20% of the bandwidth. When we consider the huge data sets that are processed and transmitted in genomics, bioinformatics, or space observation projects; the needs for fast and reliable data transfer become a key issue. In the near future, space observations, for example, using the Square Kilometer Area radio telescopes now under construction, space observation data will be generated at a rate of 1,000 Petabytes per day. This is such huge data that, despite its reduction and partial analysis “in flight”, storing it in a single Data Center is unrealistic – it is necessary to carry out partitioning, transfer and storage of data in different centers.
Data transfer is an important issue when processing huge data sets. However, let’s go one step further. Assuming that we don’t just want to store data, but also use it – we face an even bigger challenge – ensuring sufficient computing power. This problem was addressed by the InfiniCortex project conducted between 2014 and 2016 in Singapore with the participation of several dozen networking organizations, global technology companies, universities and research entities around the world. It was an unprecedented and unique event on the supercomputing scene. The team was led by Dr. Marek Michalewicz, then CEO at A*STAR Computational Resource Centre, Singapore. Due to the global nature of this project, more than forty organizations around the world were partners, among them the so-called. NRENs (National Research and Education Networks) such as Internet 2 in the US, ESNet of the Department of Energy in the US, GEANT in Europe or the Polish scientific and educational network consortium PIONIER created and managed by the Poznan Supercomputing and Networking Center, as well as commercial fiber operators worldwide such as Tata Communications and PacificWave, and a number of universities such as Tokyo Institute of Technology; Australian National University, AU; Georgia Institute of Technology, Georgie, USA; Stony Brook Univeristy in NY, USA, Georgetown University, D.C. USA; Reims University in France and Poznań Supercomputing and Networking Center and Interdisciplinary Center for Mathematical and Computer Modeling (ICM) at the University of Warsaw. Thanks to the commitment of all teams and the leadership of the A*CRC in Singapore and funding from commercial partners, surprising results were achieved. [1].
Where are supercomputers being used?
Supercomputers are now an extremely important tool in the fields of defense, climate research and weather prediction, engineering, genomics, medicine and medical diagnostics, and more recently in the field of artificial intelligence and so-called Large Language Models. They make it possible to perform advanced genomic analysis, diagnose diseases, and research new therapies. They are used to process huge amounts of data from any field of science in particular – genetic – this allows accurate analysis of patients’ genomes and in-depth study of genetic mutations associated with diseases. The merit of supercomputers is the rapid diagnosis of genetic diseases and the creation of individualized treatment plans. Medicine is a large field of development for the use of technology. Here, supercomputers are also used to simulate complex biological processes, making it possible to test new drugs and conduct research on new therapies. For this purpose, computer simulations are used to test different drug options and choose the best solution.
With the development of medical robotics and genomic research, supercomputers will be an essential tool for personalized medicine. Sequencers are increasingly appearing in every hospital, enabling genomic testing. However, in order to accurately analyze the material taken, huge computing power and a good link between the hospital and the data center are needed.
What is the InfiniCortex project?
InfiniCortex focuses on four aspects, the most important of which is high-bandwidth intercontinental connectivity between Asia (Singapore, Japan), Australia, the US and Europe (Poland, France). Also important are workflows and applications on a concurrent distributed computing infrastructure, as well as connecting separate InfiniBand subnetworks [2] with different network topologies to create a single computing resource. InfiniCortex has made it possible to create an “interconnection” of supercomputers globally – combining their computing power, increasing performance and creating a “Galaxy of Supercomputers.” [1]
At the Austin conference, the most elaborate form of InfiniCortex was presented, which included the world’s first InfiniBand lap around the globe, with most of the path running at speeds as high as 100 GB/s! The solution was tested from multiple locations around the globe and involved: genomics services, bioinformatics, medical image analysis, fusion reactor simulation of energy harvesting, asynchronous linear solvers, ADIOS framework, distributed HPC storage and resource virtualization. The concurrent supercomputing solution has the potential to revolutionize data transfer and the globalization of peri-terrestrial computing. This is the first such solution provided by civilian scientists – previously the technology was used by NASA or unofficial military projects but only for transferring big data of military importance.
Potential for the Polish market
In the interview, Dr. Marek Michalewicz repeatedly noted the scale of the project’s application possibilities. He also emphasized the potential of the Polish market in this regard:
- “Poland has fantastic opportunities in terms of supercomputing development. Already at the beginning of the 20th century, the Poznanians created the PIONIER Consortium, providing fiber-optic connections. Poland is one of the few countries that have such a solid physical foundation for data transmission. In addition, the fiber optic technology and cabling is durable – it can last for many years. This is an excellent initiative.”
Thanks to the fiber-optic network, which is spread across the country, scientific organizations in the country do not have to pay commercial rates to use this bandwidth. This makes it possible to conduct advanced scientific research, including that on supercomputers. Poland has been actively involved in European projects related to the development of supercomputers for many years, most recently under the European EuroHPC program. More recently, Poland has been involved in a European project on secure data transmission using quantum technologies.
The technology used by the Consortium is the PIONIER broadband optical backbone network, which integrates 21 academic metropolitan area networks – MAN and 5 High-Power Computer Centers [3]. There is no doubt that supercomputing will continue to grow, and with it its use in all fields of science and engineering, as well as in molecular biomedicine (so-called omics), genomic data analysis, remote diagnostics, and cloud solutions.
The development of supercomputers is extremely important for the country’s technological future. Poland has a chance to become one of the leaders in research and development and applications of supercomputers. This opens up new development opportunities for the country and allows it to attract new investments.
Bibliography:
- InfiniCortex – From Proof-of-concept to Production
https://www.researchgate.net/publication/318673017_InfiniCortex_-_From_Proof-of-concept_to_Production (accessed: 26 May 2023) - InfiniBAND, https://en.wikipedia.org/wiki/InfiniBand , (accessed: 26 May 2023)
- Konsorcjum Pionier, https://pionier.net.pl, (accessed: 26 May 2023)
- InfiniCortex: Present and Future
https://www.researchgate.net/publication/303680868_InfiniCortex_Present_and_Future (accessed: 26 May 2023) - InfiniCortex: concurrent supercomputing across the globe utilising trans-continental InfiniBand and Galaxy of Supercomputers
https://www.researchgate.net/publication/269287382_InfiniCortex_concurrent_supercomputing_across_the_globe_utilising_trans–continental_InfiniBand_and_Galaxy_of_Supercomputers (accessed: 26 May 2023) - High Performance Computing as a critical discovery and support infrastructure in Bio-Medical Research and Practice https://sano.science/seminars/high-performance-computing-as-a-critical-discovery-and-support-infrastructure-in-bio-medical-research-and-practice/ (accessed: 26 May 2023)
Zuzanna Czernicka, Joanna Rancew
