Elastisys welcomes Fog researchers Li and Mulugeta!

Elastisys is proud over its strong academic background. Our close connection to world-leading research in cloud automation and optimization gives us and our customers an edge over the competition. We are happy to welcome our newest coworkers Li Wu and Mulugeta Ayalew Tamiru! They both recently started working as data scientists at Elastisys. They have joint appointments as PhD students as part of the FogGuru project, and will be affiliated with the Technical University of Berlin and University Rennes 1, respectively.

Li Wu, automating SRE with Machine Learning

Li Wu comes from China. She has two prior academic degrees; a BSc in telecommunications engineering from Hohai University, and an MSc in information and communication engineering from Southeast University in China. She has worked as a software engineer at both Ericsson Research and IBM. Her most recent work, at IBM, was as a site reliability engineer (SRE) at IBM Cloud. There, she operated large-scale cloud systems, and identified and handled many errors in the process. Sometimes, she had to employ neural network methods to help identify problems, which she could then fix.

Mulugeta, experienced developer and DevOps engineer

Like his new colleague, Mulugeta also has two academic degrees. He holds a BSc in electrical and computer engineering from Addis Abeba University in Etiopia and an MSc in Cloud Computing and Services from the European Institute of Technology. The latter included studies at the University of Rennes 1 and the Technical University of Berlin. He has a professional background as developer, sysadmin, and DevOps engineer in Etiopia and Germany. He has worked significantly with with public clouds and container environments.

Looking ahead

Together, Li and Mulugeta will research next generation technologies for managing application resources in future Fog computing infrastructures. Li will put her SRE experience and machine learning skills to use and focus on anomaly detection. Mulugeta will work on autotuning resource management platforms, configuring them for optimal performance and reliability. Together, these projects form key building blocks for a self-driving Fog computing platform.

Business case for Fog computing automation

Fog computing is the next generation of cloud computing. It moves computation much closer to where data is generated. Your intelligent car of the future, for instance, will host many sensors. The data these generate must be processed and sent to the cloud. However, network connectivity will not always be available. This means that processing must be able to seamlessly happen on a nearby node instead. Perhaps the car itself, or perhaps somewhere along the network path. But the connected future cannot rely on merely sending all data to a public cloud provider on the same continent. The network demands would be too large for that.

At Elastisys, we already hear customers and partners describe how difficult they find running hybrid cloud solutions. That is, an on-premise cloud cluster combined with one or more public cloud regions. This shows how immature these platforms are: that less than a handful of them cause headaches for operators. Now multiply that headache by a large factor, and that gives a hint at the Fog future. It is made even more challenging by the fact that Fog nodes may range from data centers down to e.g. embedded devices, such as Raspberry Pi nodes.

To make the Fog vision come true requires much smarter infrastructure and tools for deploying, scaling, and optimizing applications. For certain, platforms as Kubernetes will require significant work to meet these demands. Stay tuned to keep up with how Li, Mulugeta, and the rest of the Elastisys team tackle these challenges!

Do you want to get a taste of the future? Check out our blog post about running Kubernetes on Raspberry Pi nodes!

Leave a Reply