Image: Huawei
The winning team of the Huawei Tech Arena competition were commended for being able to effectively communicate their results.
STEM students recently got the chance to go head-to-head in a cloud computing competition that allowed them to show off their analytical modelling and optimisation skills.
At the Huawei 2024 Ireland Tech Arena, teams were tasked with optimising data centre inventory with the top performers taking part in a two-day hackathon in Dublin.
The competition has been running since 2018. This is the first year that it has been extended to students in the UK.
Stefano Mauceri, the technical challenge author and a senior research engineer at Huawei Ireland, said the competition requires teamwork, innovation and determination, offering students a supportive environment for learning.
“Teams of students collaborate to tackle a technical challenge, requiring them to think creatively and develop solutions that could work in a real-world business environment,” he said.
“Through this year’s challenge, we aimed to demonstrate that real-world problems are often complex and open-ended, with shifting or undefined goals. As a result, there is rarely one correct answer; instead, multiple solutions may exist, each with its own set of trade-offs.”
This year’s challenge focused on server fleet management at scale. Ananth Shrinivas Srinath, CTO of cloud reliability and director of Huawei Ireland’s cloud reliability lab, said running efficient data centres with vast fleets of servers is a competitive business.
“Even small improvements in how these servers are built, operated, repaired, recycled and decommissioned can have a significant impact, not only on the cloud providers’ revenue but also on the global supply chains behind them and the environmental sustainability of this growth,” he said.
“We aimed to provide the students participating in the Tech Arena competition with a technical glimpse of the optimisation problems they might encounter in this domain, without overwhelming them with the messy complexities of real-world situations.”
The top prize was €6,000 for the winning team, while second and third place received €4,000 and €1,500 respectively.
“In addition to cash prizes, successful participants gain access to invaluable opportunities such as internship programmes, networking with industry professionals, and the possibility to visit Huawei’s headquarters in China,” added Mauceri.
The winning team
After narrowing it down to the top performing teams for the hackathon, students were able to build on the work they had done, collaborating with their team and receiving support from mentors.
At the end of the two-day event, the eight teams that reached the final had to pitch their solutions to a panel of experts for a chance to win prizes.
“What struck me most was that each team employed a completely different technical approach, such as simulated annealing, linear programming, constraint programming, demand forecasting, rule-based methods, and dynamic programming – to tackle the open-ended optimisation problems presented to them,” said Shrinivas Srinath.
“Another observation was that the teams that were well organised and communicated effectively among themselves were able to iterate more and articulate their results to the judges with greater clarity.”
This, he added, was particularly evident in the overall winning team, which comprised Boi Mai Quach, Thang-Long Nguyen-Ho, Thao-Nhu Nguyen, and Nam Trinh, all PhD students at Dublin City University, and Long Do Thanh, a master’s student at Technological University Dublin.
Trinh said the team first approached the challenge from a business insights point of view using linear programming (LP). “For example, on deciding how many servers to purchase, we had to account for slot constraints at each data centre and assess how much demand could be met,” he said.
“The key was to accurately define the constraints and explore how they correlated to other actions. By using LP, we were able to effectively manipulate variables and constraints, allowing our optimiser to maximise profit.
“Our solution has a runtime of approximately one minute per demand without any need for hyper-parameter tuning. Additionally, it’s adaptable. If the problem changes, such as adding a pricing strategy, our solution remains highly effective.”
Mauceri said the winning team stood out not just for technical expertise but for their ability to analyse and clearly communicate their results and processes.
“In many cases, business stakeholders are unfamiliar with the intricacies of complex technical solutions. Effectively explaining how a solution works and its alignment with business goals is essential for building trust and securing their support,” he said.
“It has been truly inspiring to witness such remarkable talent and dedication lead to success. I have no doubt this team has a bright future ahead, and I am eager to see how they will continue to make meaningful contributions to Ireland’s tech community.”
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
