Joe Armstrong - new adjunct professor
The inventor of the programming language Erlang, Joe Armstrong, has become an adjunct professor at the School of Information and Communication Technology and the Department of Software and Computer Systems.
Tell us about your research area.
– Currently I'm interested in the following three areas: building planetary-scale fault-tolerant distributed systems, programming massive Network On Chip (NOC) multi-core arrays and low Energy Computing. I've been involved in programming language design and implementation since 1985 when I invented Erlang. Erlang was designed for building large scale fault-tolerant distributed systems and has been used to build some of the largest distributed systems in the world. I'm now interested in extending Erlang so we can build planetary wide distributed systems, and implement these on energy efficient meshes of multi-core computers.
Which challenges will you and your research area face in the nearest future?
– We need a better understanding of where the energy is consumed in a computation. Programming large multi-cores and large networks is difficult because a lot of the assumptions we make about conventional systems are incorrect when applied to new architectures. Where to physically perform a computation becomes very important. Should we move data or programs? How and when should we turn processors on and off when performing computations. How do we deal with failures and build self-healing systems?
– Above all, how do we manage complexity? Control of complexity has always been the key problem in computer science but as systems become larger controlling complexity becomes more important. Even tiny errors can creep into the system and cause unexpected failures from which we must be able to self-repair and continue without causing catastrophic failures.
What will the adjunct professorship entail?
– I hope to be able to act as a go-between linking up research at KTH with product development at Ericsson. I've worked for 20 years in Ericsson and 4 years at SICS and started a couple of companies, so I hope I can connect together researchers with product developers and entrepreneurs and see what we can do together. Having been a researcher and entrepreneur I think I understand some of the mechanisms involved in turning research ideas into products and how problems encountered in product development can lead to fruitful new areas of research. I hope to be able to contribute to this process.