distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfmika...
TRANSCRIPT
![Page 1: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/1.jpg)
Mika Karaila
Metso Automation / PAS RTD
Distributed real-time systemarchitecture:layers & patterns in real-life
TTY 28.3.2012
![Page 2: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/2.jpg)
© Metso
metsoDNA architecture
•Background / history•General structure•Some cases (layers & patterns)•Motivation•Summary
![Page 3: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/3.jpg)
© Metso3
Compatibility with Innovative Evolution
Connectability•All existing systems can beconnected to extensionsUpgradeability•New features can be upgradedto existing systemsOpenness•Open to all solutions
Integration of•MCS, DCS, QCS, Drive and Infoapplications one system•Field & embedded solutions•Knowledge and information•User friendly, powerful, flexible andreliable platform for advanced pulp& paper andprocess & energy solutions
1979 2006
Damatic”Classic”
1988
Damatic XDDamatic XDi
metsoDNA CR
2000
100 Mb
10 Mb250 kb 2 Mb
metsoDNA
![Page 4: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/4.jpg)
© Metso
metsoDNA CR Architecture
4
Star or ring topologyredundant Ethernet network
Centralized ordistributed
Operation, Maintenance, Reporting
Engineering &Maintenance,AssetManagement,Informationservices
Reporting, Enterprise integrationRemote connections
Controls,Optimization,ConnectivityandSafetyinstrumentedsystem
Automation roomAutomation roomAutomation room
Control roomControl roomControl room
FieldFieldField
OfficeOfficeOffice
Name based communication
Ethernet andserial links
No need of internal communication linksNo need of internal servers
![Page 5: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/5.jpg)
© Metso
One Platform for All Needs
• Same platform for different types of controls• Process, machine, quality, drive controls and
optimizations• Asset management solutions
- Field device and vibration monitoring
• Integrated Safety related controls• Benefits
- One user interface and alarm handling- No links between systems- One history data collection and trending- Common engineering tools- Spare parts and trainings
5
Quality controlsand optimizations
Process controls
Drive controls
Machine controls
Integrated SafetyInstrumentedSystem
- MachineConditionMonitoring-Field deviceassetmanagement
![Page 6: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/6.jpg)
© Metso
Scalable metsoDNA from small to hugePowerful Architecture
6
Large mill network• 17 system• Different versions• Different system
generations• ~ 100 000 IO’s
Tiny embedded application• 10 - 250 IO’s• typically installed to field or
embedded to machines
Damatic Classic, XD, XDi andmetsoDNA-networks
Mill Ethernet
autom. Ethernet
metsoDNA-network
KA1
KA2
KA4
metsoDNA-network
KK2
ST2
SK5
Router/Switch
Damatic XD
PK6
metsoDNA-network
Damatic XD
PK8
PASTA
CTMP
KA5
Damatic CL + XD
Damatic XD
Damatic XD
KU1
KRO
SK6
VPU
Damatic CL
Damatic XD
Damatic XD
Damatic XD/CL
metsoDNA-network
metsoDNA-network
Damatic XD
TAKE
CIS
CIS
CIS
INFO
F
N
C
B
M
A
L D
D
E
K
G
G
Mill 1 H
Mill network which connects Damatic XD- ja metsoDNA networks
Damatic-XD
INFOCIS
VL1
HO5
KL1
Damatic CL
Damatic CL
Damatic CL
Mill 2
Mill 3
J
O
Damatic CL
TLA
Damatic CL/XD
C
Benefits• One system grows according
the needs• Seamless communication• For all types of controls• Asset management
![Page 7: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/7.jpg)
© Metso
metsoDNA NetworksTopology / design
MAY 8, 2009 / MY
7
Recommended topology: Ring topology network ( Turbo Ring)• Metso brand labelled switches (Moxa).• Simple structure.• Expected longer switch life span•Can be expanded to cover large and very distributed mill’s (sub-rings).• One switch model covers basically whole structure.• Use for new mills
• also where <300ms recovery times needed forredundancy.
RSTP (tree-like) solution• Cisco, Hewlett Packard and Metso (Moxa brand labeled)switches.• Three different level of configuration templates exists forsupported devices.•Can be expanded to cover large and very distributed mill’s
•need for high density fiber switches (rare - no options -expensive)•need to use couple of different switch models mill wide.
• Existing RSTP solutions should be primarily expanded usingRSTP
Fibers
4U2U1U
2U2U2U1U1U1U
2U1U
2U2U2U1U1U1U
1U1U1U1U
2U1U
2U2U1U1U
AKKUAC/DC-POWERSUPPLIES(2) 4U
2U1U
2U2U2U1U1U1U
2U1U
2U2U2U1U1U1U
1U1U1U1U
2U1U
2U2U1U1U
AKKUAC/DC-POWERSUPPLIES(2)
4U2U1U
2U2U2U1U1U1U
2U1U
2U2U2U1U1U1U
1U1U1U1U
2U1U
2U2U1U1U
AKKUAC/DC-POWERSUPPLIES(2) 4U
2U1U
2U2U2U1U1U1U
2U1U
2U2U2U1U1U1U
1U1U1U1U
2U1U
2U2U1U1U
AKKUAC/DC-POWERSUPPLIES(2) 4U
2U1U
2U2U2U1U1U1U
2U1U
2U2U2U1U1U1U
1U1U1U1U
2U1U
2U2U1U1U
AKKUAC/DC-POWERSUPPLIES(2)
![Page 8: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/8.jpg)
© Metso
Redundancy patterns
• Redundant functionality• Redundancy switch• Centralized synchronization
8
![Page 9: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/9.jpg)
© Metso
Hardware changes: CPU models
• Damatic Classic- Zilog Z80 based 8-bit CPU
• Damatic XD- Motorola 68k based 16-bit CPU
• metsoDNA- Intel x86, powerPC, VIA, Motorola (all still supported)
9
![Page 10: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/10.jpg)
© Metso
Operating systems
• Operating systems:- Embedded kernel- HP-UX- Sco Xenix- Sco Unix- Windows NT, XP, Vista, Windows7- Linux
10
![Page 11: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/11.jpg)
© Metso
Operating system patterns
• Hardware abstraction layer (HAL)• Device proxy• Concurrent execution• Operating system abstraction
11
![Page 12: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/12.jpg)
© Metso
Applications implemented by Visual Language
12
![Page 13: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/13.jpg)
© Metso
Application execution
• Applications from late 1980 still possible to run...
• Operating system:- Hardware abstraction layer (BSP: Board Specific Porting, see Linux examples)- Operating system abstraction (Windows / Linux, POSIX interface)
• Applications:- Visual language compiled to byte code (Domain Specific Visual Language)- Virtual machine runs byte code (like Java virtual machine) for Function Block
applications- Subroutines encapsulated into Function Blocks
13
![Page 14: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/14.jpg)
© Metso
Real-time separation patterns
• Separate real-time• Early work• Operator profile• Opportunistic delegation• Third-Party confinement• Static resource allocation• Static scheduling• Limp home• Isolate control algorithm
14
![Page 15: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/15.jpg)
© Metso
Safety-oriented patterns
• Safe state• Next stable state• Distributed safety• Heart beat• Watchdog
15
![Page 16: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/16.jpg)
© Metso
Field bus Solutions, ACN node
16
PROFIBUS PA
max. 30 units
max. 125 units
ACTUATOR- SENSOR- INTERFACE
max. 62 units
PROFIBUS DP
max. 12 units
FOUNDATION FIELDBUS
H1
PROFIBUS DP max 12 Mbits/s(single / fiber ring)fiber 2 km, twisted pair max 1 km
PROFIBUS PA 31.25 kbits/stwisted pair 1900 mpower and data via same cable
AS-i156 kbit/stwisted pair 100 mpower and data via same cable(max. 62 devices in v2.1max. 31 devices in v.2.0)
Foundation FieldbusHSE 100 Mbits/stwisted pair 100 m /fiber 2 kmH1 31.25 kbits/stwisted pair 1900 mpower and data via same cable
Three PROFIBUS-DP buses can be connected to one ACN PCS. SeveralDP/PA couplers and AS-i gateways can be connected to PROFIBUS-DP. Alsoseveral FF Linking Devices with four H1 segments can be connected to ACNPCS.
![Page 17: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/17.jpg)
© Metso
Communication
• Several levels:- Devices connected to I/O-units- I/O-units to rack controllers- Rack controllers via fieldbus to CPU- CPUs are connected to system bus- One system to own ”segment”
• Fieldbus protocols:- Profibus DP/PA- Foundation fieldbus HSE/H1- Ethernet/IP, EtherCAT
17
![Page 18: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/18.jpg)
© Metso
Communication principles
• Deterministic• Robust• Diagnostic data coming more important
• Master / slave• Boardcast• Heart beat• Watch dog (fail safe function)
18
![Page 19: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/19.jpg)
© Metso
Messaging patterns
• Isolate functionalities(subsystems)
• Bus abstraction (change, noactual location)
• Message queue (acyclic,performance)
• One at a time (master / slaveprinciple)
• Priorized messages(alarms/events,...., background)
• Early warning• Converting message filter• Distributed transaction• Message channel selection• Vector clock• Permission request• Unique confirmation• Locker key
19
![Page 20: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/20.jpg)
© Metso
metsoDNA CR – User Interaction Activity20
Firewall
Automation roomAutomation roomAutomation room
Star or ring topologyredundant Ethernet network
Control roomControl roomControl room
Centralized ordistributed
FieldFieldField
OfficeOfficeOfficeOperation, Maintenance, Reporting
Engineering &Maintenance,Asset Management,Information Services
Reporting, Enterprise integration
andSafetyinstrumentedsystem
Remote connections
Controls,Optimization,Connectivity
XML
![Page 21: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/21.jpg)
© Metso
• For collected tags- History replay from any
picture- Drag ’n Drop to analysis
tools
21
Switch to history modeHistory mode indication
Select time in history and replay
DNAuse – Efficient Problem SolvingTime machine feature for any picture
![Page 22: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/22.jpg)
© Metso
• Analysis Tools interact withDNAuse- Fast analysis by one click- Predefined parameters are
passed between tools
• USE CASE DNAtracer:- Open tool from Action Menu- Loop tag name automatically as
parameter- Trends and related events
combined automatically- Further analysis from event row
in DNAtracer• Alarms and Events
analysis• Diary entries
22
Deeper alarms &events analysisException
entry usingDNAdiary
Trends with alarms & events
One Click
Integration Increases Awareness of SituationOptional Analysis Tools with DNAuse
![Page 23: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/23.jpg)
© Metso
DNAals - Event Browser Filtering Feature
23
Helps focusing on situation at hand – Refiners area
Updating value
Alarm limits
![Page 24: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/24.jpg)
© Metso
DNAuse Replay – Alarm & Events Included
24
Something wrong with the Feed Pump?
DNAmachineAssessorknows history mode too!
![Page 25: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/25.jpg)
© Metso
DNAhelp – Functional DescriptionsHow the Feed Pump should work?
Online usercomments possible!
![Page 26: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/26.jpg)
© Metso
DNAdiary – Disturbance Entry
26
Report faulty Feed Pump and send notification to CMMS
![Page 27: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/27.jpg)
© Metso
Key Strengths of User Interaction
• Industrial design with high usability comes from the experience- Application of systematic usability methods and usability field tests- Hundreds of studies and thousands of deliveries through 20 years
• Integration of real-time and history data- Exactly the same user interface is able to use real-time and history information- No need to do the system setup, training or maintenance twice
• Tools for analyzing, reporting and log book integrated- All tools can be opened from Operator Interface with tag as a parameters- Drag n’ drop from pictures to trends
• Exactly the same tools for Control Room and Office users- Everyone is using the same system, one version of the truth
27
Focus areas in User Interfaces
![Page 28: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/28.jpg)
© Metso
Key Features to Boost Daily OperationsTransparently integrated history functions
• DNAuse TEA brings time machine to the mill or plant
• DNAuse with History Mode and Replay- Ultimate troubleshooting tool - metsoDNA CR has a memory!- Replay can be set up completely and fast by using Metso standard
• My Community Tools for Analyzing and Reporting- Power full trending and reporting tools to Control Room
• Replace alarm printers with DNAreport AE• Replace manual log book with DNAdiary• Replace short DCS trends with DNAtracer
• Bottom line – today a production organization cannot survive globalcompetition without help of history functions in every day operations
28
![Page 29: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/29.jpg)
© Metso
User Interaction is Easy to Set up and MaintainEfficient engineering tools and install free user tools
• Setting up data collection with familiar tools- No need for OPC gateways, interfaces to Metso systems are in-build- Configuration is done using regular metsoDNA engineering tools
(DNAexplorer)
• Web based My Community tools are installation free- Longer life cycle since the operating system is not a critical issue- If client software update is needed, only Server side needs to be updated.- No need to select dedicated end users – only web browser is needed
• Non-web clients – DNAtracer and DNAview are Click Once -applications- Installed upon first usage and updated automatically- Firm look and feel of native windows application
29
![Page 30: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/30.jpg)
© Metso
User interfaces summary
• Operating system neutral• Both Windows & Linux supported (QT)
• Common communication for all components- Data from real-time environment- Data from different databases
• Feature rich client applications• Interaction between applications
30
![Page 31: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/31.jpg)
© Metso
Motivation
• C / C++ used as programming language in embedded environments• Java in some optimization / web applications• C# is coming more common in UI / web applications• Security coming very important• Concurrent / parallel execution due multi-core CPUs
- Thread Building Blocks (TBB ), http://threadingbuildingblocks.org/- MPAPI, http://mpapi.codeplex.com/
• Continuous Integration, CI- Cruise Control, http://cruisecontrol.sourceforge.net/- Hudson -> Jenkins, http://jenkins-ci.org/
• Automated TestingSoftware production methods needed for efficient & high quality
systems.
31
![Page 32: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/32.jpg)
© Metso
Summary
• Principles done in 1980 are still valid
• It is easier to use them as they are commonly ”known” solutions
• Some parts of implementations can be hidden, compilers can domore
- More libraries are supporting abstraction like OpenMP, QT
32
![Page 33: Distributed real-time system architecture: layers ...ohar/luennot/luennot2011/ohar_karaila.pdfMika Karaila Metso Automation / PAS RTD Distributed real-time system architecture: layers](https://reader033.vdocuments.mx/reader033/viewer/2022052710/5aad955f7f8b9a2b4c8eb85f/html5/thumbnails/33.jpg)
© Metso