Ttulo do Trabalho

1
Catholic University, PUCRS (Brazil) Faculty of
Informatics and Faculty of Engineering Embedded
Systems Research Group lthttp//www.inf.pucrs.br/gs
egt
CCSDS TC/TM Interface Targeting Reconfigurable
Logic by Eduardo Augusto Bezerra CTA/ITA,
São José dos Campos, 23-24 November 2004.
2
Summary

• Supporting Team GSE and Collaborators
• Motivation, Objectives and Contributions
• Design Decisions
• Design Methodology and Strategy
• Drawbacks
• Conclusions
• Contact Details

3
Supporting Team The GSE Group
Application Area Emphasis embedded software and
hardware for sensors and dedicate computing
systems
Research Emphasis Embedded Systems
Space Systems
Mobility and Pervasive Computing
Precision Farming
4
Supporting Team The Space Science Centre
5
Supporting Team The GAPH Group
Application Area Emphasis Telecom Systems
Research Emphasis SoC
Modelling, design and validation at systemic
levels
Modular Systems Construction
IP Construction
Reconfigurable Hardware Construction
Hardware Virtualisation
IP cores
Reconfigurable Systems
6
Supporting Team Collaborators
The University of Sussex (UK) Space Science
Centre, EIT http//www.sussex.ac.uk/space-science
Faculty of Engineering, PUCRS Systems, Computing
and Communications Group http//www.ee.pucrs.br/s
isc
IPCT, PUCRS Microgravity Laboratory http//www.ipc
t.pucrs.br/microg
Faculty of Informatics, PUCRS Hardware Design Aid
Group http//www.inf.pucrs.br/gaph
7
Summary

• Supporting Team GSE and Collaborators
• Motivation, Objectives and Contributions
• Design Decisions
• Design Methodology and Strategy
• Drawbacks
• Conclusions
• Contact Details

8
Motivation, Objectives and Contributions

• Brazilian space program is interested in the use
  of FPGAs for the implementation of on-board
  processing systems
• Main objective Design and implementation of a
  CCSDS telecommand and telemetry system in an
  FPGA
• Investigation of the use of configurable
  computing (FPGAs) to substitute the traditional
  approach (microprocessors) employed in the design
  of on-board computers

9
Motivation, Objectives and Contributions

• Development of a configurable system to be used
  in space applications
• Investigation of high level languages for
  hardware description (e.g. Java, C)
• Investigation/definition of fault-tolerance
  strategies for configurable computers
• High performance with low clock and less power
  consumption

10
Motivation, Objectives and Contributions

• Flexibility of developing hardware in a similar
  way as SW (bug fixes, system upgrades), with no
  HW changes
• Microprocessors applications have to adapt to
  the architecture available. FPGAs architecture
  adapts to the application
• Flexibility to define as many independent memory
  modules (distributed memory) as necessary,
  reducing the von Neumann bus bottleneck
• Modern FPGAs have architectures suitable for
  arithmetic and DSP applications.

11
Motivation, Objectives and Contributions

• HW/SW co-design is too complex ? investigation of
  pure hardware solutions.

Modelling/System Description
(
SystemC
,
SDL, CSP, Java, ...)

Validation

(Co
-
Simulation,
Formal Verification)

HW/SW Partitioning

HW (HDL, occam)

SW (C, occam)


Co
-
Synthesis
Synthesis

Compilation


Software
Prototiping
components

Hardware
(Library)

components

FPGA
FPGA

FPGA

µ
P


(IP cores)



FPGA

analogic

memory

memory

SoC

12
Summary

• Supporting Team GSE and Collaborators
• Motivation, Objectives and Contributions
• Design Decisions
• Design Methodology and Strategy
• Drawbacks
• Conclusions
• Contact Details

13
Design Decisions The Simulation Environment
Prototyping platform selection target FPGA,
prototyping board, hardware description language,

14
Design Decisions The Description Language Issue

• The first activity in a project should be to
  identify the problems complexity.
• In complex (and large) projects, the solution
  will certainly involve some sort of systemic
  approach.
• In mid-size projects a behavioural HDL solution
  may be sufficient.
• In small projects a schematic entry or structural
  HDL solution may be a better option, but just in
  cases where designers have a hardware background.

15
Design Decisions Target Device Selection

• Manufacturers of military and/or rad-hard FPGA
  devices
• Actel Corp. lthttp//www.actel.comgt
• Altera Corp. lthttp//www.altera.comgt
• Atmel Corp. lthttp//www.atmel.comgt
• Chip Express lthttp//www.chipexpress.comgt
• Gatefield Corp. lthttp//www.gatefield.comgt
• Lucent Corp. lthttp//www.lucent.com/micro/fpgagt
• Mission Research Corp. lthttp//www.mrcabq.comgt
• Quicklogic Corp. lthttp//www.quicklogic.comgt
• Xilinx Corp. lthttp//www.xilinx.comgt

16
Summary

• Supporting Team GSE and Collaborators
• Motivation, Objectives and Contributions
• Design Decisions
• Design Methodology and Strategy
• Drawbacks
• Conclusions
• Contact Details

17
Design Methodology and Strategy

• CCSDS
• Consultative Committee for Space Data Systems
• Employed in a large number of scientific and
  commercial spacecraft.
• Allows reduced costs for on-board, ground and
  test equipment, as well as for spacecraft testing
  in in-orbit operation.
• Standard space industry communication protocol.

Telecommand Transmit
Telecommand Receive
Frame Processing
Frame Processing
Telemetry Receive
Telemetry Transmit
R-S Encoder
R-S Decoder
Space Segment
Ground Segment
18
Design Methodology and Strategy
FPGA
AMBA bus
Leon 2 Sparc V8 microprocessor
HurriCANe CAN interface
First step A soft core microprocessor (Leon 2)
runs a minimal CCSDS TC/TM implementation. The
system uses the AMBA bus for cores
interconnection, and a CAN core for connecting
external modules.
19
Design Methodology and Strategy
FPGA
?
?
?
AMBA bus
Leon 2 Sparc V8 microprocessor
HurriCANe CAN interface
End objective Selection of some modules of the
protocol suitable for HW implementation. The
modules should be designed as peripherals for the
Leon microprocessor.
20
Design Methodology and Strategy
21
Design Methodology and Strategy
22
Design Methodology and Strategy
23
Design Methodology and Strategy
Dependability Improvement Signature
Analysis-Driven Refresh Without FPGA Replication
24
Design Methodology and Strategy
Dependability Improvement Masking Connectivity
Faults
25
Summary

• Supporting Team GSE and Collaborators
• Motivation, Objectives and Contributions
• Design Decisions
• Design Methodology and Strategy
• Drawbacks
• Conclusions
• Contact Details

26
Drawbacks

• Synthesis tools for high level languages are
  still inefficient
• VHDL is device independent, but it is not
  synthesis tool independent
• For good results, the designer is required to
  have a good knowledge of the FPGA architecture,
  synthesis tools functionality, and how to program
  in HDL for synthesis
• An acceptable VHDL description, suitable for
  synthesis, is in a level of abstraction lower
  then Assembly language, since the designer has to
  think in terms of flip-flops and multiplexers
• FPGAs are still expensive
• Skillful people in this field is hard to find.

27
Summary

• Supporting Team GSE and Collaborators
• Motivation, Objectives and Contributions
• Design Decisions
• Design Methodology and Strategy
• Drawbacks
• Conclusions
• Contact Details

28
Conclusions

• The proposed project is feasible, but the right
  personal must be found
• A close contact with Surrey and Sussex Space
  groups is important for the project success
• The whole CCSDS TM/TC interface is not proper
  for a pure hardware implementation
• In a first moment, the whole system will be
  prototyped in an FPGA as a pure software
  implementation for a soft core RISC
  microprocessor
• Gradually, modules of the system will be
  converted into hardware as IP cores.

29
Summary

• Supporting Team GSE and Collaborators
• Motivation, Objectives and Contributions
• Design Decisions
• Design Methodology and Strategy
• Drawbacks
• Conclusions
• Contact Details

30
Contact Details
Eduardo Augusto Bezerra Embedded Systems Research
Group (GSE) Faculdade de Informática, PUCRS Av.
Ipiranga, 6681, Porto Alegre, RS, Brazil Email
bezerra_at_inf.pucrs.br Web http//www.inf.pucrs.br/
eduardob Web http//www.inf.pucrs.br/gse Tel.
55 (51) 3320 3611 or 55 (51) 3320 3695
31
Useful Links
CCSDS http//www.ccsds.org Sussex University
http//www.sussex.ac.uk/space-science Surrey
University http//www.ee.surrey.ac.uk/SSC/ SSTL
http//www.sstl.co.uk/ SatServ
http//www.satserv.nl ESA http//www.esa.int Leon
http//www.gaisler.com/products/leon2/leon.html
GSE http//www.inf.pucrs.br/gse