Postdoctoral stipend (2 years) in Autonomous Resource Allocation for Rack-Scale Systems
Umeå University is dedicated to providing creative environments for learning and work. We offer a wide variety of courses and programs, world leading research, and excellent innovation and collaboration opportunities. More than 4 300 employees and more than 30 000 students have already chosen Umeå University. We welcome your application!
Umeå University announces one stipend for postdoctoral research in autonomous resource allocation for rack-scale systems. The stipend is part of a massive effort on autonomous systems for industry and society of the future, with eight postdoctoral stipends in eight separate subprojects. Welcome with your application!
The Department of Computing Science (www.cs.umu.se) is a dynamic department with about 100 employees from over twenty countries. We are providing research and education within a broad spectrum of areas, and offer education on basic, advanced, and PhD levels. The research is internationally well recognized and includes basic research, methods development, and software development, but also research and development within various application domains.
The currently announced stipend is associated with the research group in Distributed Systems, whose research primarily addresses autonomous management systems for cloud infrastructures and applications. The group includes around 20 people of which 7 reseachers and 8 PhD students. The research is to a large extent funded by eternal resources, including The Swedish Research Council, H2020, and the Wallenberg Foundation. The research is frequently performed in collaboration with leading industrial partners like Google, IBM, Intel, Ericsson, SAP, and Red Hat. For a presentation of the research group, please see www.cloudresearch.org.
We are opening a postdoctoral stipend for research on autonomous resource allocation for rack-scale systems.
The recent development of ever-faster computer networks but stagnating performance of CPUs and storage has led to the creation of a novel virtualization paradigm: rack-scale computing. Instead of subdividing a physical server into multiple smaller virtual machines as in traditional virtualization, rack-scale systems will combine resources from whole server racks into a single, virtually large, machine. This paradigm will revolutionize management of performance sensitive applications that cannot be provisioned in a single server, e.g., large-scale databases. Future rack-scale systems will have increased flexibility as virtual machines can be created by combining, e.g., CPUs from one server with memory from a second and networking from a third. However, how to best allocate hardware resources to applications in such environments is more challenging due to this increased flexibility.
The goal of this project is to create an autonomous system for resource allocation is rack-scale systems. The overall challenge is how much capacity to allocate to each virtual machine at each point in time, and in which server(s) to allocate this capacity. A solution to this dual optimization problem should maximize both application performance and overall system utilization. Research methods that could be used to solve this problem include application performance analysis, optimization to solve scheduling and matching problems during resource allocation, as well as feedback control to make the system robust to disturbances due to changes application performance requirements and/or interference between collocated applications.
The project is part of a major effort including eight postdoctoral stipends in autonomous systems. For more information, see the following link.
Each stipend is for two years with a starting date to be negotiated. The stipend, provided by the Kempe Foundations, amounts to 300 000 SEK per year. The stipend is not subject to tax.
A qualified applicant is required to have a PhD degree or a foreign degree that is deemed equivalent to a PhD in Computer Science, Computer Engineering, or another subject of relevance for the project. The PhD degree shall not be more than three years old by the application deadline unless there are special reasons. The applicant should be strongly motivated and interested to develop new competencies, as well as to act in an international environment.
Documented knowledge and proven research experiences in at least one of the fields machine learning, operating systems, or cloud resource management systems is required. Good research merits and scientific publications in the area of the position are strongly meriting. International research experience is also a merit.
A successful candidate should be capable of performing practical implementations of new algorithms, as well as producing scientific publications in English. Very good knowledge in the English language, both spoken and written, is required.
The application, preferably written in English, should include:
- An introductory letter summarizing your qualifications, research interests, and motivation for applying (max 2 pages)
- CV with list of publications
- Copies of PhD thesis and relevant publications
- Copies of degree certificates
- Name and contact information for two or three reference persons
Umeå university is an equal opportunity employer, therefore, we welcome female applicants in particular.
Further information is provided by Assoc. Prof. Johan Tordsson, firstname.lastname@example.org.
Your complete application, marked FS 2.1.6-2286-16, should be sent electronically (in pdf-format) to email@example.com (provide the reference number on the email’s subject line). Application deadline is 2017-02-15.
We look forward to your application!