Benchmarking SpaceWire Networks Asaf Baron, Isask'har Walter, Ran Ginosar, Isaac Keslassy EE Department, Technion, Haifa, Israel Ofer Lapid Israel Ministry.

Benchmarking Benchmarking SpaceWire NetworksSpaceWire Networks

Asaf Baron, Isask'har Walter, Ran Ginosar, Isaac Keslassy

EE Department, Technion, Haifa, Israel

Ofer Lapid

Israel Ministry of Defense

International SpaceWire Conference 200717-19 September 2007

The problemThe problem Multiple proposals for SpaceWire Vendors design new SpaceWire parts

BUT …

How can we evaluate proposals? How can we compare parts from

different vendors?

BenchmarkBenchmark

Benchmark purposesBenchmark purposes Evaluate protocol proposals Compare network configurations

Cost Performance Reliability

Compare components Formalize requirements Improve the standard

Benchmark ContentsBenchmark Contents End units

Redundancy Traffic specifications

Unit X sends to unit Y Quality of Service Requirements

Bandwidth, real time constraints, deadlines, Priorities, etc.

Benchmark does NOT include implementation details No topology, no specific components

End units (examples)End units (examples)

Unit#

Units

# Active Units

Data Handling System (DHS) 2 1

Reconfiguration Unit (RU) 1 1

Solar Cells Control 2 1

Payload Sensor (Camera) 3 1

Downlink 2 1

Star Tracker 3 3

Traffic matrixTraffic matrix

Traffic matrix (examples)Traffic matrix (examples) High bandwidth

From To Type BWB/sec

Latency Recommended

PrioritySensor Storage Payload 100M L

Low bandwidth

DHS Sensor Control 1000 Low H

Using Using the the

BenchmarkBenchmark

Evaluating SpaceWire featuresEvaluating SpaceWire features Standard: Packet-level priority (PLP) Non-Standard: N-Char interleaving

(NCI)

N-Char InterleavingN-Char Interleaving Not in SpaceWire standard But, Very useful in Wormhole

networks Requires multiple VCs VC code added to each N-Char

Ideally, SPW would allow longer flits

Delay

No-NCI ExampleNo-NCI Example 20181614121086428078767472706866646260585654525048464442403836343230282624220

Red:

Yellow:

Green: 69

Router #1

Router #2

400 Mb/s

100 Mb/s

400 Mb/s

48

26

Delay

NCI Example (4 VC)NCI Example (4 VC)2421.619.216.814.4129.67.24.82.49693.691.288.886.48481.679.276.874.47269.667.264.862.46057.655.252.850.44845.643.240.838.43633.631.228.826.40

Red (26):

Yellow (48):

Green (69):

31

38

45

Router #1

Router #2

400 Mb/s

100 Mb/s

400 Mb/s

Network topologyNetwork topology

DHS

storagesensor

downlink

router

Only active units shown and simulatedRedundant units, routers, links hidden and ignored

Simulator displaySimulator display Simulated using OPNET A commercial simulator for networks

Simulations: Low priority payload trafficSimulations: Low priority payload traffic

1. SIMPLE NET

2. ADD PLP

3. ADD NCI

4. NCI+PLP

5. NCI+PLP,more VCs

Simulations Results – Other Simulations Results – Other PrioritiesPriorities

High Priority Medium Priority

NCI with PLP is beneficial Low ETE delay of med+high priority packets High ETE delay of low priority packets

SummarySummary A new SpaceWire benchmark (open

source) A new simulator for SpaceWire An example of using the benchmark Priorities help Non-standard NCI helps even more