As associate professor of computer science David Bindel takes on the mantle of Associate Dean for Diversity and Inclusion for Computing and Information Science (CIS), we acknowledge the contributions of associate professor of computer science, Hakim Weatherspoon, as he completes his term as Interim Associate Dean. Looking ahead to 2021, Weatherspoon will be on sabbatical, in part, to pursue an entrepreneurial venture with professor of computer science, Robbert van Renesse and Zhiming Shen, Ph.D., '17, currently a postdoctoral researcher.
Exotanium, a start-up launched in 2018, is now—after a fresh round of funding—poised to grow. In her profile of the effort, "Cloud software startup raises $1M, signs pilot deal," Melanie Lefkowitz of the Chronicle traces the origins of the company, its technological offering, and its plans for the near future.
The cloud was initially viewed as an opportunity for cost savings, since companies would no longer have to own and maintain physical servers to run their applications. But the cost of renting servers in the cloud—paid to cloud providers such as Google, Microsoft and Amazon—can consume as much as 50% of businesses’ revenue, Weatherspoon said.
Because cloud companies must have application servers available when clients need them, cloud providers often have temporarily unused servers, known as spot server instances, which they make available for a fraction of the cost of their regular service. Although, significantly less expensive, these spot server instances can be terminated at any time, making the cloud spot market unusable for most kinds of application servers.
Exotanium’s technology—called “X-spot” and based on Shen’s doctoral thesis—allows businesses to effectively navigate what’s known as the spot market, so they can take advantage of the low prices when temporary spot servers are available and then move to more reliable cloud servers when it isn’t.
“The underlying technology is called a container, meaning that you have your application and its environment, everything that it needs, packaged together, which makes it easy to test and deploy in different places,” Weatherspoon said. “And then we have an artificial intelligence base scheduler to optimize where we’re actually placing your application.”
Their research shows that the method could save businesses up to 90% on cloud application server fees, giving the company a huge advantage in a market estimated to be worth $26 billion by 2024.