Using Network Calculus to compute end-to-end delays in SpaceWire networks

Abstract

The SpaceWire network standard is promoted by the ESA and is scheduled to be used as the sole on-board network for future satellites. This network uses a wormhole routing mechanism that can lead to packet blocking in routers and consequently to variable end-to-end delays. As the network will be shared by real-time and non realtime traffic, network designers require a tool to check that temporal constraints are verified for all the critical messages.Network Calculus can be used for evaluating worstcase end-to-end delays. However, we first have to model SpaceWire components through the definition of service curves. In this paper, we propose a new Network Calculus element that we call the Wormhole Section. This element allows us to better model a wormhole network than the usual multiplexer and demultiplexer elements used in the context of usual Store-and-Forward networks.

NoC contention analysis using a branch-and-prune algorithm

“…But Ferrandiz et al [2011] clearly showed with an illustrative example the flaws of the design and that such a modeling is pessimistic, leading to overdimensioning of resources. Ferrandiz et al [2011] consider a space wire network topology and introduced a special network element called the "wormhole section" to describe the wormhole routing with the network-calculus terminology.…”

“…But Ferrandiz et al [2011] clearly showed with an illustrative example the flaws of the design and that such a modeling is pessimistic, leading to overdimensioning of resources. Ferrandiz et al [2011] consider a space wire network topology and introduced a special network element called the "wormhole section" to describe the wormhole routing with the network-calculus terminology. This element envelopes a set of routers lying in the shared path between an analyzed flow and the blocking flow(s): the analyzed and the blocking flows enter the first router and exit through the last router of the wormhole section, with no additional blocking flows either entering or leaving from any other link within the wormhole section.…”

EDF as an arbitration policy for wormhole-switched priority-preemptive NoCs

“…Several techniques have been proposed to provide an upper-bound on the worst-case delays of individual traffic flows [5,6,7,15], however, due to complex interference patterns, these methods yield pessimistic results [5].…”

“…Assuming that a packet is released at the critical instant t ∈ T , its worst-case traversal time Li(t) (expressed in absolute values) is computed by solving Equation 7.…”

Worst-case memory traffic analysis for many-cores using a limited migrative model

“…Several methods have been proposed to obtain upper-bounds on worst-case latencies ( [8], [10], [11]), but they yield either pessimistic results [10], [11], or have complexity and scalability issues [8]. Conversely, if a platform contains multiple virtual channels ( [6], [7]), the benefits are twofold: (i) by avoiding idle routers, the performance of the wormhole switching technique is significantly improved ( [6], [7]) and (ii) preemptions can be implemented [24].…”