Page 1
Standards
Certification
Education & Training
Publishing
Conferences & Exhibits
DNP3 Implementation
The Do’s and Don’ts
2014 ISA Water / Wastewater and Automatic Controls SymposiumAugust 5-7, 2014 – Orlando, Florida, USA
Speaker:Jeff M. Miller, P.E., ENV SP
Page 2
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA 2
Presenter
• Jeff M. Miller, P.E., ENV SP– Jeff is a Water Solutions Architect for Schneider Electric’s
Water Wastewater Competency Center.
– Jeff has a B.S. in Electrical Engineering and has workedas an engineering consultant and a control systemsintegrator in the water industry for 24 years.
Page 3
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA 3
Presentation Outline
• Understanding DNP3
• Planning DNP3
• Executing DNP3
• Summary
How come youdon’t have
comfortableseats like ours?
Page 4
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
DNP3
UNDERSTANDING
Page 5
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3 History
• DNP3 : Distributed Network Protocol– Originally developed by Westronic (now GE) in 1990
– Owned and governed by DNP3 Users Group since 1993
– Widely accepted advanced communications protocol that is…
– A robust and secure protocol over unreliable communications media
– Open to allow interoperability between many different vendorhardware and software solutions
– Efficient for data and application transmission on many types ofmedia and communication frameworks
– Richly featured with built in capabilities preventing the customcoding or configuration needed with other protocols
– A live standard that evolves as technology changes but still allowsbackward compatibility
5
Page 6
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Informative References
• www.dnp.org ~ DNP3 Users Group
• www.TriangleMicroWorks.com ~ DNP3 Training
6
Page 7
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3Traditional Polling Protocol Architecture
• Master controlscommunications andprevents collisions
• Data Concentrators may beneeded for repeaters andpossibly for the Master
• Repeaters avoided
• No advanced featureswithout custom programming
– Current data only
– Inefficient, all datatransferred with possibleseparate transmissionsfor different data types
7
DataConcentrator
SlaveSlave
Slave
Slave
Slave
Slave
Slave
MasterRadio
Master DataConcentrator
SCADA
– No time stamping
– Little verification ofactions or data quality
Page 8
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3 ConceptsTerms
• Masters and Outstations
• Static and Event Data
• Routing vs. Data Concentrator
• Reporting ~ one or a combination of the following:– Polling
– Unsolicited
– Unsolicited Report By Exception
– 2-Pass Control
• Classes
• Implementation Levels
8
Page 9
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3 ConceptsKey Concepts
• Static Data = present value
• Event Data = triggered data
• DNP3 Outstations can buffer event data and backfillbetween communication connections
• DNP3 typically has a larger message size– Can transfer multiple data types in single message
– Additional point and message data for quality, security, andconfirmation
• DNP3 includes time stamped data and automatic timesyncing
• DNP3 Masters and Outstations have communicationdatabases that need to be configured
9
Page 10
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3Store and Forward
10
100%
0%
50%
TIME
100%
0%
50%
TIME
100%
0%
50%
TIME
Page 11
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3Polling
• Possible long delaysbetween updates but data is“backfilled’
• Prevents collisions if onlymethod used
• Different data classes maybe polled at different rates
• Infrastructure avoidsrepeaters in architecture
11
DataConcentrator
OutstationOutstation
Outstation
Outstation
Outstation
Outstation
Outstation
MasterRadio
SCADAMaster
Page 12
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3Routing Without a Data Concentrator
• Known as Message Pass-Through or Routing
• Data concentrator may stillbe needed for outstation-to-outstation control
12
Outstation
OutstationOutstation
Outstation
Outstation
Outstation
Outstation
Outstation
MasterRadio
SCADAMaster
ROUTING
Page 13
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3Unsolicited
• Unsolicited messages sentfrom Outstation by time and /or events
• Very efficient but collisionsmay occur on simultaneouscommunications
• Classes determineresponses
• Data may be sent out on astate change or when adeadband is exceeded alongwith a hold time and countvalues to prevent spurioustransmissions
13
Outstation
OutstationOutstation
Outstation
Outstation
Outstation
Outstation
Outstation
MasterRadio
SCADAMaster
COLLISION
REPORTED BY EXCEPTION
Page 14
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3 ConceptsKey Concepts ~ Class
• Class 0: Polled Data– Typically infrequent polls (e.g. daily)
– Typical Data
– Integrity
– Transfer of non crucial stored data
– Run times, etc.
• Classes 1 → 3: Event Data– User configurable by event and response time
– Typically:
– Class 1 immediate unsolicited response (e.g. alarms, controlactions)
– Class 2 delayed unsolicited response (e.g. monitored values)
– Class 3 further delayed unsolicited response (e.g. collected values)
14
CLASSES NOT CLASSROOM
Page 15
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP32-Pass Control
• Simple built-in handshake to verify control action andresponse
15
Outstation
MasterRadio
SCADAMaster
OFF ON FAIL
Page 16
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Understanding DNP3 ConceptsImplementation Levels
• Level 1: Typical of simple IED (Intelligent Electronic Device)
– Simple reads and writes, no analog time stamping
– Possible simple unsolicited messages
• Level 2: Typical of small RTU– Supports Level 1 plus
– more types of messages
• Level 3: Typical of large RTU or Data Concentrator– Supports Level 1 & 2 plus
– Fully configurable unsolicited messages by class
• Level 4: Advanced “Smart” RTU– Supports Level 1, 2 & 3 plus
– Additional communication and network options
16
Page 17
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Why is DNP3 not used all the timeon Water / Wastewater Implementations?
If I answerthis correctlywill I get free
stuff?
With all theseadvantages I
don’t seewhy it isn’t?
Page 18
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Roadblocks to DNP3 Usage inthe Water / Wastewater Industry
• Started in energy industries
• Mindset of old technologies and designmethodologies
• Not connecting the advanced featureswith application advantages
• Unfamiliar or lack of experience
• Misapplication ~ configuring the old way
• Understanding update rates versesresponse time
• Poor configurations lead to poorperformance and give DNP3 a badreputation
18
Page 19
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
DNP3
PLANNING
Page 20
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Do’s Don’ts
20
Planning DNP3
• Leverage features to lowersystem infrastructure costs
– Lower antenna heights
– Use of repeaters
• Leverage features to help inthe choosing of communicationmediums
– Fiber, Radio, Ethernet, Cell,Satellite, land line, etc.
• Consider the classes of data indesign
• Understand update ratesverses response times
• Design like a traditional staticdata system
• Dictate a fast response time forall data
• Classify all data as crucial
Page 21
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
DNP3
EXECUTING
Page 22
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Do’s Don’ts
22
Executing DNP3
• Plan your data reporting andtake advantages of classes
• Keep network traffic low toavoid collisions
• Leverage features to lowerdata costs on cell or satellitelinks
• Train users to understand thedifference between dataupdate rates and responsetime
• Minimal configuration mayprovide minimal performance
• Assume effort is like otherprotocols you have usedbefore
• Implementation withoutplanning
• Configure for the fastest dataupdate rate possible
Page 23
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Executing DNP3Don’ts ~ Common Configuration Errors
• Too many critical class data points
• Unnecessary data transmitted. Configure only what is needed.
• Hold Time & Hold Count values poorly set– Can cause too much network traffic or can cause too much latency
• Inappropriate or missing deadband & debounce values
• Requesting Confirmation messages at both Application & Data Linklayers is redundant & inefficient
• Confirmation is not required in a Master– Master Writes already require confirmation
– Master Requests require a reply anyway. If no response, retries automaticallyoccur
• Data Link confirmation may be used on very noisy networks, or setFragment Size below 249 bytes and enable Application Layerconfirmation
23
Page 24
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Executing DNP3Don’ts ~ Common Configuration Errors
• Data object type buffer sizes improperly set– Too large – wastes limited event buffer space
– Too small – lost events due to buffer overflow
• Multiple Masters can cause poor communications – events will notbe removed from Outstation buffers until all Masters receive
• Poor choice of Network mode – can cause low network efficiency orlatency issues
• Incorrect choice of Operating mode – know when to use Routing(message pass through) or a Data Concentrator
• Understand the implications of choosing Ethernet TCP verses UDP– Ethernet TCP has additional overhead
– Ethernet UDP’s lack of message validation is covered by DNP3 protocol features
24
Page 25
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
DNP3
SUMMARY
Page 26
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Wisdom and Knowledge
• DNP3 is an advanced protocolthat takes knowledge, planning,and effort to implementsuccessfully
• The cost savings andperformance advantages of DNP3need to be accounted for early inthe design of the network in orderto take advantage of them
• Data must be organized by itsneed and cruciality
Page 27
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Questions?
Jeff M. Miller, PE, ENV SP
Solutions Architect
Water Wastewater Competency Center
8001 Knightdale Blvd.Knightdale, NC 27545-9023Office: 919.266.8360 | Mobile: 919.824.9114JeffM.Miller@Schneider-Electric.comwww.schnedier-electric.com
Page 28
2014 ISA WWAC SymposiumAug 5-7, 2014 – Orlando, Florida, USA
Bibliographical References
1. "Home - DNP.org - Distributed Network Protocol." Home - DNP.org - DistributedNetwork Protocol. DNP Users Group, n.d. Web. 02 May 2014.
2. Curtis/Woodland Engineering, Ken. A DNP3 Protocol Primer. Calgary, AB, Canada:DNP Users Group, 20 Mar. 2005. PDF.
3. DNP3 Overview. Raleigh, NC, USA: Triangle Microworks, Inc., 22 Feb. 2002. PDF.
4. Aubin, Philip, PEng. DNP3 Communications for Water and Wastewater Applications.Kanata, Ontario, Canada: Schneider Electric, 28 Aug. 2010. PDF.
28