Further information can be found in the course description and the current Moodle course.
After this course students will have an understanding of networks beyond basic principles with the focus on the specific challenges, approaches and technologies of storage and data-centre networks. In order to achieve this the deployed protocols and topologies in the field of storage networks (e.g. iSCSI, SAN, iSER) and for the realization of large computing facilities and data centers (e.g. Infiniband) are covered. Furthermore the basic design principles and architectures for large scale data centers, the available technologies and procedures (e.g. benchmarking, performance and power efficiency indicators, ...) to optimize such architectures are covered. The primary teaching goals are to allow course participants to name and describe the technologies and protocols used for storage networks and in data centers. Participants will understand the pros and cons of different approach and learn how to evaluate their applicability in different context and examples. The participants will learn to apply these technologies and background knowledge to design data centre networks and architectures for different applications (e.g. Big Data Analytics, Scientific Computing, Server Farming, Cloud Computing, ...) As background material mostly protocol specification and primary literature and publications are used. The exams repeat the topics from the lecture with a more practical viewpoint and validate the knowledge based on exercises. The exercises will enable the participants to solve smaller design challenges for storage networks and will evolve over the course up to design of complete data centre systems using the technology introduced in the course for different application domains.
The area of data centre and storage networks is characterized by a large focus on optimization for high speed and bandwidth as well as extreme requirements for low latencies. In particular for very large data centers in the field of Big Data analytics, scientific computing/simulation or server farming these requirements combined with the boundary conditions of cost optimizations and performance/cost balance are very challenging. The design of such systems does not only demand for selecting the most appropriate protocols but similarly asks for creativity in selecting metrics and benchmarks, topologies and architecture. In this course mature and emerging technologies in the field of storage and data centre networks are presented. The capabilities of the different technologies will be discussed along specific use cases e.g. for realizing high performance parallel filesystems, data centers to support large scale data analytics or compute intensive simulation applications from different fields such as automotive or chemistry.