SDN for Dynamic Reservations on Real-Time Networks
Ref: CISTER-TR-200904 Publication Date: 21, Jan, 2020
SDN for Dynamic Reservations on Real-Time Networks
Ref: CISTER-TR-200904 Publication Date: 21, Jan, 2020Abstract:
Recent growing frameworks such as the IoT, IIot, Cloud/Fog/Edge computing, CPS, etc, bring the networking platforms on which they rely to the spotlight, as first class citizens of an increasingly software-dependent landscape. As a result, networks play an increasingly central role in supporting the needed system-wide properties. In particular, we have been working to provide openness and adaptivity together with timeliness guarantees. This combination seems fundamental to support inherently dynamic applications in a resource efficient way, covering not only the cases of systems of systems, systems with variable number of users, components or resources but also systems that undergo frequent live maintenance and even reconfiguration during their lifetime. Examples range from autonomous vehicles to collaborative robotics, remote interactions, fog/edge computing, flexible manufacturing, etc. We postulate that combining openess and adaptivity with guaranteed timeliness can only be achieved with an adequate communication abstraction supported on adequate protocols. To this end, we have been proposing channel reservation-based communication as a means to provide scalable and open latency-constrained communication and thus enable a more efficient system design. In this talk we will show our recent work in using Software-Defined Networking (SDN) to provide standard interfaces for traffic flexibility. We proposed extending the SDN OpenFlow protocol with adequate services to take advantage of flexible real-time communication protocols and thus provide standard interfaces for flexible real-time reservations, too. We call it the Real-Time OpenFlow framework (RTOF). We end describing and assessing a prototype implementation based on the HaRTES Ethernet switches.
Events:
Document:
Workshop on Next Generation Real-Time Embedded Systems (NG-RES 2020), Dagstuhl Publishing, Edited: Marko Bertogna and Federico Terraneo, pp 1:1-1:1.
Bologna, Italy.
DOI:10.4230/OASIcs.NG-RES.2020.1.
ISBN: 978-3-95977-136-8.
ISSN: 2190-6807.
Notes: Co-located with HiPEAC 2020.
Record Date: 29, Sep, 2020