Iec 62439 3 2016

A source DANP sen s the same frame over b th LANs an a destination DANP receives it from b th LANs within a certain time, con umes the first frame an dis ard the d pl cate.. Fig re 3 – P

Industrial communication network s – High avai abi ity automation network s –

Part 3: Paral el Redundancy Protocol (PRP) and High-avai abi ity Seamless

Redundancy (HSR)

disponibi ité –

THIS PUBLICATION IS COPYRIGHT PROT CTED

Copyright © 2 16 IEC, Ge e a, Switzerla d

Al rig ts reserv d Unles oth rwise sp cifie , n p rt of this p blc tio ma b re ro u e or uti ze in a y form

or b a y me ns,ele tro ic or me h nic l in lu in p oto o yin a d microfim, with ut p rmis io in writin from

eith r IEC or IEC's memb r Natio al Commite in th c u try of th re u ster If y u h v a y q estio s a o t IEC

c p rig t or h v a e q iry a o t o tainin a ditio al rig tsto this p blc tio , ple se c nta t th a dres b low or

y ur lo al IEC memb r Natio al Commite for furth r informatio

Droits d re ro u tio réserv s Sa f in ic tio c ntraire, a c n p rtie d c te p blc tio n p ut être re ro uite

ni uti sé so s q elq e forme q e c soit et p r a c n pro é é, éle tro iq e o mé a iq e, y c mpris la p oto o ie

et les microfims, sa s la c rd é rit d l EC o d Comité n tio al d l EC d p ys d d ma d ur Si v usa ez d s

q estio s sur le c p rig t d l EC o si v us d sirezo te ir d s droits su pléme taires sur c te p blc tio , uti sez

les c ord n é s ci-a rès o c nta tez le Comité n tio al d l EC d v tre p ysd résid n e

Th Intern tio al Ele trote h ic l Commis io (IEC) is th le din glo al org nizatio th t pre ares a d p blsh s

Intern tio al Sta d rds for al ele tric l ele tro ic a d relate te h olo ies

Ab ut IEC publ c tio s

Th te h ic l c nte t of IEC p blc tio s is k pt u d r c nsta t re iew b th IEC Ple se ma e sure th t y u h v th

latest e itio , a c rig n a or a ame dme t mig t h v b e p blsh d

IEC Catalog e - webstore.ie c / catalog e

Th sta d-alo e a plc tio for c nsultin th e tire

biblo ra hic l infor matio o IEC Inter natio al Sta d rds,

Te h ic l Sp cific tio s, Te h ic l Re orts a d oth r

d c me ts Av ia le for PC, Ma OS, An r oid Ta lets a d

iPa

IEC publc tio s s arc - w w.ie c /se rc pub

Th a v n e se rc e a les to fin IEC p blc tio s b a

v riety of crite a (r efer en e n mb r, te t, te h ic l

c mmite ,…) It also giv s informatio o pr oje ts, re la e

a d w ith r awn p blc tio s

IEC J st Publs ed - webstore.ie c / j stpubls ed

Sta u to d te o al n w IEC p blc tio s Just Pu lsh d

d tais al n w p blc tio s rele se Av ia le o ln a d

also o c a mo th b emai

Ele to edia - ww w.ele to edia.org

Th w or l 's le din o ln dictio ary of ele tro ic a d

ele tr i al terms c ntainin 2 0 0 ter ms a d d finitio s in

En lsh a d Fre c , w ith e uiv le t terms in 15 a ditio al

la g a es Also k now n as th Inter natio al Ele tr ote h ic l

Vo a ulary (IEV) o ln

IEC Glos ary - std.ie c / glos ary

6 0 0 ele trote h ic l ter min lo y e tr i s in En lsh a d

Fre c e tr acte from th Terms a d Definitio s cla se of

IEC p blc tio s is u d sin e 2 0 Some e tr i s h v b e

c le te fr om e r lier p blc tio s of IEC TC 3 , 7 , 8 a d

CIS R

IEC Cu tomer Serv ic Cente - webstore.ie c / cs

If y u w ish to giv us y our fe d a k o this p blc tio or

n e furth r as ista c ,ple se c nta t th Customer Ser vic

Ce tr e: csc@ie c

A pro os de lIEC

L Commis io Ele trote h iq e Intern tio ale (IEC) est la première org nisatio mo diale q i éla ore et p ble d s

Normes intern tio ales p ur to t c q i a trait à léle tricité, à léle tro iq e et a x te h olo ies a p re té s

A pro os de publc tio s IEC

L c nte u te h iq e d s p blc tio s IEC est c nstamme t re u Ve i ez v us as urer q e v us p s é ez lé itio la

plus ré e te, u c rig n um o ame d me t p ut a oir été p blé

Catalog e IEC - webstore.ie c / catalog e

Ap lc tio a to ome p ur c nsulter to s les r enseig eme ts

biblo ra hiq es sur les Normes intern tio ales,

Sp cific tio s te h iq es, Ra p r ts te h iq es et a tr es

d c me ts d l EC Disp nible p ur PC, Ma OS, ta letes

An roid et iPa

Re h rc e de publc tio s IEC - w w.ie c / se rc pub

L r ec er ch a a c e p rmet d tr ou er d s p blc tio s IEC

e uti sa t difér ents c tères (n mér o d référe c , te te,

c mité d’étu es,…) Ele d n e a s i d s infor matio s sur les

pr ojets et les p blc tio s rempla é s o r etir ées

IEC J st Publs ed - webstore.ie c / j stpubls ed

Restez infor mé sur les n u eles p blc tio s IEC Just

Pu lsh d d tai e les n u eles p blc t io s p r ues

Disp nible e lg e et a s i u e fois p r mois p r emai

Ele to edia - ww w.ele to edia.org

L pr emier dictio n ir e e lg e d ter mes éle tro iq es et

éle tr i u s I c ntie t 2 0 0 termes et d finitio s e a glais

et e fr an ais, ainsi q e les ter mes é uiv le ts d ns 15

la g es a ditio n les Eg leme t a p lé Vo a ulair e

Ele tr ote h iq e Inter natio al (IEV) e lg e

Glos aire IEC - std.ie c / glos ary

6 0 0 e tré s termin lo iq es éle tr ote h iq es, e a glais

et e fr an ais, e tr aites d s articles Ter mes et Définitio s d s

p blc tio s IEC p r ues d p is 2 0 Plus c rtain s e tr ées

a té e res e traites d s p blc tio s d s CE 3 , 7 , 8 et

CIS R d l EC

Serv ic Clents - webstore.ie c / cs

Si v us d sir ez n us d n er d s c mme tair es sur c te

p blc tio o si v us a ez d s q estio s c nta tez-n us:

csc@ie c

Industrial communication network s – High avai abi ity automation network s –

Part 3: Paral el Redundancy Protocol (PRP) and High-avai abi ity Seamless

Redundancy (HSR)

disponibi ité –

® Register edtr ademar k of the Inter national Electr otec ical Com ision

Mar que déosé de la Com ision Electr ot ec hniq e Inter nationale

®

Warnin ! Mak e s re th t y ou o tain d this publc tion from a a thorize distributor

Ate tion! Ve i ez v ou a s rer qu v ou av ez o te u c te publc tion via u distribute r a ré

CONTENTS

FOREWORD 8

INTRODUCTION 10 0.1 General 1

0 0.2 Chan es with resp ct to the previou edition 10 0.3 Patent declaration 10 1 Sco e 12 2 Normative referen es 1

2 3 Terms, definition , a breviation , acron ms, an con ention 13 3.1 Terms an definition 13 3.2 Ab reviation an acron ms 13 3.3 Con ention 14 4 Paral el Red n an y Protocol (PRP) 15 4.1 PRP prin iple of o eration 15 4.1.1 PRP network to olog 15 4.1.2 PRP LANs with l ne r or bu to olog 15 4.1.3 PRP LANs with rin to olog 16 4.1.4 DANP node stru ture 16 4.1.5 PRP at ac ment of sin ly at ac ed nodes 17 4.1.6 Comp tibi ty b twe n sin ly an doubly atac ed nodes 18 4.1.7 Network management 18 4.1.8 Implcation on a pl cation 18 4.1.9 Tran ition to non-red n ant networks 18 4.1.10 Dupl cate han l n 19 4.1.1 Network s p rvision 2

4.1.12 Red n an y management interface 2

4.2 PRP protocol sp cification 24 4.2.1 In tal ation, config ration an re air g idelnes 2

4.2.2 Unicast MAC ad res es 2

4.2.3 Multicast MAC ad res es 2

4.2.4 IP ad res es 2

4.2.5 Nodes 2

4.2.6 Dupl cate Ac e t mode ( estin only) 2

4.2.7 Dupl cate Dis ard mode 2

4.3 PRP_ up rvision frame 3

4.3.1 PRP_ up rvision frame format 3

4.3.2 PRP_ up rvision frame contents 31

4.3.3 PRP_ up rvision frame for RedBox 3

4.3.4 Rece tion of a PRP_ up rvision frame an NodesTa le 3

4.4 Brid in node 3

4.5 Con tants 3

4.6 PRP service sp cification 33 5 Hig -avai a i ty Se mles Red n an y (HSR) 3

5.1 HSR o jectives 3

5.2 HSR prin iple of o eration 34 5.2.1 Basic o eration with a rin to olog 3

5.2.2 DANH node stru ture 3

5.2.3 To olog 3

5.2.4 RedBox stru ture 4

5.3 HSR node sp cification 4

5.3.1 HSR o eration 4

5.3.2 DANH receivin from its l nk layer interface 4

5.3.3 DANH receivin from an HSR p rt 4

5.3.4 DANH forwardin rules 4

5.3.5 CoS 4

5.3.6 Clock s n hronization 5

5.3.7 Deterministic medium ac es 5

5.4 HSR RedBox sp cification 5

5.4.1 RedBox pro erties 5

5.4.2 RedBox receivin from interl nk 5

5.4.3 RedBox forwardin on the rin 5

5.4.4 RedBox receivin from an HSR p rt 5

5.4.5 RedBox receivin from its l nk layer interface 5

5.4.6 Redb x Prox NodeTa le han l n 5

5.4.7 RedBox CoS 5

5.4.8 RedBox clock s n hronization 5

5.4.9 RedBox medium ac es 5

5.5 QuadBox sp cification 5

5.6 Dupl cate Dis ard method 5

5.7 Frame format for HSR 5

5.7.1 Frame format for al frames 5

5.7.2 HSR_ up rvision frame 5

5.8 Con tants 5

5.9 HSR service sp cification 60 6 Protocol Implementation Conforman e Statement (PICS) 61

7 PRP/HSR Management Information Base (MIB) 6

An ex A (normative) Clocks s n hronization over red n ant p th in IEC 6 4 9-3 7

A.1 Overview 7

A.2 At ac ment to red n ant LANs by a b u dary clock 7

A.3 At ac ment to red n ant LANs by doubly atac ed ordinary clocks 8

A.4 PRP ma pin to PTP 8

A.4.1 Scenarios an device roles 8

A.4.2 Op ration in PRP 8

A.4.3 Config ration sp cification 8

A.4.4 Sp cification of DANP as DAC 8

A.4.5 Clock model of a RedBox for PTP 8

A.5 HSR Ma pin to PT 10

A.5.1 PTP trafic in HSR 10

A.5.2 HSR nodes sp cification 10

A.5.3 Red n ant clocks in HSR 10

A.5.4 At ac ment of an MC to an external LAN 10

A.6 PRP to HSR Ma pin 10

A.6.1 Con ection method 10

A.6.2 PRP-HSR con ection by BC 10

A.6.3 PRP-HSR con ection by TCs 10

A.7 Doubly atac ed clock model 1 0

A.7.1 State mac ine 1 0

A.7.2 Sup rvision of the p rt 1 3

A.7.3 BMCA for p ired p rts 1 4

A.7.4 Selection of the p rt state 1 5

A.8 PTP datasets for hig avai a i ty 1 5

A.8.1 General 1 5

A.8.2 Data typ s 1 5

A.8.3 Datasets for ordinary or bou dary clocks 1 6

A.8.4 Object for tran p rent clocks 12

An ex B (normative) PT profi e for Power Uti ty Automation – Red n ant clock

C.4 Protocol sp cification common 12

C.5 Protocol sp cification for L E2E automation profie 12

C.6 Protocol sp cification for L P2P automation profie 12

C.7 Timin req irements 12

C.7.1 Me s rement con ition 12

C.7.2 Network time inac urac 12

C.7.3 Network elements 12

C.7.4 Req irements for gran masters 12

C.7.5 Req irements for TCs 12

C.7.6 Req irements for BCs 12

C.7.7 Req irements for media con erters 12

C.7.8 Req irements for l nks 12

D.5 L yer 2 an layer 3 commu ication 13

D.6 1-ste an 2-ste cor ection 13

D.6.1 Time cor ection in TCs 13

D.6.2 2-ste to 1-ste tran lation 13

D.7 En -To-En l nk delay me s rement 13

D.7.1 General method 13

D.7.2 En - o-En l nk delay me s rement with 1-ste clock cor ection 13

D.7.3 En - o-En l nk delay me s rement with 2-ste clock cor ection 13

D.7.4 En - o-En l nk delay calc lation by Delay_ Req/Delay_ Resp 14

D.8 Pe r- o-Pe r lnk delay calc lation 14

D.8.1 Pe r- o-Pe r lnk delay calc lation with 1-ste cor ection 14

D.8.2 Pe r- o-Pe r lnk delay calc lation with 2-ste cor ection 141

An ex E (normative) Management Information b se for sin ly an doubly at ac ed

clocks 14

Bibl ogra h 16

Fig re 1 – PRP example of general red n ant network 15

Fig re 2 – PRP example of red n ant network as two LANs (bu to olog ) 16

Fig re 3 – PRP example of red n ant rin with SANs an DANPs 16

Fig re 4 – PRP with two DANPs commu icatin 17

Fig re 5 – PRP RedBox, tran ition from sin le to double LAN 19

Fig re 6 – PRP frame exten ed by an RCT 2

Fig re 7 – PRP VLAN- ag ed frame exten ed by an RCT 21

Fig re 8 – PRP p d ed frame closed by an RCT 21

Fig re 9 – Dupl cate Dis ard algorithm b u daries 2

Fig re 10 – HSR example of rin config ration for multicast traf ic 34 Fig re 1 – HSR example of rin config ration for u icast traf ic 3

Fig re 12 – HSR stru ture of a DANH 3

Fig re 13 – HSR example of to olog u in two in e en ent networks 3

Fig re 14 – HSR example of p er coupl n of two rin s 3

Fig re 15 – HSR example of con ected rin s 3

Fig re 16 – HSR example of coupl n two red n ant PRP LANs to a rin 4

Fig re 17 – HSR example of coupl n from a rin node to red n ant PRP LANs 41

Fig re 18 – HSR example of coupl n from a rin to two PRP LANs 42 Fig re 19 – HSR example of coupl n thre rin s to one PRP LAN 43 Fig re 2 – HSR example of mes ed to olog 4

Fig re 21 – HSR stru ture of a RedBox 4

Fig re 2 – HSR frame without a VLAN tag 5

Fig re 2 – HSR frame with VLAN tag 5

Fig re 2 – HSR node with management cou ters 6

Fig re 2 – HSR RedBox with management cou ters 61

Fig re A.1 – Doubly Atac ed Clock as BC (MCA is b st master) 79 Fig re A.2 – Doubly Atac ed Clock when MCA is b st master 81 Fig re A.3 – Doubly atac ed clocks when OC1 is b st master 82 Fig re A.4 – Elements of PRP networks 8

Fig re A.5 – Con ection of a master clock to an ordinary clock over PRP 8

Fig re A.6 – PRP RedBox as BCs (OC3 an BC7 are best masters) 8

Fig re A.7 – RedBox DABC clock model 8

Fig re A.8 – PRP RedBoxes as DABC with E2E – BC7 is master 8

Fig re A.9 – PRP RedBoxes as DABC with E2E – timin 9

Fig re A.10 – PRP RedBoxes as DABC with P2P – OC5 is b st master 91

Fig re A.11 – PRP RedBoxes as DABC with P2P – timin 9

Fig re A.12 – PRP RedBox as DATC with E2E –sig al flow 9

Fig re A.13 – PRP RedBox as DATC with E2E – timin 9

Fig re A.14 – PRP RedBox as DATC with P2P 9

Fig re A.15 – PRP RedBox as DATC with P2P – timin 9

Fig re A.16 – PRP RedBox as SLTC with E2E 10 Fig re A.17 – PRP RedBox as SLTC with E2E – timin 101 Fig re A.18 – PRP RedBox as SLTC with P2P 10 Fig re A.19 – HSR with one GMC 10 Fig re A.2 – PT mes ages sent an received by an HSR node (1-ste ) 10 Fig re A.21 – PT mes ages sent an received by an HSR node (2-ste ) 10 Fig re A.2 – Atac ment of a GMC to an HSR rin throu h a RedBox as TC 10 Fig re A.2 – PRP to HSR coupl n by BCs 10 Fig re A.2 – PRP to HSR coupl n by TCs 1 0 Fig re A.2 – Port states in lu in tran ition for red n ant o eration 1 1 Fig re A.2 – BMCA for red n ant masters 1 4 Fig re D.1 –Precision an ac urac example 13 Fig re D.2 – Precision Time Protocol prin iple 13 Fig re D.3 – Precision Time Protocol elements 13 Fig re D.4 – Delay an time-stampin logic in TCs 13 Fig re D.5 – Cor ection of the Sy c mes age by 1-ste an 2-ste (p er- o-p er) 13 Fig re D.6 – Tran lation from 2-ste to 1-ste in TCs 13 Fig re D.7 – Tran lation from 2-ste to 1-ste – mes age view 13 Fig re D.8 – En - o-en l nk delay me s rement with 1-ste clock cor ection 13 Fig re D.9 – En - o-en delay me s rement with 2-ste clock cor ection 14 Fig re D.10 – Pe r- o-p er l nk delay me s rement with 1-ste clock cor ection 141 Fig re D.1 – Pe r- o-p er l nk delay me s rement with 2-ste clock cor ection 14 Ta le 1 – Dupl cate dis ard cases 2

Ta le 2 – Monitorin data set 26 Ta le 3 – NodesTa le atributes 2

Ta le 4 – PRP_ up rvision frame with no VLAN tag 3

Ta le 5 – PRP_ up rvision frame with (o tional) VLAN tag 31

Ta le 6 – PRP_ up rvision frame contents 3

Ta le 7 – PRP_ up rvision TLV for Redb x 3

Ta le 8 – PRP con tants 3

Ta le 9 – HSR_ up rvision frame with no VLAN tag 5

Ta le 10 – HSR_ up rvision frame with o tional VLAN tag 58 Ta le 1 – HSR Con tants 6

Ta le A.1 – States 1 2

Ta le A.2 – Tran ition 1 3

Ta le A.3 – Varia les 1 3

Ta le C.1 – PTP atributes for the In u trial Automation profi e 12

INTERNATIONAL ELECTROTECHNICAL COMMISSION

1) Th Intern tio al Ele trote h ic l Commis io (IEC) is a worldwid org nizatio for sta d rdizatio c mprisin

al n tio al ele trote h ic l c mmite s (IEC Natio al Commite s) Th o je t of IEC is to promote

intern tio al c -o eratio o al q estio s c n ernin sta d rdizatio in th ele tric l a d ele tro ic fields To

this e d a d in a ditio to oth r a tivities, IEC p blsh s Intern tio al Sta d rds, Te h ic l Sp cific tio s,

Te h ic l Re orts, Pu lcly Av ia le Sp cific tio s (PAS) a d Guid s (h re fer refere to as “IEC

Pu lc tio (s)”) Th ir pre aratio is e truste to te h ic l c mmite s; a y IEC Natio al Commite intereste

in th su je t d alt with ma p rticip te in this pre aratory work Intern tio al g v rnme tal a d n

n-g v rnme tal org nizatio s laisin with th IEC also p rticip te in this pre aratio IEC c la orates closely

with th Intern tio al Org nizatio for Sta d rdizatio (ISO) in a c rd n e with c n itio s d termin d b

a re me t b twe n th two org nizatio s

2) Th formal d cisio s or a re me ts of IEC o te h ic l maters e pres , as n arly as p s ible, a intern tio al

c nse sus of o inio o th rele a t su je ts sin e e c te h ic l c mmite h s re rese tatio from al

intereste IEC Natio al Commite s

3) IEC Pu lc tio s h v th form of re omme d tio s for intern tio al use a d are a c pte b IEC Natio al

Commite s in th t se se Whie al re so a le eforts are ma e to e sure th t th te h ic l c nte t of IEC

Pu lc tio s is a c rate, IEC c n ot b h ld resp nsible for th wa in whic th y are use or for a y

misinterpretatio b a y e d user

4) In ord r to promote intern tio al u iformity, IEC Natio al Commite s u d rta e to a ply IEC Pu lc tio s

tra sp re tly to th ma imum e te t p s ible in th ir n tio al a d re io al p blc tio s An div rg n e

b twe n a y IEC Pu lc tio a d th c r esp n in n tio al or re io al p blc tio sh l b cle rly in ic te in

th later

5) IEC itself d es n t pro id a y atestatio of c nformity In e e d nt c rtific tio b dies pro id c nformity

as es me t servic s a d, in some are s, a c s to IEC marks of c nformity IEC is n t resp nsible for a y

servic s c rie o t b in e e d nt c rtific tio b dies

6) Al users sh uld e sure th t th y h v th latest e itio of this p blc tio

7) No la i ty sh l ata h to IEC or its dire tors, emplo e s, serv nts or a e ts in lu in in ivid al e p rts a d

memb rs of its te h ic l c mmite s a d IEC Natio al Commite s for a y p rso al injury, pro erty d ma e or

oth r d ma e of a y n ture wh tso v r, wh th r dire t or in ire t, or for c sts (in lu in le al fe s) a d

e p nses arisin o t of th p blc tio , use of, or rela c u o , this IEC Pu lc tio or a y oth r IEC

Pu lc tio s

8) Ate tio is drawn to th Normativ refere c s cite in this p blc tio Use of th refere c d p blc tio s is

in isp nsa le for th c re t a plc tio of this p blc tio

International Stan ard IEC 6 4 9-3 has b en pre ared by s bcommite 6 C: In u trial

networks, of IEC tec nical commite 6 : In u trial-proces me s rement, control an

automation

This third edition can els an re laces the secon edition publ s ed in 2 12 This edition

con titutes a tec nical revision

This edition in lu es the fol owin sig ificant tec nical c an es with resp ct to the previou

edition:

a) tec nical cor ection an exten ion of sp cification ;

b) con ideration of IEC 615 8 clock s n hronization with en - o-en delay me s rement

alon side the existin p er- o-p er delay me s rement in PRP

The text of this stan ard is b sed on the fol owin doc ments:

Ful information on the votin for the a proval of this stan ard can b fou d in the re ort on

votin in icated in the a ove ta le

This publcation has b en drafed in ac ordan e with the ISO/IEC Directives, Part 2

This International Stan ard is to b re d in conju ction with IEC 6 4 9-1

A l st of al p rts in the IEC 6 4 9 series, publ s ed u der the general title Industrial

communicato n tworks – Hig a aiab il y au tomato n tworks, can b fou d on the IEC

we site

The commit e has decided that the contents of this publ cation wi remain u c an ed u ti

the sta i ty date in icated on the IEC we site u der "htp:/we store.iec.c " in the data

related to the sp cific publ cation At this date, the publ cation wi b

indicates that it contain colours whic are considered to be useful for the cor ect

understan ing of its contents Users s ould therefore print this publ cation using a

0.1 General

IEC 6 4 9-3 b lon s to the IEC 6 4 9 series “Industrial commu nicato n tworks – High

a aiab il y au tomato n tworks” It sp cifies the PRP an HSR se mles red n an y

protocols It was ado ted by IEC TC5 WG10 as the red n an y method for deman in

s bstation automation networks o eratin on layer 2 networks, ac ordin to IEC 618 0-8-1

an IEC 618 0-9-2

The se mles red n an y prin iple has b en exten ed to clocks o eratin ac ordin to the

Precision Time Protocol (IEC 615 8) an atac ed to red n ant networks Two variants are

sp cified: L E2E for clocks whic o erate on layer 3 networks with en - o-en l nk delay

me s rement (E2E) an L P2P for clocks that o erate on layer 2 with p er- o-p er lnk delay

– An ex A has b en redrafed as a general con e t for doubly atac ed clocks a pl ca le to

en - o-en (E2E) an to p er- o-p er (P2P) l nk delay me s rement; the prin iple of

p ired p rt o eration has now b en sp cified in terms of a state mac ine b sed on

IEC 615 8:2 0

– An ex B of IEC 6 4 9-3:2 12 has b en deleted; its pro erties are mentioned in 5.3.7

– An ex B (new) ma es the s p ort of red n an y man atory for IEC/IEEE 618 0-9-3 that

sp cifies doubly atac ed clocks on layer 2, with p er- o-p er delay me s rement

– An ex C sp cifies two profi es of a precision clock for in u trial automation:

L E2E for layer 3, en - o-en delay me s rement an

L P2P for layer 2, p er- o-p er delay me s rement

– An ex D contain the tutorial information on IEC 615 8:2 0 for u derstan in the a ove

an exes It was contained in IEC 6 4 9-3:2 12 An ex A

– An ex E (MIB) contain the SNMP Management Information Base to b u ed for sin ly

an doubly atac ed clocks in al profi es

0.3 Patent declaration

The International Electrotec nical Commis ion (IEC) draws atention to the fact that it is

claimed that compl an e with this doc ment may in olve the u e of a p tent con ernin

fi terin of red n ant frames in a network node (Siemen Aktien esel s haf – EP 212 3 9,

US 818 6 0, CN 10161 615B) given in 5.2.3.3

IEC ta es no p sition con ernin the eviden e, val dity an s o e of this p tent rig t

The holder of this p tent rig t has as ured the IEC that he/s e is wi in to negotiate lcen es

u der re sona le an non-dis riminatory terms an con ition with a plcants throu hout the

world In this resp ct, the statement of the holder of this p tent rig t is registered with IEC

Information may be o tained from:

Siemen Aktien esel s haf

Oto-Hah -Rin 6

813 9 Mu ic , German

The International Electrotec nical Commis ion (IEC) draws atention to the fact that it is

claimed that complan e with this doc ment may in olve the u e of p tents con ernin

Rece tion of red n ant an non-red n ant frames (ABB Rese rc Ltd – EP 18 5 5 ,

US 8 8 4 6, CN 1010 7 8 , IN 2 4 2 ) given in 4.2.7, con ernin Identifyin impro er

ca l n of devices (ABB Tec nolog AG – EP 216 0 4, US 8 4 7 6, CN 1016 9 8 ) given in

4.3, con ernin Critical device with in re sed avaia i ty (ABB Rese rc Ltd – EP 2 9 9 0)

given in 4.4, con ernin Rin coupl n nodes for hig avai a i ty networks (ABB Rese rc Ltd

– US 8 8 4 4, EP 2 2 18 , CN 10 10 121) given in 5.2.3

IEC ta es no p sition con ernin the eviden e, val dity an s o e of these p tent rig ts

The holder of these p tent rig ts has as ured the IEC that he/s e is wi in to negotiate

l cen es u der re sona le an non-dis riminatory terms an con ition with a pl cants

throu hout the world In this resp ct, the statement of the holder of these p tent rig ts is

registered with IEC Information may b o tained from:

Atention is drawn to the p s ibi ty that some of the elements of this doc ment may b the

s bject of p tent rig ts other than those identified a ove IEC s al not b held resp n ible for

identifyin an or al s c p tent rig ts

ISO (www.iso.org/p tents) an IEC (htp:/ p tents.iec.c ) maintain on-l ne data b ses of

p tents relevant to their stan ard Users are en ouraged to con ult the data b ses for the

most up to date information con ernin p tents

This p rt of IEC 6 4 9 sp cifies two red n an y protocols desig ed to provide se mles

recovery in case of sin le faiure of an inter-brid e l nk or brid e in the network, whic are

b sed on the same s heme: p ral el tran mis ion of d pl cated information

2 Normativ referenc s

The folowin doc ments, in whole or in p rt, are normatively referen ed in this doc ment an

are in isp n a le for its a pl cation For dated referen es, only the edition cited a pl es For

u dated referen es, the latest edition of the referen ed doc ment (in lu in an

IEC TR 618 0-9 -4:2 13, Commu nicato n tworks a d systems forp owe r u ti ty au tomato

– Part 90-4: Network e gin e rin guide lin s

IEC 6 4 9-1, Industrial commu nicato n tworks – Hig a aiab il y au tomato n tworks –

Part 1: Gen ral co ce pts a d calculato meth ds

IEC/IEEE 618 0-9-3:—, Communicato n tworks a d systems forp owe r u ti ty automato -

Part 9-3: Pre cisio tme protocol p rofie for p oweru ti ty automato (prop osed IEC 618 0

-9-3)

1

ISO/IEC/IEEE 8 0 -3:2 14, Sta dard for Eth rn t

IEEE 8 2.1D:2 0 , IEEE Sta dard forLocal a d metrop ol a are a n tworks – Media Ac es

Co trol (MAC) Bridg s

IEEE 8 2.1Q:2 14, IEEE Sta dard for Local a d metrop ol a are n tworks – Media Ac e ss

Co trol (MAC) Bridges a d Virtual Bridge Local Are Network

IET RFC 2 7 , Structure ofMa a eme t Informato Versio 2 (SMIv )

IET RFC 3 18, Ma a eme t Informato Base (MIB) for th Simple Network Ma ageme t

Protocol (SNMP)

3 Terms, definitions, abbreviations, acronyms, and conventions

3.1 Terms and definitions

For the purp ses of this doc ment, the terms an definition given in IEC 6 0 0-191 an in

IEC 6 4 9-1, as wel as the fol owin , a ply

3.1.1

grandmaster-ca a le

ordinary clock or b u dary clock that is a le to ta e the role of a gran master

Note 1 to e try: A gra dmaster-c p ble clo k is n t n c s ariy c n e te to a re o nize time so rc

time inac urac evaluated or me s red b twe n a p rtic lar slave clock an the time

referen e sig al of its gran master clock(s), con iderin the worst network to olog

q adruple p rt device con ectin two p er HSR rin s, whic b haves as an HSR node in

e c rin an is a le to fi ter the trafic an forward it from rin to rin

time er or not ex e ded by 9 ,7% of the me s rements, evaluated over a series of 1 0 0

me s rements (a out 2 min) in ste d state

Note 1 to e try: In this d c me t, th words “brid e” a d“brid in ” are sy o ymo s with th words “switc ” a d

“switc in ” resp ctiv ly,wh n th y a ply to la er 2 c n e tivity

3.2 Abbreviations an acronyms

For the purp ses of this doc ment, the a breviation an acron ms given in IEC 6 4 9-1, as

wel as the fol owin a ply

ARP Ad res Resolution Protocol

BC Bou dary Clock, as defined in IEC 615 8:2 0

BMCA Best Master Clock Algorithm, as defined in IEC 615 8:2 0

BOC Bou dary Ordinary Clock

CoS Clas of Service (layer 2 q alty of service)

MIB Management Information Base

OC Ordinary Clock, as defined in IEC 615 8:2 0

P2P Pe r to Pe r (delay me s rement

PICS Protocol Implementation Conforman e Statement

TC Tran p rent Clock, as defined in IEC 615 8:2 0

VDAN Virtual Doubly At ac ed Node (SAN as visible throu h a RedBox)

4 Paral el Redundancy Protocol (PRP)

4.1 PRP principle of operation

4.1.1 PRP network to ology

This red n an y protocol implements red n an y in the nodes rather than in the network (se

the definition in IEC 6 4 9-1), u in Doubly At ac ed Nodes o eyin to PRP (DANPs)

A DANP is at ac ed to two in e en ent L cal Are Networks (LANs) of simi ar to olog ,

named LAN_A an LAN_ , whic o erate in p ralel A source DANP sen s the same frame

over b th LANs an a destination DANP receives it from b th LANs within a certain time,

con umes the first frame an dis ard the d pl cate

Fig re 1 s ows a red n ant network con istin of two LANs, e c of whic can have an

to olog , e.g tre , rin or mes ed

Figure 1 – PRP example of general redundant network

The two LANs are identical in protocol at the MAC-L C level, but they can difer in

p rforman e an to olog Tran mis ion delay may also b diferent, esp cial y if one of the

networks reconfig res itself, e.g u in RSTP, to overcome an internal fai ure

The two LANs fol ow config ration rules that al ow the network management protocols s c as

Ad res Resolution Protocol (ARP) an Simple Network Management Protocol (SNMP) to

o erate cor ectly

The two LANs have no con ection b twe n them an are as umed to b fai -in e en ent

Red n an y can b defe ted by sin le points of faiure, s c as a common p wer s p ly or a

direct con ection whose fai ure brin s b th networks down In tal ation g idel nes in this

doc ment provide g idan e to the in tal er to ac ieve fai -in e en en e

Figure 2 – PRP example of redun ant network as two LANs (bus topology)

4.1.3 PRP LANs with ring topology

The two LANs can have a rin to olog , as Fig re 3 s ows

Red n an y within e c rin is provided by a dif erent mec anism, for in tan e as defined in

IEEE 8 2.1D or in other IEC 6 4 9 protocols

Fig re 3 – PRP example of red ndant ring with SANs an DANPs

4.1.4 DANP node stru ture

Eac node has two p rts that o erate in p ral el an that are atac ed to the same up er

layers of the commu ication stack throu h the Link Red n an y Entity (LRE), as Fig re 4

Figure 4 – PRP with two DANPs communicating

For the b sic commu ication, the LRE presents toward its up er layers the same interface as

a non-red n ant network ada ter, so the up er layers are u aware of red n an y

The LRE has two tasks: han l n of d pl cates an management of red n an y

When receivin a frame from the node’s up er layers, the LRE a p n s to the frame a

Red n an y Check Trai er (RCT) containin a seq en e n mb r an sen s the frame throu h

b th its p rts at ne rly the same time The two frames are ne rly identical ex e t for the LAN

identifier (an the c ecks m)

The two frames tran it throu h the two LANs with diferent delay , ide l y they ar ive at ne rly

the same time at the destination node

When receivin frames from the network, the LRE forward the first received frame of a p ir

to its node’s up er layers an dis ard the d pl cate frame (if it ar ives) It removes the RCT if

req ired

For management of red n an y an c eckin of the presen e of other DANPs, an LRE

p riodicaly sen s PRP_ Sup rvision frames an can evaluate the PRP_ up rvision frames

sent by other DANPs

4.1.5 PRP at ac ment of singly at ac ed nodes

Sin ly at ac ed nodes (SANs) can b at ac ed in two way :

SANs can b atac ed directly to one LAN only Su h SANs can only commu icate with other

SANs on the same LAN For in tan e, in Figure 1, SAN A1 can commu icate with SAN A2,

but not with SAN B1 or SAN B2 SANs can commu icate (not red n antly) with al DANPs

SANs can b atac ed over a Red n an y Box (RedBox) to b th LANs, as Fig re 1 s ows for

SAN R1 an SAN R2 (se also 4.1.9) Su h SANs can commu icate with al DANP an SANs,

for in tan e SAN A1 an SAN R1 can commu icate

IEC

por tALin Re u dan y Entity

NOT SA s are n t aware of PRP; th y are for insta c of th -sh lf c mp ters or printers.

In some a pl cation , only avai a i ty-critical devices ne d a double at ac ment, for in tan e

the o erator work laces, whi e the majority of the devices are SANs Ta in ad antage of the

b sic infrastru ture of PRP, a DANP can b at ac ed to two diferent brid es of the same LAN

(e.g a rin ) an u e protocols diferent from PRP to reconfig re the network in case of fai ure

The DANP then b haves as a brid e ac ordin to IEEE 8 2.1D For in tan e, the brid e

element may implement the RST protocol or a s bset of RSTP where there is no forwardin

of traf ic b twe n the p rts These a i ties are o tional an not detai ed in this doc ment The

s p orted mode is sp cified in Clau e 6

4.1.6 Compatibi ity b twe n singly and doubly at ac ed nodes

Sin ly at ac ed nodes (SAN), for in tan e maintenan e la to s or printers that b lon to one

LAN, can b con ected to an LAN Brid es are alway SANs A SAN con ected to one LAN

can ot commu icate directly to a SAN con ected to the other LAN These SANs are not

aware of PRP red n an y, but DANPs generate a trafic that these SANs u derstan

The con ition is however that the SANs ig ore the RCT in the frames, whic is the case sin e

a SAN can ot distin uis the RCT from the Ethernet p d in

Con ersely, DANPs u derstan the trafic generated by SANs, sin e these do not a p n an

RCT They only forward one frame to their up er layers sin e the SAN trafic u es one LAN

only

4.1.7 Network management

A node has the same MAC ad res on b th p rts an has only one set of IP ad res es

as ig ed to that ad res This ma es red n an y tran p rent to the up er layers an

esp cial y al ows ARP to work as with a SAN Brid es in a LAN are not doubly at ac ed

devices, an therefore al managed brid es have dif erent IP (an MAC) ad res es A

network management to l is prefera ly a DANP that can ac es nodes an brid es in b th

LANs

4.1.8 Impl cation on a pl cation

When the same frame comes from the two p rts with sig ificant time diferen e, a frame

could, in ex e tional situation , b wron ly rejected as d plcate Therefore, there exists a

mec anism to age out frames This agein mec anism cau es in seldom cases that b th

d pl cates are ac e ted The d pl cate mu t then b fitered out by the tran p rt layer (e.g

TCP) or tolerated by the a pl cation This is a normal case sin e a LAN do s not g arante

a sen e of d pl cates

The agein mec anism con iders b th the maximum re etition rate of the frames an the

worst case dif eren e in pro agation delay, as des rib d in 4.1.10.3

4.1.9 Transition to non-redu dant networks

The mec anism of d plcate rejection can b implemented by a RedBox that do s the

tran ition b twe n a SAN an the doubled LANs, as Fig re 5 s ows The RedBox mimic the

SANs con ected b hin it (caled VDAN or virtual DANs) an multicasts s p rvision frames

on their b half The RedBox is itself a DANP an has its own IP ad res for management

purp ses, but it may also p rform a pl cation fu ction

Figure 5 – PRP RedBox, transition from single to double LAN

4.1.10 Dupl cate han l n

4.1.10.1 Dupl cate Ac e t mode ( esting only)

The Dupl cate Ac e t mode is u ed for testin purp ses, to verify that d pl cates are in e d

dis arded by the l nk layer an not by hig er layer protocols

In the Dupl cate Ac e t mode, a sen er is config red to sen a normal frame over b th p rts

In the Duplcate Ac e t mode, a receiver is config red to ac e t b th frames an forward b th

of them (if b th ar ive) to its up er layers

A node in icates its mode (Dupl cate Ac e tDuplcate Dis ard) in its s p rvision frames

NOT Up er la ers also disc rd d plc tes, b t les efe tiv ly th n th ln la er

4.1.10.2 Dupl cate Dis ard mode

4.1.10.2.1 Principle

To alow the receivers to detect d plcates, the sen er LRE a p n s a six-octet field that

contain a seq en e n mb r, the Red n an y Control Trai er (RCT), to b th frames it sen s,

as Fig re 6 s ows

The receiver LRE u es the seq en e n mb r of the RCT an the source MAC ad res to

detect d plcates It forward only the first frame of a p ir to its up er layers To oflo d the

a pl cation proces or, the LRE can b implemented with an in e en ent pre-proces or, an

intel gent Ethernet control er or in hardware

Ke pin track of d pl cates al ows at the same time to improve the red n an y s p rvision.

As an o tion, the receiver may remove the RCT b fore forwardin the ( irst frame to its up er

The LRE sen s b th (ne rly identical) frames over b th LANs

The receivin LRE can then detect d pl cates b sed on the RCT

NOT 1 This meth d c nsid rs th t SA s also e ist o th n twork, a d th t frames se t b SA s c uld b

wro gly reje te as d plc tes b c use th y h p e to h v a trai n field with th c re t PRP sufix, same

se u n e n mb r a d th same size Howe er, SA s se d o o e LA o ly, th refore th ir so rc MA a dres

wi n t b th same as th t of a frame o th oth r LA , so th t a frame from a SA wi n v r b disc rd d

NOT 2 Re e tio of a wel- orme R T is n pro f th t th so rc n d is a D NP Ev n if th p th is wro g, a

DA P a c pts th frame

4.1.10.2.2 Use of PRPs f ix

The PRPs f ix identifies PRP frames an distin uis es PRP frames from other protocols that

also a p n a trai er to their u eful data

4.1.10.2.3 Use of LSDUsize

The 12-bit LSDUsize field contain the len th of the lnk service data u it (LSDU) in octets

The LSDUsize field al ows the receiver to distin uis PRP frames from ran om, non

-red n ant frames as an ad itional c eck

In VLANs, frame VLAN tag can b ad ed or removed d rin tran it throu h a brid e To

ma e the len th field in e en ent of VLAN tag in , only the LSDU an the RCT are

con idered in the LSDUsize, as Fig re 7 s ows

IEC time

Fig re 7 – PRP VLAN- agged frame extended by an RCT

If the receiver detects that the frame is closed by the cor ect PRP s fix, that the 12 bits

b fore the PRP s f ix cor esp n to the LSDU size, an that the LAN identifier matc es the

identifier of the LAN of the receivin p rt, the frame is a can idate for dis ardin

Sin e s ort frames ne d p d in to ac ieve the minimum frame size of 6 octets, the sen er

alre d in lu es the p d in to avoid s an in the frame b ckward , as Fig re 8 s ows The

minimum frame size is exten ed to 7 octets sin e the LRE or a RedBox can remove the

RCT

Fig re 8 – PRP pad ed frame closed by an RCT

The p d in size is in re sed to 7 octets when a VLAN tag is u ed

NOT A VLA -a g d frame c n p s se eral brid es, whic remo es or inserts VLA ta s If th se d r

o serv s th Eth rn t rule to se d a minimum frame size of 6 o tets for a VLA - a g d frame a d of 6 o tets for

aVLA -u ta g d frame, th re wi n v r b a situ tio in whic th re is p d in b fore a d afer th R T

4.1.10.2.4 Use of LanId

The 4 bit L nId field of the RCT car ies a distin t identifier for LAN_A or LAN_ , sp cificaly

the codes 1010 ( A”) an 101 ( B”) Therefore, the A- rame an the B- rame difer in one bit

(an in the FCS) The receiver c ecks that the frame comes from the cor ect LAN It do s not

reject a frame that comes from the wron LAN, sin e this could b a legitimate frame whic

ha p n to have the PRP s fix an the len th information in its last 2 bits, but it in rements

the er or cou ters CntEr Wron L nA or CntEr Wron L nB sin e this hints at a config ration

er or Sin e this kin of er or is p rmanent, it is detected ra idly

4.1.10.2.5 Use of SeqNr

The 16-bit field SeqNr is in remented by one for e c frame a DANP sen s The doublet

{source MAC ad res , seq en e n mb r} u iq ely identifies co ies of the same frame

4.1.10.3 Dupl cate dis ard algorithm an configuration

The algorithm for dis ardin d pl cates is not sp cified For in tan e, node ta les, has

ta les, fifos or seq en e n mb r trackin can b u ed Whic ever the algorithm, it is

desig ed s c that it never rejects a legitimate frame, whi e oc asional ac e tan e of a

d pl cate can b tolerated Sin e the 16-bit SeqNr wra s arou d afer 6 5 6 frames an the

IEC time

pre mble

PRP

sufixLSD

seq en e n mb rs are not neces ari y contig ou , entries older than a sp cified time

EntryForgetTime are purged

Fig re 9 s ows the con ition for dis ardin or ac e tin frames

The time t

k w

is the time diferen e b twe n the ar ivals of two co ies of the same frame

( rom LAN A to LAN B an from LAN B to LAN A) in a p rtic lar case

The maximum time t

k wMa

b twe n two co ies is a network pro erty, estimated by the

network desig er b sed on the n mb r of brid es an the trafic for a p rtic lar a pl cation,

e.g 12 ms

The time t

alasMin

is the minimum p s ible interval b twe n two dif erent, legitimate frames

with the same seq en e n mb r (afer 6 5 6 in rements) in a p rtic lar a pl cation

NOT Howe er, a n d th t re o ts with a arbitrary se u n e n mb r c uld c use a sh rte in of time t

alasMin,

as wi b d taie b low

In a 10 Mbits network, t

alasMin

is the the retical minimum frame re etition interval of the

smalest frame (6 octets) from the same source (6 5 6 × 6,7 µs) = 4 0 ms

In a 1 Gbit s, t

alasMin

is 4 ms

This is a worst case sin e a pl cation rarely sen minimum size frames contig ou ly

EX MPLE: An IEC 618 0-9-2 d vic se din sample v lu s frames at 4,8 kHz wi e hibit a t

.These values de en on the implementation an are

adeq ate for a n mb r of a pl cation , they can de en on the traf ic if the l st size is b u d

It is recommen ed to ma e t

resid Min

resid Ma

adju ta le de en in on the data rate

The folowin cases of Fig re 9 are con idered in Ta le 1:

Th d plc te is sometimes n t disc rd d if th e try was p rg d.

This situ tio is tolera le if infre u nt

frame a d cre tes a n w e try in th d plc tes lst

This situ tio n v r o c rs if th n twork was pro erly dime sio e

with (t

k wMax

< t

resideMax)

Th d plc te is n v r disc rd d, b t a le itimate frame is disc rd d

sin eth e try cre te b LA B c nflcts with th alas o LA A

This situ tio n v r o c rs if th n d s are pro erly d sig e

This situ tio n v r o c rs if th n d s are pro erly d sig e

The con ition for cor ect o eration are as fol ows:

is enforced by a timer This value is adju ta le throu h

the management set in lreDup istResideMaxTime (se Clau e 7)

A node that re o ts could generate al ases at a lower interval than t

alasMin

sin e the

seq en e n mb r at restart is arbitrary To al ow al p s ible destination nodes time to flu h

al entries of this node from the d pl cates l st, a node may not re o t faster than

NodeRe o tInterval (se Ta le 8) NodeReb otInterval hold for al nodes in the network an

therefore it is in e en ent from the l nk sp ed

Therefore, al nodes flu h their entries b fore NodeRe o tInterval The maximum residen e

time of an entry regardles of the sp ed (an of the value of t

Ap en in the RCT could generate oversize frames that ex e d the maxValdSize forese n

by ISO/IEC/IEEE 8 0 -3:2 14 (1 5 2 octets)

Hen e, it is exp cted that every network comp nent in the LANs s p orts oversize frames up

to maxVal dSize forese n by ISO/IEC/IEEE 8 0 -3:2 14 plu 6 bytes of RCT (1 5 8 octets)

NOT Ma y brid es are dime sio e for d u le-VLA - a g d (Q-in-Q) frames th t h v a ma imum size of

1 5 6 o tets Most Eth rn t c ntrolers are c rtifie u to 1 5 8 o tets Most brid es forward c r e tly frames of u

4.1.1 Network s pervision

The he lth statu of e c LAN an its at ac ed devices (nodes an brid es) is monitored,

otherwise red n an y helps l t le

The receiver c ecks that frames are cor ectly received over b th c an els It maintain er or

cou ters that network management can re d

To this ef ect, sen ers an receivers can maintain ta les that record the last time a frame was

received from another node, the time a multicast or bro d ast frame was sent an other

protocol information

Sup rvision rel es on e c DANP sen in p riodical y a PRP_ up rvision frame that al ows

c eckin the integrity of the network an the presen e of the nodes At the same time, these

frames al ow c eckin whic devices are DANP, the MAC ad res es they u e an whic

o eratin mode they s p ort: Dupl cate Ac e t or Dupl cate Dis ard

4.1.12 Red ndanc management interface

Red n ant devices an l nks are u eles without network management s p rvisin

red n an y an cal n for maintenan e action

The LRE presents a network management interface that alows to track the he lth of e c

LAN, an esp cialy to detect fai ures e rly when the er or rate in re ses To this efect, the

LRE k e s for e c p rt a cou ter of received mes ages an of mes ages received with an

The fol owin g idel nes a ply at in tal ation time They do not a ply to conforman e testin

of the devices

4.2.1.2 LANs layout

The network s al con ist of two LANs that have simi ar prop rties, i.e e c one is a le to

car y the traf ic that would exist in the a sen e of red n an y

DANP s al b at ac ed to b th LANs

SANs that ne d to commu icate with one another s al b at ac ed to the same LAN or to

b th LANs throu h a RedBox

4.2.1.3 La el ng ca les

The two LANs s al u e ca les distin tly identified as A an B

If colours are u ed to identify the ca les, A s ould b red an B s ould b blue

NOT Th inte tio is th t p rso s with c lo r disa i ty distin uish th m

SANs con ected to one LAN only s al not have the same MAC ad res as another node

within the whole network (LAN_A plu LAN_ )

NOT 1 Uniq e es a ples to al IP a dres es of a d vic

A DANP s al have the same IP ad res (es) when se n from either LAN_A or LAN_

Brid es on LAN_A an LAN_ are con idered as SANs an s al have diferent IP ad res es

for the purp se of network management, even if they oc upy the same location

NOT 2 A p s ible sc eme is to distin uish IP a dres es b o e bit, e.g alo ate e e IP a dres es to LA _A

d vic s a d o d IP a dres es to LA _ d vic s

4.2.5 Nodes

4.2.5.1 Node types

Doubly atac ed nodes ac ordin to the p ral el red n an y protocol (DANP) s al have two

p rts (p rtA an p rtB) that have the same a i ties, an in p rtic lar could b u ed

alternatively if only one LAN would b con ected, p rtA b in con ected to LAN_A an p rtB

to LAN_

Sin ly At ac ed Nodes (SAN) have only one p rt for the purp se of this protocol, no sp cial

req irements a ply to them

4.2.5.2 La el ng connectors

Subclau e 4.2.5.2 a pl es to a DANP u in two LANs of simi ar nature

The con ectors for e c LAN s al b la el ed distin tly as A an B

When con ectors are ordered vertical y, LAN_A s ould b the up er con ector an LAN_

the lower con ector in the normal p sition of the device

When con ectors are ordered horizontaly, the lef con ector s ould b the LAN_A an the

rig t con ector the LAN_ B, as se n from the side where the ca les or fibres are plu ged

The red n ant con ectors s al b in e en ently remova le an in erta le

4.2.6 Dupl cate Ac e t mode ( esting only)

4.2.6.1 Sen ing

A sen er in the Dupl cate Ac e t mode s al sen the frame it receives from its up er layers

u c an ed over b th its p rts so that the two frames ap e r on the resp ctive LANs

Activ A a d Activ B o tio al For e c p rt A a d B, th status of th c mmu ic tio ln

No esTa le o tio al A ta le of al oth r n d s th t this n d d te ts

4.2.7.2 NodesTa le

A node may maintain a NodesTa le for the purp se of network s p rvision If a NodesTa le

is u ed, it s al con ist of the elements given in Ta le 3 for e c entry

Ta le 3 – NodesTa le at ributes

CntRe eiv dA Numb r of frames re eiv d from th t n d o er LA _A Unsig e 3

CntRe eiv dB Numb r of frames re eiv d from th t n d o er LA _ Unsig e 3

CntErWro g a A Numb r of frames th t were re eiv d from th t n d

with th wro g LA id ntifier o LA _A

Unsig e 3

CntErWro g a B Numb r of frames th t were re eiv d from th t n d

with th wro g LA id ntifier o LA _

A node s al cle r a NodesTa le entry of a given other node when the time ela sed sin e

rece tion of a frame from that node over b th Time astSe nA an Time astSe nB ex e d

NodeForgetTime (se 4.5)

NOT 1 Th No esTa le is p p late b th re eiv d frames a d distin uish s SA s from DA Ps Th

No esTa le alows a n d to id ntify a oth r n d as SA o o e LA o ly a d in this c se th n d c uld se d

frames with ut a R T o er th p rt of th t LA o ly A n d with ut a No esTa le alwa s se ds al frames with

a R T o er b th p rts, also wh n th d stin tio is a SA In a plc tio s wh re bro d ast trafic is th b lk, th

imp ct o n twork lo d is smal

NOT 2 Th k y atrib te of th No esTa le is Ma Ad res as re eiv d in th PRP Su ervisio frame se t b a

DA P or in a y frame re eiv d from a SA

NOT 3 Most of th se atrib tes e ist n t o ly in o e insta c p r p ysic l remote n d , b t also as se arate

insta c s for e c multi/bro d ast a dres use b th t n d , a d some also for e c multi/bro d ast a dres

use b this (lo al) n d

NOT 4 TimeTic s is a lo al time u it, n t a glo al time

4.2.7.3 Redundan y Control Trai er (RCT)

The Red n an y Control Trai er (RCT) a p n ed to DANP frames s al con ist of six octets,

stru tured in the folowin way (in the order of tran mis ion):

• a 16-bit seq en e n mb r (SeqNr) tran mited with the most sig ificant 8 bits in the first

octet, whic reflects the cou ter Sen Seq (se 4.2.7.1);

• a 4-bit LAN identifier (L nId) tran mited as the most sig ificant 4 bits of the third octet,

whic car ies the seq en e “1010” for LAN_A an the seq en e “101 ” for LAN_ ,

resp ctively;

• a 12 bit LSDU size (LSDUsize) whose most sig ificant 4 bits are tran mit ed in the le st

sig ificant 4 bits of the third octet an the le st sig ificant 8 bits in the fourth octet, that

in icates the size in octets of the LSDU startin from the en of the Protocol Typ (PT)

field as defined in ISO/IEC/IEEE 8 0 -3:2 14 an IEEE 8 2.1Q (octet of set 12-13 without

LAN he der or 16-17 with VLAN he der) to the RCT, ex lu in the PT an in lu in

NOT 2 Th total p d in to 7 o tets (n VLA ta ) c nsid rs th t th LRE or a Re Bo c uld remo e th

6-o tet RC Pa din to 7 o tets (with VLA ta ) c nsid rs th t in a ditio , a e g p rt in a brid e c uld remo e

th 4-o tet priority ta This wa , th frame giv n to th d stin tio resp cts th minimum frame size of 6 o tets

a d th RCT is alwa s just b fore th FCS

4.2.7.4.2 Timing

The time sk w b twe n the sen in of b th frames of a p ir s ould b k pt at a minimum to

red ce as mmetry of the tran mis ion delay

In p rtic lar, frames waitin for sen in s ould b dro p d from the sen in q eue if the

cor espon in p rt is not o erational

Frames with a destination MAC ad res ac ordin to IEEE 8 2.1Q:2 14, Ta le 8-1, Ta le 8-2

an Ta le 8-3 are tre ted at the l nk layer an are not exp cted to car y an RCT, even if sent

by a DANP

4.2.7.5.2 Identification of frames as ociated with the Dupl cate Dis ard mode

A receiver s al identify a d pl cate can idate as a val d frame:

• whose last 16 bits b fore the c ecks m matc the 16-bit PRPs f ix,

• whose 12 bits b fore the PRPs fix matc the ph sical size of the LSDU (ex lu in the

ProtocolTyp PT, but in lu in the RCT) an

• whose L nId matc es the LAN identity of the receivin p rt, 1010 (A) or 101 (B)

4.2.7.5.3 Dupl cate Dis ard method

The Dupl cate Dis ard method s al not dis ard a frame sent as sin le or b th frames of a

p ir; ex e tional y, b th frames of a p ir may b p s ed to the hig er protocol layers

The Dupl cate Dis ard method s al forget an entry afer EntryForgetTime (se Ta le 8)

4.2.7.5.4 Transparent rece tion

A receiver config red for Tran p rentRece tion (default s al not remove the RCT b fore

tran fer in the frame to its up er layers

A receiver not config red for Tran p rentRece tion s al remove the RCT on frames where it

has identified the presen e of the RCT

4.2.7.5.5 Receiving and NodesTa le

If the node s p orts a NodesTa le, it s al , at rece tion of a frame:

If t i f am e i n t a PR P_S p rvi i n f am e o a H R S p rvi i n f ame a d i s

NOT 2 Re e tio of a wel-orme RCT is n t a suficie t criterio to d clare its so rc as D NP, sin e some

proto ols re ly with th same frame as re eiv d

4.3 PRP_ upervision frame

4.3.1 PRP_ upervision frame format

Eac DANP s al multicast a PRP_ up rvision frame over b th its p rts with the format

sp cified in Ta le 4 (no VLAN) an Ta le 5 (with VLAN) resp ctively, every LifeCheckInterval

NOT 1 Octets with ofset 14to 17 are inserte o ly if a VLA ta a c rdin to IE E 8 2.1Q is use

NOT 2 Th frame h s a size of 7 o tets (n VLA ta ) or 7 o tets (with VLA ta ) to a oid p d in if a

e tity remo es th VLA ta or th RCT

NOT 3 Th Se Nr, L nId a d LSD size are re se wh n th frame is p s e to a HSR n twork Th

Su Eth rTy e is th same as th PRP sufix It is also use for HSR

Impleme tatio of v rsio X of th proto ol sh l interpret v rsio >X asif th y

were v rsio X, ig orin a y p rameters a d/or fla s a d d b th more re e t

v rsio , a d interpret v rsio < X PRP Su ervisio framese a tly as sp cifie

for th v rsio c n ern d

Su Se u n eNumb r Se u n e n mb r in reme te b 1 for e c su ervisio frame se t

TLV1.Ty e Op ratio mo e, a v lu of 2 in ic tes th t th n d su p rts th Du lc te

Disc rd or a v lu of 21 in ic tes th t it impleme ts Du lc te Ac e t

Oth r v lu sarereserv d

TLV1.L n th L n th of th folowin MA a dres in o tets

A RedBox, i.e a node actin as a prox for one or several SANs (caled VDAN or virtual DAN)

s al a p n to the TLV1 field a further TLV2 field with the content given in Ta le 7:

TLV2.Ty e Re Bo id ntifier with a v lu of 3 to in ic te th t th n d is a Re Bo

TLV2.L n th L n th of th folowin MA a dres in o tets

Re Bo Ma Ad res MA a dres of th Re Bo th t a ts as pro y for th oth r d vic (V A )

TLV0.Ty e Closin TLV, set to 0

TLV0.L n th Closin TLV, set to 0

4.3.4 Rece tion of a PRP_ up rvision frame and NodesTa le

When receivin a PRP_ up rvision frame over an LAN, a node s al cre te an entry in the

NodesTa le cor esp n in to the MacAd res of that source as in icated in the mes age

b d , not in the source ad res , with the Dupl cate Ac e t or Dupl cate Dis ard mode as

If a node ce ses to receive PRP_ up rvision frames from a source for a time lon er than

NodeForgetTime, but receives frames from that source over one LAN only, it s al c an e the

statu of this node to SanA, resp ctive SanB, de en in on the LAN from whic frames are

received

NOT 1 This rule alows mo in a SA b twe n LA _A a d LA _ , a d also to o tain th o eratin mo e for a

SA if it was first re istere at se din a d n t at re eivin ,sin e a D NP starts b se din o b th LA s

NOT 2 This rule alows distin uishin a SA from a D NP in Du lc te Ac e t mo e with o e ln

NOT 1 Th brid in n d setin su p rtsata hme t of a D NP to two brid es of th same LA to impleme t a

p rtial re u d n y to olo y Th re is n re uireme t th t n rmal frames b brid e in c se of a d u le faiure, b t

impleme ters are fre to in lu e this fe ture

NOT 2 Th prese c of n d s in brid in mo e is p te tialy d n ero s for a PRP n twork sin e it bre c es th

as umptio of fai-in e e d n e of th two LA s

4.5 Constants

The con tant p rameters are s own in Ta le 8

Other values may b defined at the u er’s resp n ibi ty

No eForg tTime Time afer whic a n d e try iscle re from th No esTa le

afer framesfrom this n d c ase to b re eiv d

Objects in the LRE, e.g er or cou ters, can b ac es ed from an a plcation ru nin on the

same node u in direct cal s or from an a pl cation on another node u in a management

protocol s c as SNMP or a sp cific protocol s c as IEC 618 0

Al LRE o jects are des rib d in Clau e 7 as a Management Information Base (MIB)

conformant to SNMP (RFC 2 7 , RFC 3 18), whic is common for PRP an HSR

5 High-availability Seamle s Re undanc (HSR)

5.1 HSR objectives

Clau e 5 des rib s the a pl cation of the PRP prin iples of Clau e 4 to implement a Hig

-avai a i ty Se mles Red n an y (HSR), retainin the PRP pro erty of zero recovery time,

NOT With resp ct to PRP, HSR alows to ro g ly h lv th n twork infrastru ture With resp ct to rin s b se

o IE E 8 2.1D (RS P), IEC 6 4 9-2 (MRP), IEC 6 4 9-6 (DRP) or IEC 6 4 9-7 (R P), th n twork trafic is

a o t th same

Nodes within the rin are restricted to b HSR-ca a le brid in nodes, th s avoidin the u e

of dedicated brid es Sin ly Atac ed Nodes (SANs) s c as la to s or printers can ot b

at ac ed directly to the rin , but ne d at ac ment throu h a Red n an y Box (RedBox)

5.2 HSR principle of operation

5.2.1 Basic op ration with a ring topology

As in PRP, a node has two p rts o erated in p ral el; it is a DANH (Doubly At ac ed Node

with HSR protocol)

A simple HSR network con ists of doubly atac ed brid in nodes, e c havin two rin

p rts, intercon ected by ful-d plex l nks, as s own in the example of Fig re 10 (multicast

an Fig re 1 (u icast for a rin to olog

Figure 10 – HSR example of ring config ration for multicast traf ic

A source DANH sen s a frame p s ed from its up er layers ( C” frame), prefixes it by an HSR

tag to identify frame d pl cates an sen s the frame over e c p rt ( A”- rame an “B”- rame)

A destination DANH receives, in the fault fre state, two identical frames from e c p rt within

a certain interval, removes the HSR tag of the first frame b fore p s in it to its up er layers

( D”- rame) an dis ard an d pl cate

The nodes s p ort the IEEE 8 2.1D brid e fu ctional ty an forward frames from one p rt to

the other, ac ordin to four rules:

1) a node wi not forward a frame that it injected into the rin or mes ;

2) a node wi not forward a frame for whic it is the u iq e destination (ex e t for sp cial

a pl cation s c as red n an y s p rvision);

3) a p rt wi not sen a frame that is a d pl cate of a frame that it alre d sent into that

same direction;

4) a p rt wi (o tional y) refrain from sen in a frame that is a d pl cate of a frame that it

alre d received from the o p site direction (ex e t for s p rvision an timin frames)

Figure 1 – HSR example of rin configuration for unicast traf ic

Frames circ latin in the rin car y the HSR tag in erted by the source, whic contain a

seq en e n mb r The doublet {source MAC ad res , seq en e n mb r} u iq ely identifies

co ies of the same frame

NOT 1 Th ma imum time sk w b twe n two frames of a p ir d p n s o th relativ p sitio of th re eivin

n d a d of th se din n d As umin a worst c se in whic e c n d in th rin is tra smitin at th same

time its own frame with th larg st size of 1 5 6 o tets (ma imum le gth su p rte b th Eth rTy e d fin d in

ISO/IEC/IE E 8 0 -2:19 8), e c n d c uld intro u e 12 µs of d la at 10 Mbits With 5 n d s, th time

sk w c n e c e 6 ms

NOT 2 Th minimum d la b fore a frame returns to th n d th t inje te it d p n s o th rin size

As umin a worst c se in whic th re is n oth r trafic a d in whic e c n d in th rin o erates in c t hro g

with 4 µs resid n e time, in a rin of 3 n d s, a frame with th ma imum size of 1 5 6 o tets (12 µs at

10 Mbits) returns to th inje tin n d afer 12 µs, i e b fore th inje tin n d finish d tra smitin it

5.2.2 DANH node stru ture

Fig re 12 s ows a con e tual view of the stru ture of a DANH implemented in hardware;

practical implementation can b diferent The two HSR p rts A an B an the device p rt C

are con ected by the LRE, whic in lu es a switc in matrix al owin to forward frames from

one p rt to the other The switc in matrix al ows c t throu h brid in The LRE presents to

the hig er layers the same interface as a stan ard Ethernet tran ceiver would do

The input circ it c ecks if this node is the destination of the frame an p s ibly do s VLAN

an multicast fi terin to oflo d the proces or The Dupl cate Dis ard is implemented in the

output q eues

Figure 12 – HSR structure of a DANH

5.2.3 Top log

5.2.3.1 At ac ment of singly at ac ed nodes

Sin ly atac ed nodes (SAN), for in tan e maintenan e la to s or printers can ot b in erted

directly into the rin sin e they have only one p rt an can ot interpret the HSR tag in the

frames SANs commu icate with rin devices throu h a RedBox that acts as a prox for the

SANs atac ed to it, as s own in Fig re 10 an Fig re 1 The RedBox is detai ed in 5.2.3.8

Con ectin non-HSR nodes to rin p rts, bre kin the rin , is alowed to ena le

config ration Non-HSR trafic within the closed rin is s p orted in an o tional mode

5.2.3.2 Use of HSR with se arate LANs

HSR nodes can b con ected in the same way as PRP nodes To this efect, an HSR node

may b set to “no forwardin ” (mode N) Unl k PRP, SANs can ot b at ac ed directly to

s c a d plcated network u les they are a le to interpret the HSR tag

Figure 13 – HSR example of top logy usin two inde endent networks

5.2.3.3 Pe r coupl n of rings

Two HSR rin s may b con ected by q adruple p rt devices with forwardin ca a i ties,

cal ed QuadBoxes, as Fig re 14 s ows This is ad antage u when the traf ic flow ex e d

the ca a i ties of a sin le rin However, tran mis ion delay from en to en are not

improved

Althou h one QuadBox is s ficient to con u t the trafic in the fault fre state of the network,

two QuadBoxes are u ed to prevent a sin le p int of fai ure

A Quadb x forward frames over e c rin as an HSR node, an p s es the frames

u c an ed to the other rin , ex e t if the frame can b identified as a frame not to b

forwarded to the other rin To this ef ect, a QuadBox is exp cted to fi ter trafic b sed for

in tan e on multicast fi terin or on VLAN fi terin There is no le rnin of MAC ad res es in

a QuadBox, sin e the le rnin of MAC ad res es on sp cific p rts of a QuadBox device could

le d to a s ort bre k in commu ication, if the QuadBox that has le rned an ad res an is

forwardin network trafic, fais

With QuadBoxes re l zed as sin le ph sical entities, the two intercon ected rin s s are the

same redu dan y domain con ernin fault toleran e If one QuadBox bre ks down, b th

intercon ected rin s are in a degraded state an can ot tolerate a further fault

Therefore, con tru tin QuadBoxes in the same way as a RedBox can help k e the

red n an y in e en ent The QuadBox then con ists of two devices con ected by an

interl nk For this re son, the RedBox sp cification in lu e the HSR con ection

Figure 14 – HSR example of pe r coupl ng of two rings

The presen e of two QuadBoxes on the same rin cau es that two co ies of the same frame

are tran fer ed from the first rin to the secon , e c generatin another two co ies

This do s not cau e four frames to circ late on the secon rin , sin e, when a co y from a

first QuadBox re c es the secon QuadBox on the same secon rin , the secon QuadBox

wi not forward it if it alre d sent a co y that came from its interl nk Con ersely, if the

secon QuadBox did not yet receive a co y from its interl nk, it wi forward the frame, but not

the co y that comes later from the interl nk

When a QuadBox receives a frame that it itself injected into the rin or a frame that the other

QuadBox in erted into the rin , it forward it to the interl nk an to its other p rt if it did not

alre d sent a co y This d plcate wi b dis arded at the other en of the interl nk This

s heme may cau e some ad itional trafic on the interlnk, but it al ows to simpl fy the desig