Marios Kogias awarded 2019 IBM PhD Fellowship

Marios Kogias studies tail-latency optimizations in datacenter systems. © Marios Kogias

Marios Kogias studies tail-latency optimizations in datacenter systems. © Marios Kogias

Marios Kogias, a doctoral student in the EPFL School of Computer and Communication Sciences (IC) Data Center Systems Laboratory (DCSL) led by Professor Edouard Bugnion, has been selected to receive a competitive two-year IBM PhD Fellowship Award.

Kogias’s research in the DCSL focuses on datacenter systems, and specifically tail-latency optimizations. While typical communication latencies inside datacenters consist of a few microseconds, existing systems have been designed with millisecond-scale assumptions, which requires redesigning the entire software stack specifically for latency-critical applications. Kogias is currently working on transport protocol support for remote procedure calls and in-network computation, taking advantage of the programmability introduced by emerging network appliances.

Since the 1950s, the IBM PhD Fellowship program has honored and supported “exceptional PhD students who have an interest in solving problems that are important to IBM and fundamental to innovation in many academic disciplines and areas of study,” including artificial intelligence, blockchain, cloud computing, data science, quantum computing, security, and the internet of things (IoT). In addition to support for living expenses, travel, and attending academic conferences, fellowships provide the opportunity to work with an IBM mentor.