Heads in the cloud
Monday 9 September 2019
This article originally featured in Swinburne's Research Impact magazine.
Swinburne research into artificial intelligence has been used to develop a suite of software called Smart Cloud Broker, designed to give individuals, online retailers and other internet-based service providers the tools to manage resources on ‘the cloud’ simply and efficiently.
In the past, computer servers sometimes crashed during demand surges — during an online Christmas rush, for example. Roughly 10 years ago, servers began to be replaced by a vast network of computers that provide computing resources via the internet, known as the cloud.
But a challenge for businesses and individuals is selecting the right service configuration so their cloud resources don’t become overloaded: for example, do you purchase resources as you need them, buy a monthly subscription, or buy in advance to get a lowest price?
For medium to large companies that use cloud-based computing, database hosting, searching and more, the number of possibilities becomes mind-boggling: the cloud-management industry in Australia alone is worth $4.4 billion.
In 2012, Professor Ryszard Kowalczyk from Swinburne’s Software Systems Lab, part of the Swinburne Digital Research Innovation Capability Platform, realised that artificial intelligence (AI) could help assess myriad scenarios and even negotiate with other computers.
“We realised there was opportunity to add value with the AI technology we’re developing, to not only save money, but to better utilise resources as a whole,” explained Kowalczyk.
The group developed Smart Cloud Broker, which uses AI to help both users and businesses make the most of cloud resources. One component is Smart Cloud Bench, which compares cloud services by performing real-time cost and service benchmarking and helps select the best match for a businesses’ needs. Another is Smart Cloud Marketplace, which is a platform on which cloud consumers and providers can trade resources, thereby balancing supply and demand.
Smart Cloud Purchaser then negotiates the best deal, using AI to forecast spot prices, and buys the resource. Finally, Smart Cloud Manager looks after a business’s cloud resources in real-time, for example using AI to predict sudden demand and avert overload by purchasing more resources.