-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdraft-ietf-bier-te-arch-00.xml
1403 lines (1104 loc) · 60.6 KB
/
draft-ietf-bier-te-arch-00.xml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<?xml version="1.0" encoding="US-ASCII"?>
<!-- This template is for creating an Internet Draft using xml2rfc,
which is available here: http://xml.resource.org. -->
<!DOCTYPE rfc SYSTEM "rfc2629.dtd" [
<!-- One method to get references from the online citation libraries.
There has to be one entity for each item to be referenced.
An alternate method (rfc include) is described in the references. -->
<!ENTITY RFC2119 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2119.xml">
<!ENTITY RFC2629 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2629.xml">
<!ENTITY RFC3552 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3552.xml">
<!ENTITY RFC5226 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5226.xml">
<!ENTITY RFC8279 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8279.xml">
<!ENTITY RFC8296 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8296.xml">
]>
<?xml-stylesheet type='text/xsl' href='rfc2629.xslt' ?>
<!-- used by XSLT processors -->
<!-- For a complete list and description of processing instructions (PIs),
please see http://xml.resource.org/authoring/README.html. -->
<!-- Below are generally applicable Processing Instructions (PIs) that most I-Ds might want to use.
(Here they are set differently than their defaults in xml2rfc v1.32) -->
<?rfc strict="yes" ?>
<!-- give errors regarding ID-nits and DTD validation -->
<!-- control the table of contents (ToC) -->
<?rfc toc="yes"?>
<!-- generate a ToC -->
<?rfc tocdepth="4"?>
<!-- the number of levels of subsections in ToC. default: 3 -->
<!-- control references -->
<?rfc symrefs="yes"?>
<!-- use symbolic references tags, i.e, [RFC2119] instead of [1] -->
<?rfc sortrefs="yes" ?>
<!-- sort the reference entries alphabetically -->
<!-- control vertical white space
(using these PIs as follows is recommended by the RFC Editor) -->
<?rfc compact="yes" ?>
<!-- do not start each main section on a new page -->
<?rfc subcompact="no" ?>
<!-- keep one blank line between list items -->
<?rfc iprnotified="no" ?>
<!-- Change to "yes" if someone has disclosed IPR for the draft -->
<!-- end of list of popular I-D processing instructions -->
<rfc ipr="trust200902" docName="draft-ietf-bier-te-arch-00" category="exp">
<front>
<title abbrev="BIER-TE ARCH">Traffic Engineering for Bit Index Explicit Replication (BIER-TE)</title>
<author role="editor" fullname="Toerless Eckert" initials="T.T.E." surname="Eckert">
<organization abbrev="Huawei">Huawei USA - Futurewei Technologies Inc.</organization>
<address>
<postal>
<street>2330 Central Expy</street>
<city>Santa Clara</city>
<code>95050</code>
<country>USA</country>
</postal>
<email>[email protected]</email>
</address>
</author>
<author fullname="Gregory Cauchie" initials="G.C." surname="Cauchie">
<organization>Bouygues Telecom</organization>
<address>
<email>[email protected]</email>
</address>
</author>
<author fullname="Wolfgang Braun" initials="W.B." surname="Braun">
<organization>University of Tuebingen</organization>
<address>
<email>[email protected]</email>
</address>
</author>
<author fullname="Michael Menth" initials="M.M." surname="Menth">
<organization>University of Tuebingen</organization>
<address>
<email>[email protected]</email>
</address>
</author>
<date day="23" month="Jan" year="2018"/>
<abstract>
<t>
This document proposes an architecture for BIER-TE: Traffic
Engineering for Bit Index Explicit Replication (BIER). </t>
<t> BIER-TE shares part of its architecture with BIER as
described in <xref target="RFC8279"/>.
It also proposes to share the packet format with BIER.</t>
<t> BIER-TE forwards and replicates packets like BIER based on a
BitString in the packet header but it does not require an IGP.
It does support traffic engineering by explicit hop-by-hop forwarding
and loose hop forwarding of packets. It does support Fast ReRoute (FRR)
for link and node protection and incremental deployment. Because BIER-TE
like BIER operates without explicit in-network tree-building but also
supports traffic engineering, it is more similar to SR than RSVP-TE. </t>
</abstract>
</front>
<middle>
<section anchor="intro" title="Introduction">
<section anchor="overview" title="Overview">
<t> This document specifies the architecture for BIER-TE: traffic
engineering for Bit Index Explicit Replication BIER.</t>
<t> BIER-TE shares architecture and packet formats with BIER as
described in <xref target="RFC8279"/>. </t>
<t> BIER-TE forwards and replicates packets like BIER based on a BitString
in the packet header but it does not require an IGP. It does support
traffic engineering by explicit hop-by-hop forwarding and loose hop
forwarding of packets. It does support incremental deployment and a Fast
ReRoute (FRR) extension for link and node protection is given in
[I-D.eckert-bier-te-frr]. Because BIER-TE like BIER operates
without explicit in-network tree-building but also supports traffic
engineering, it is more similar to Segment Routing (SR) than RSVP-TE. </t>
<t> The key differences over BIER are: </t>
<t><list style="symbols">
<t> BIER-TE replaces in-network autonomous path calculation by explicit
paths calculated offpath by the BIER-TE controller host. </t>
<t> In BIER-TE every BitPosition of the BitString of a BIER-TE packet
indicates one or more adjacencies - instead of a BFER as
in BIER.</t>
<t> BIER-TE in each BFR has no routing table but only a BIER-TE Forwarding
Table (BIFT) indexed by SI:BitPosition and populated with only those
adjacencies to which the BFR should replicate packets to. </t>
</list></t>
<t>BIER-TE headers use the same format as BIER headers.</t>
<t>BIER-TE forwarding does not require/use the BFIR-ID. The BFIR-ID can
still be useful though for coordinated BFIR/BFER functions, such as
the context for upstream assigned labels for MPLS payloads in MVPN
over BIER-TE. </t>
<t>If the BIER-TE domain is also running BIER, then the BFIR-ID in
BIER-TE packets can be set to the same BFIR-ID as used with BIER
packets.</t>
<t>If the BIER-TE domain is not running full BIER or does not
want to reduce the need to allocate bits in BIER bitstrings for
BFIR-ID values, then the allocation of BFIR-ID values in BIER-TE packets can
be done through other mechanisms outside the scope of this document,
as long as this is appropriately agreed upon between all BFIR/BFER.</t>
<!-- <t> Currently, this specification has no considerations for BIER sub-domains.</t> -->
</section>
<!-- overview -->
<section anchor="requirements" title="Requirements Language">
<t>The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in <xref
target="RFC2119">RFC 2119</xref>.</t>
</section>
<!-- requirements -->
</section>
<!-- intro -->
<section anchor="layering" title="Layering">
<t>End to end BIER-TE operations consists of four components:
The "Multicast Flow Overlay", the "BIER-TE Controller Host",
the "Routing Underlay" and the "BIER-TE forwarding layer".</t>
<figure> <artwork align="left"><![CDATA[
Picture 2: Layers of BIER-TE
<------BGP/PIM----->
|<-IGMP/PIM-> multicast flow <-PIM/IGMP->|
overlay
[Bier-TE Controller Host]
^ ^ ^
/ | \ BIER-TE control protocol
| | | eg.: Netconf/Restconf/Yang
v v v
Src -> Rtr1 -> BFIR-----BFR-----BFER -> Rtr2 -> Rcvr
|--------------------->|
BIER-TE forwarding layer
|<- BIER-TE domain-->|
|<--------------------->|
Routing underlay
]]></artwork></figure>
<section anchor="flow-overlay" title="The Multicast Flow Overlay">
<t>The Multicast Flow Overlay operates as in BIER. See
<xref target="RFC8279"/>. Instead of
interacting with the BIER layer, it interacts with the BIER-TE Controller
Host</t>
</section>
<!-- flow-overlay -->
<section anchor="controller" title="The BIER-TE Controller Host">
<t>The BIER-TE controller host is representing the control plane of
BIER-TE. It communicates two sets of information with BFRs:</t>
<t>During bring-up or modifications of the network topology, the controller discovers
the network topology, assigns BitPositions to adjacencies and signals the resulting
mapping of BitPositions to adjacencies to each BFR connecting to the adjacency.</t>
<t>During day-to-day operations of the network, the controller signals to
BFIRs what multicast flows are mapped to what BitStrings.</t>
<t>Communications between the BIER-TE controller host to BFRs is ideally
via standardized protocols and data-models such as Netconf/Retconf/Yang.
This is currently outside the scope of this document. Vendor-specific CLI
on the BFRs is also a possible stopgap option (as in many other SDN solutions lacking
definition of standardized data model).</t>
<t>For simplicity, the procedures of the BIER-TE controller host are described in
this document as if it is a single, centralized automated entity, such as an SDN
controller. It could equally be an operator setting up CLI on the BFRs. Distribution
of the functions of the BIER-TE controller host is currently outside the scope of this
document.</t>
<section anchor="assignment" title="Assignment of BitPositions to adjacencies of the network topology">
<t>The BIER-TE controller host tracks the BFR topology of the
BIER-TE domain. It determines what adjacencies require
BitPositions so that BIER-TE explicit paths can be built
through them as desired by operator policy.</t>
<t>The controller then pushes the BitPositions/adjacencies to the BIFT of
the BFRs, populating only those SI:BitPositions to the BIFT of each
BFR to which that BFR should be able to send packets to - adjacencies
connecting to this BFR.</t>
</section>
<!-- assignment -->
<section anchor="changes-in-topo" title="Changes in the network topology">
<t>If the network topology changes (not failure based) so that adjacencies
that are assigned to BitPositions are no longer needed, the controller can
re-use those BitPositions for new adjacencies. First, these BitPositions
need to be removed from any BFIR flow state and BFR BIFT state, then they
can be repopulated, first into BIFT and then into the BFIR.</t>
</section>
<!-- changes-in-topo -->
<section anchor="setup" title="Set up per-multicast flow BIER-TE state">
<t>The BIER-TE controller host tracks the multicast flow overlay
to determine what multicast flow needs to be sent by a BFIR
to which set of BFER. It calculates the desired distribution
tree across the BIER-TE domain based on algorithms outside the
scope of this document (eg.: CSFP, Steiner Tree,...). It then
pushes the calculated BitString into the BFIR.</t>
</section>
<!-- setup -->
<section anchor="failures" title="Link/Node Failures and Recovery">
<t>When link or nodes fail or recover in the topology, BIER-TE can quickly
respond with the optional FRR procedures described in
[I-D.eckert-bier-te-frr]. It can also more slowly react by
recalculating the BitStrings of affected multicast flows. This reaction is
slower than the FRR procedure because the controller needs to receive
link/node up/down indications, recalculate the desired BitStrings and push
them down into the BFIRs. With FRR, this is all performed locally on a BFR
receiving the adjacency up/down notification.</t>
</section>
<!-- failures -->
</section>
<!-- controller -->
<section anchor="forwarding-layer" title="The BIER-TE Forwarding Layer">
<t>When the BIER-TE Forwarding Layer receives a packet, it simply looks
up the BitPositions that are set in the BitString of the packet in the
Bit Index Forwarding Table (BIFT) that was populated by the BIER-TE controller
host. For every BP that is set in the BitString, and that has one or
more adjacencies in the BIFT, a copy is made according to the type
of adjacencies for that BP in the BIFT. Before sending any copy, the
BFR resets all BitPositions in the BitString of the packet to which it
can create a copy. This is done to inhibit that packets can loop.</t>
</section>
<!-- forwarding-layer -->
<section anchor="routing-underlay" title="The Routing Underlay">
<t>BIER-TE is sending BIER packets to directly connected
BIER-TE neighbors as L2 (unicasted) BIER packets without requiring a
routing underlay. BIER-TE forwarding uses the Routing underlay for
forward_routed adjacencies which copy BIER-TE packets to not-directly-connected
BFRs (see below for adjacency definitions).
</t>
<t>If the BFR intends to support FRR for BIER-TE, then the BIER-TE
forwarding plane needs to receive fast adjacency up/down notifications:
Link up/down or neighbor up/down, eg.: from BFD. Providing these notifications
is considered to be part of the routing underlay in this document.</t>
</section>
<!-- routing-underlay -->
</section>
<!-- layering -->
<section anchor="forwarding" title="BIER-TE Forwarding">
<section anchor="btft" title="The Bit Index Forwarding Table (BIFT)">
<t>The Bit Index Forwarding Table (BIFT) exists in every BFR. For every
subdomain in use, it is a table indexed by SI:BitPosition and is populated by the
BIER-TE control plane. Each index can be empty or contain a list of one or more
adjacencies.</t>
<t>BIER-TE can support multiple subdomains like BIER. Each one with a separate BIFT</t>
<t>In the BIER architecture, indices into the BIFT are explained to be both
BFR-id and SI:BitString (BitPosition). This is because there is a 1:1 relationship
between BFR-id and SI:BitString - every bit in every SI is/can be assigned to
a BFIR/BFER. In BIER-TE there are more bits used in each BitString than there are
BFIR/BFER assigned to the bitstring. This is because of the bits required to express
the (traffic engineered) path through the topology. The BIER-TE forwarding definitions
do therefore not use the term BFR-id at all. Instead, BFR-ids are only used as required
by routing underlay, flow overlay of BIER headers. Please refer to <xref target="mgmt-stuff"/>
for explanations how to deal with SI, subdomains and BFR-id in BIER-TE.</t>
<figure> <artwork align="left"><![CDATA[
------------------------------------------------------------------
| Index: | Adjacencies: |
| SI:BitPosition | <empty> or one or more per entry |
==================================================================
| 0:1 | forward_connected(interface,neighbor,DNR) |
------------------------------------------------------------------
| 0:2 | forward_connected(interface,neighbor,DNR) |
| | forward_connected(interface,neighbor,DNR) |
------------------------------------------------------------------
| 0:3 | local_decap([VRF]) |
------------------------------------------------------------------
| 0:4 | forward_routed([VRF,]l3-neighbor) |
------------------------------------------------------------------
| 0:5 | <empty> |
------------------------------------------------------------------
| 0:6 | ECMP({adjacency1,...adjacencyN}, seed) |
------------------------------------------------------------------
...
| BitStringLength | ... |
------------------------------------------------------------------
Bit Index Forwarding Table
]]></artwork></figure>
<t>The BIFT is programmed into the data plane of BFRs by the BIER-TE
controller host and used to forward packets, according to the rules
specified in the BIER-TE Forwarding Procedures.</t>
<t>Adjacencies for the same BP when populated in more than one BFR
by the controller do not have to have the same adjacencies. This is
up to the controller. BPs for p2p links are one case (see below).</t>
</section>
<!-- btft -->
<section anchor="atypes" title="Adjacency Types">
<section anchor="forward-connected" title="Forward Connected">
<t>A "forward_connected" adjacency is towards a directly connected
BFR neighbor using an interface address of that BFR on the connecting
interface. A forward_connected adjacency does not route packets
but only L2 forwards them to the neighbor.</t>
<t>Packets sent to an adjacency with "DoNotReset" (DNR) set in the
BIFT will not have the BitPosition for that adjacency reset when the
BFR creates a copy for it. The BitPosition will still be reset for
copies of the packet made towards other adjacencies. The can be
used for example in ring topologies as explained below.</t>
</section>
<!-- forward-connected -->
<section anchor="forward-routed" title="Forward Routed">
<t>A "forward_routed" adjacency is an adjacency towards a BFR that
is not a forward_connected adjacency: towards a loopback address
of a BFR or towards an interface address that is non-directly
connected. Forward_routed packets are forwarded via the Routing
Underlay.</t>
<t>If the Routing Underlay has multiple
paths for a forward_routed adjacency, it will perform ECMP independent
of BIER-TE for packets forwarded across a forward_routed adjacency. </t>
<t>If the Routing Underlay has FRR, it will perform FRR independent
of BIER-TE for packets forwarded across a forward_routed adjacency.</t>
</section>
<!-- forward-routed -->
<section anchor="forward-ecmp" title="ECMP">
<t>The ECMP mechanisms in BIER are tied to the BIER BIFT and are are therefore
not directly useable with BIER-TE. The following procedures describe ECMP
for BIER-TE that we consider to be lightweight but also well manageable.
It leverages the existing entropy parameter in the BIER header to keep
packets of the flows on the same path and it introduces a "seed" parameter
to allow engineering traffic to be polarized or randomized across multiple
hops.</t>
<t>An "Equal Cost Multipath" (ECMP) adjacency has a list of two or
more adjacencies included in it. It copies the BIER-TE to
one of those adjacencies based on the ECMP hash calculation.
The BIER-TE ECMP hash algorithm must select the same adjacency
from that list for all packets with the same "entropy" value in
the BIER-TE header if the same number of
adjacencies and same seed are given as parameters. Further use of the
seed parameter is explained below.</t>
</section>
<!-- forward-ecmp -->
<section anchor="forward-local" title="Local Decap">
<t>A "local_decap" adjacency passes a copy of the payload of
the BIER-TE packet to the packets NextProto within the BFR (IPv4/IPv6, Ethernet,...).
A local_decap adjacency turns the BFR into a BFER for matching
packets. Local_decap adjacencies require the BFER to support
routing or switching for NextProto to determine how to further
process the packet.</t>
</section>
<!-- forward-local -->
</section>
<!-- atypes -->
<section anchor="encapsulation" title="Encapsulation considerations">
<t>Specifications for BIER-TE encapsulation are outside the scope of this document.
This section gives explanations and guidelines.</t>
<t>Because a BFR needs to interpret the BitString of a BIER-TE packet differently
from a BIER packet, it is necessary to distinguish BIER from BIER-TE packets. This
is subject to definitions in BIER encapsulation specifications.</t>
<t>MPLS encapsulation <xref target="RFC8296"/> for
example assigns one label by which BFRs recognizes BIER packets for every
(SI,subdomain) combination. If it is desirable that every subdomain can
forward only BIER or BIER-TE packets, then the label allocation could stay the
same, and only the forwarding model (BIER/BIER-TE) would have to be defined
per subdomain. If it is desirable to support both BIER and BIER-TE forwarding
in the same subdomain, then additional labels would need to be assigned for
BIER-TE forwarding.</t>
<t>"forward_routed" requires an encapsulation permitting to unicast BIER-TE packets
to a specific interface address on a target BFR. With MPLS encapsulation, this can
simply be done via a label stack with that addresses label as the top label - followed
by the label assigned to (SI,subdomain) - and if necessary (see above) BIER-TE.
With non-MPLS encapsulation, some form of IP tunneling (IP in IP, LISP, GRE) would
be required.</t>
<t>The encapsulation used for "forward_routed" adjacencies can equally support
existing advanced adjacency information such as "loose source routes" via eg: MPLS
label stacks or appropriate header extensions (eg: for IPv6).</t>
</section>
<!-- encapsulation -->
<section anchor="basic" title="Basic BIER-TE Forwarding Example">
<t>Step by step example of basic BIER-TE forwarding. This does not
use ECMP or forward_routed adjacencies nor does it try to minimize
the number of required BitPositions for the topology.</t>
<figure> <artwork align="left"><![CDATA[
Picture 1: Forwarding Example
[Bier-Te Controller Host]
/ | \
v v v
| p13 p1 |
+- BFIR2 --+ |
| | p2 p6 | LAN2
| +-- BFR3 --+ |
| | | p7 p11 |
Src -+ +-- BFER1 --+
| | p3 p8 | |
| +-- BFR4 --+ +-- Rcv1
| | | |
| |
| p14 p4 |
+- BFIR1 --+ |
| +-- BFR5 --+ p10 p12 |
LAN1 | p5 p9 +-- BFER2 --+
| +-- Rcv2
|
LAN3
IP |..... BIER-TE network......| IP
]]></artwork></figure>
<t>pXX indicate the BitPositions number
assigned by the BIER-TE controller host to adjacencies in the
BIER-TE topology. For example, p9 is the adjacency towards BFR9
on the LAN connecting to BFER2.</t>
<figure> <artwork align="left"><![CDATA[
BIFT BFIR2:
p13: local_decap()
p2: forward_connected(BFR3)
BIFT BFR3:
p1: forward_connected(BFIR2)
p7: forward_connected(BFER1)
p8: forward_connected(BFR4)
BIFT BFER1:
p11: local_decap()
p6: forward_connected(BFR3)
p8: forward_connected(BFR4)
]]></artwork></figure>
<t>...and so on.</t>
<t>Traffic needs to flow from BFIR2 towards Rcv1, Rcv2.
The controller determines it wants it to pass across
the following paths:</t>
<figure> <artwork align="left"><![CDATA[
-> BFER1 ---------------> Rcv1
BFIR2 -> BFR3
-> BFR4 -> BFR5 -> BFER2 -> Rcv2
]]></artwork></figure>
<t>These paths equal to the following BitString:
p2, p5, p7, p8, p10, p11, p12.</t>
<t>This BitString is set up in BFIR2. Multicast packets
arriving at BFIR2 from Src are assigned this BitString.</t>
<t>BFIR2 forwards based on that BitString.
It has p2 and p13 populated. Only p13 is in BitString
which has an adjacency towards BFR3. BFIR2 resets p2 in BitString
and sends a copy towards BFR2.</t>
<t>BFR3 sees a BitString of p5,p7,p8,p10,p11,p12.
It is only interested in p1,p7,p8. It creates a copy of the
packet to BFER1 (due to p7) and one to BFR4 (due to p8). It
resets p7, p8 before sending.</t>
<t>BFER1 sees a BitString of p5,p10,p11,p12.
It is only interested in p6,p7,p8,p11 and therefore considers
only p11. p11 is a "local_decap" adjacency installed
by the BIER-TE controller host because BFER1 should pass
packets to IP multicast. The local_decap adjacency instructs
BFER1 to create a copy, decapsulate it from the BIER header
and pass it on to the NextProtocol, in this example IP multicast.
IP multicast will then forward the packet out to LAN2 because
it did receive PIM or IGMP joins on LAN2 for the traffic. </t>
<t>Further processing of the packet in BFR4, BFR5 and BFER2
accordingly.</t>
</section>
<!-- basic -->
<section anchor="fwd-comparison" title="Forwarding comparison with BIER">
<t>Forwarding of BIER-TE is designed to allow common forwarding hardware
with BIER. Like BIER, the core of BIER-TE forwarding are BIFTs with bitstring size
number of entries: One for each bit of the bitstring in the processed packet
(consider that 256 is the most common size).</t>
<t>When a packet is received, the BIFT to process needs to be selected.
This is based on SI and subdomain like in BIER. How SI and subdomain are
indicated is subject to the BIER-TE encapsulation, but not BIER-T itself.
It is expected that the mechanisms for encapsulation will be very similar
if not the same to BIER, but this is subject to followup work.</t>
<t>There are some key difference between the BIFT in BIER and BIER-TE:</t>
<t>In BIER-TE, each entry in the BIFT can have a list of 0 or more adjacencies.
A separate copy of the packet is made for each adjacency. In BIER, each BIFT
entry has at most one adjacency (BFR-NBR). In BIER, different bits
can not be processed independently directly: Only one packet copy is
to be sent for all bits in the packet with the same adjacency, which is
why the forwarding procedure specifies how to sequentially identify
those bits and avoid duplication. In BIER-TE there are no mutual dependencies
between bit adjacencies, so all bits of a BIER-TE bitstring could be
procssed independently in parallel.</t>
<t>In BIER the BIFT has adjacencies for all BFR-ids assigned to BFER
and reachable in the IGP. In BIER-TE the BIFT only has adjacencies for
bits that are adjacent hops - intermediate or BFER. In forwarding, this
can be treated via the same lookup logic except that in BIER-TE there
is no step modifyin the original packet and the packet copy
bitstring with the FBM. Instead, all the bits locally processed are
reset in the original packet before looking up bits in the BIFT (~MyBitsOfInterest).
Only for an adjacency with the "DNR" (Do Not Reset) bit set would the
bit in the bitstring not be set again as part of processing of the
adjacency.</t>
<t>In summary, implementations of BIER forwarding that are to be extended
to also support BIER-TE forwarding primarily need to consider how they
can ensure that individual bit lookups can result in a sequence of
more than one copy to be made (as opposed to one in BIER), and they need
to see that they can accordingly reset bits in the bitstring differently
for BIER (per-packet) vs. BIER-TE (per-paket-copy).</t>
</section>
</section>
<!-- forwarding -->
<section anchor="bitpositions" title="BIER-TE Controller Host BitPosition Assignments">
<t>This section describes how the BIER-TE controller host can use the
different BIER-TE adjacency types to define the BitPositions of a BIER-TE domain.</t>
<t>Because the size of the BitString is limiting the size of the
BIER-TE domain, many of the options described exist to support larger
topologies with fewer BitPositions (4.1, 4.3, 4.4, 4.5, 4.6, 4.7, 4.8).</t>
<section anchor="p2p-links" title="P2P Links">
<t>Each P2p link in the BIER-TE domain is assigned one unique BitPosition
with a forward_connected adjacency pointing to the neighbor on the
p2p link.</t>
</section>
<!-- p2p-links -->
<section anchor="bfer" title="BFER">
<t>Every BFER is given a unique BitPosition with a local_decap adjacency.</t>
</section>
<!-- bfer -->
<section anchor="bfirs" title="Leaf BFERs">
<t>Leaf BFERs are BFERs where incoming BIER-TE packets never need to
be forwarded to another BFR but are only sent to the BFER
to exit the BIER-TE domain. For example, in networks where PEs
are spokes connected to P routers, those PEs are Leaf BFIRs unless
there is a U-turn between two PEs.</t>
<t>All leaf-BFER in a BIER-TE domain can share a single BitPosition.
This is possible because the BitPosition for the adjacency to reach the BFER
can be used to distinguish whether or not packets should reach the BFER.</t>
<t>This optimization will not work if an upstream interface of the BFER
is using a BitPosition optimized as described in the following two
sections (LAN, Hub and Spoke).</t>
</section>
<!-- bfirs -->
<section anchor="lans" title="LANs">
<t>In a LAN, the adjacency to each neighboring BFR on the LAN
is given a unique BitPosition. The adjacency of this BitPosition
is a forward_connected adjacency towards the BFR and this BitPosition
is populated into the BIFT of all the other BFRs on that LAN.</t>
<figure> <artwork align="left"><![CDATA[
BFR1
|p1
LAN1-+-+---+-----+
p3| p4| p2|
BFR3 BFR4 BFR7
]]></artwork></figure>
<t>If Bandwidth on the LAN is not an issue and most BIER-TE traffic
should be copied to all neighbors on a LAN, then BitPositions
can be saved by assigning just a single BitPosition to the LAN
and populating the BitPosition of the BIFTs of each BFRs on
the LAN with a list of forward_connected adjacencies to all other
neighbors on the LAN.</t>
<t>This optimization does not work in the face of BFRs redundantly
connected to more than one LANs with this optimization because
these BFRs would receive duplicates and forward those duplicates into
the opposite LANs. Adjacencies of such BFRs into their LANs still
need a separate BitPosition.</t>
</section>
<!-- lans -->
<section anchor="hubnspoke" title="Hub and Spoke">
<t>In a setup with a hub and multiple spokes connected via separate
p2p links to the hub, all p2p links can share the same BitPosition.
The BitPosition on the hubs BIFT is set up with a list of
forward_connected adjacencies, one for each Spoke.</t>
<t>This option is similar to the BitPosition optimization in
LANs: Redundantly connected spokes need their own BitPositions.</t>
</section>
<!-- hubnspoke -->
<section anchor="rings" title="Rings">
<t>In L3 rings, instead of assigning a single BitPosition for
every p2p link in the ring, it is possible to save BitPositions by
setting the "Do Not Reset" (DNR) flag on forward_connected adjacencies.</t>
<t>For the rings shown in the following picture, a single BitPosition
will suffice to forward traffic entering the ring at BFRa or BFRb
all the way up to BFR1:</t>
<t>On BFRa, BFRb, BFR30,... BFR3, the BitPosition is populated with
a forward_connected adjacency pointing to the clockwise neighbor
on the ring and with DNR set. On BFR2, the adjacency also points
to the clockwise neighbor BFR1, but without DNR set.</t>
<t>Handling DNR this way ensures that copies forwarded from any BFR in
the ring to a BFR outside the ring will not have the ring BitPosition set,
therefore minimizing the chance to create loops.</t>
<figure> <artwork align="left"><![CDATA[
v v
| |
L1 | L2 | L3
/-------- BFRa ---- BFRb --------------------\
| |
\- BFR1 - BFR2 - BFR3 - ... - BFR29 - BFR30 -/
| | L4 | |
p33| p15|
BFRd BFRc
]]></artwork></figure>
<t>Note that this example only permits for packets to enter the ring at
BFRa and BFRb, and that packets will always travel clockwise. If
packets should be allowed to enter the ring at any ring BFR, then one
would have to use two ring BitPositions. One for clockwise, one for
counterclockwise.</t>
<t>Both would be set up to stop rotating on the same link, eg: L1. When the
ingress ring BFR creates the clockwise copy, it will reset the counterclockwise
BitPosition because the DNR bit only applies to the bit for which the
replication is done. Likewise for the clockwise
BitPosition for the counterclockwise copy. In result, the ring ingress
BFR will send a copy in both directions, serving BFRs on either side of the
ring up to L1.</t>
</section>
<!-- rings -->
<section anchor="ecmp" title="Equal Cost MultiPath (ECMP)">
<t>The ECMP adjacency allows to use just one BP per link
bundle between two BFRs instead of one BP for each p2p member
link of that link bundle. In the following picture, one BP
is used across L1,L2,L3 and BFR1/BFR2 have for the BP</t>
<figure> <artwork align="left"><![CDATA[
--L1-----
BFR1 --L2----- BFR2
--L3-----
BIFT entry in BFR1:
------------------------------------------------------------------
| Index | Adjacencies |
==================================================================
| 0:6 | ECMP({L1-to-BFR2,L2-to-BFR2,L3-to-BFR2}, seed) |
------------------------------------------------------------------
BIFT entry in BFR2:
------------------------------------------------------------------
| Index | Adjacencies |
==================================================================
| 0:6 | ECMP({L1-to-BFR1,L2-to-BFR1,L3-to-BFR1}, seed) |
------------------------------------------------------------------
]]></artwork></figure>
<t>In the following example, all traffic from BFR1 towards BFR10 is
intended to be ECMP load split equally across the topology. This
example is not mean as a likely setup, but to illustrate that ECMP can
be used to share BPs not only across link bundles, and it explains
the use of the seed parameter.</t>
<figure> <artwork align="left"><![CDATA[
BFR1
/ \
/L11 \L12
BFR2 BFR3
/ \ / \
/L21 \L22 /L31 \L32
BFR4 BFR5 BFR6 BFR7
\ / \ /
\ / \ /
BFR8 BFR9
\ /
\ /
BFR10
BIFT entry in BFR1:
------------------------------------------------------------------
| 0:6 | ECMP({L11-to-BFR2,L12-to-BFR3}, seed) |
------------------------------------------------------------------
BIFT entry in BFR2:
------------------------------------------------------------------
| 0:6 | ECMP({L21-to-BFR4,L22-to-BFR5}, seed) |
------------------------------------------------------------------
BIFT entry in BFR3:
------------------------------------------------------------------
| 0:6 | ECMP({L31-to-BFR6,L32-to-BFR7}, seed) |
------------------------------------------------------------------
]]></artwork></figure>
<t> With the setup of ECMP in above topology, traffic would not be
equally load-split. Instead, links L22 and L31 would see no traffic
at all: BFR2 will only see traffic from BFR1 for which the ECMP
hash in BFR1 selected the first adjacency in a list of 2 adjacencies:
link L11-to-BFR2. When forwarding in BFR2 performs again an ECMP
with two adjacencies on that subset of traffic, then it will
again select the first of its two adjacencies to it: L21-to-BFR4. And
therefore L22 and BFR5 sees no traffic. </t>
<t>To resolve this issue, the ECMP adjacency on BFR1 simply needs to
be set up with a different seed than the ECMP adjacencies on BFR2/BFR3</t>
<t>This issue is called polarization. It depends on the
ECMP hash. It is possible to build ECMP that does not have
polarization, for example by taking entropy from the actual
adjacency members into account, but that can make it harder to
achieve evenly balanced load-splitting on all BFR without making
the ECMP hash algorithm potentially too complex for fast forwarding
in the BFRs.</t>
</section>
<!-- ecmp -->
<section anchor="routed" title="Routed adjacencies">
<section anchor="reducing" title="Reducing BitPositions">
<t>Routed adjacencies can reduce the number of BitPositions
required when the traffic engineering requirement is not hop-by-hop
explicit path selection, but loose-hop selection.</t>
<figure> <artwork align="left"><![CDATA[
............... ...............
BFR1--... Redundant ...--L1-- BFR2... Redundant ...---
\--... Network ...--L2--/ ... Network ...---
BFR4--... Segment 1 ...--L3-- BFR3... Segment 2 ...---
............... ...............
]]></artwork></figure>
<t>Assume the requirement in above network is to explicitly engineer
paths such that specific traffic flows are passed from segment 1
to segment 2 via link L1 (or via L2 or via L3).</t>
<t>To achieve this, BFR1 and BFR4 are set up with a forward_routed
adjacency BitPosition towards an address of BFR2 on link L1
(or link L2 BFR3 via L3).</t>
<t>For paths to be engineered through a specific node BFR2 (or BFR3),
BFR1 and BFR4 are set up up with a forward_routed adjacency BitPosition
towards a loopback address of BFR2 (or BFR3).</t>
</section>
<!-- reducing -->
<section anchor="without" title="Supporting nodes without BIER-TE">
<t>Routed adjacencies also enable incremental deployment of BIER-TE.
Only the nodes through which BIER-TE traffic needs to be steered -
with or without replication - need to support BIER-TE. Where
they are not directly connected to each other, forward_routed
adjacencies are used to pass over non BIER-TE enabled nodes.</t>
</section>
<!-- without -->
</section>
<!-- routed -->
</section>
<!-- bitpositions -->
<section anchor="avoiding" title="Avoiding loops and duplicates">
<section anchor="loops" title="Loops">
<t>Whenever BIER-TE creates a copy of a packet, the BitString of
that copy will have all BitPositions cleared that are associated
with adjacencies in the BFR. This inhibits looping of packets.
The only exception are adjacencies with DNR set.</t>
<t>With DNR set, looping can happen. Consider in the ring picture
that link L4 from BFR3 is plugged into the L1 interface of
BFRa. This creates a loop where the rings clockwise BitPosition is
never reset for copies of the packets traveling clockwise
around the ring.</t>
<t>To inhibit looping in the face of such physical misconfiguration,
only forward_connected adjacencies are permitted to have DNR set,
and the link layer destination address of the adjacency (eg.: MAC address)
protects against closing the loop. Link layers without port unique
link layer addresses should not used with the DNR flag set.</t>
</section>
<!-- loops -->
<section anchor="duplicates" title="Duplicates">
<t>Duplicates happen when the topology of the BitString is not a
tree but redundantly connecting BFRs with each other. The controller
must therefore ensure to only create BitStrings that are trees in
the topology.</t>
<t>When links are incorrectly physically re-connected before the
controller updates BitStrings in BFIRs, duplicates can happen.
Like loops, these can be inhibited by link layer addressing
in forward_connected adjacencies.</t>
<t>If interface or loopback addresses used in forward_routed adjacencies
are moved from one BFR to another, duplicates can equally happen.
Such re-addressing operations must be coordinated with the controller.</t>
</section>
<!-- duplicates -->
</section>
<!-- avoiding -->
<section anchor="pseudocode" title="BIER-TE Forwarding Pseudocode">
<t>The following sections of Pseudocode are meant to illustrate the
BIER-TE forwarding plane. This code is not meant to be normative
but to serve both as a potentially easier to read and more precise
representation of the forwarding functionality and to illustrate
how simple BIER-TE forwarding is and that it can be efficiently
be implemented.</t>
<t>The following procedure is executed on a BFR whenever the BIFT is
changed by the BIER-TE controller host:</t>
<figure> <artwork align="left"><![CDATA[
global MyBitsOfInterest
void BIFTChanged()
{
for (Index = 0; Index++ ; Index <= BitStringLength)
if(BIFT[Index] != <empty>)
MyBitsOfInterest != 2<<(Index-1)
}
]]></artwork></figure>
<t>The following procedure is executed whenever a BIER-TE
packet is to be forwarded:</t>
<figure> <artwork align="left"><![CDATA[
void ForwardBierTePacket (Packet)
{
// We calculate in BitMask the subset of BPs of the BitString
// for which we have adjacencies. This is purely an
// optimization to avoid to replicate for every BP
// set in BitString only to discover that for most of them,
// the BIFT has no adjacency.
local BitMask = Packet->BitString
Packet->BitString &= ~MyBitsOfInterest
BitMask &= MyBitsOfInterest
// Replication
for (Index = GetFirstBitPosition(BitMask); Index ;
Index = GetNextBitPosition(BitMask, Index))
foreach adjacency BIFT[Index]
if(adjacency == ECMP(ListOfAdjacencies, seed) )
I = ECMP_hash(sizeof(ListOfAdjacencies),
Packet->Entropy, seed)
adjacency = ListOfAdjacencies[I]
PacketCopy = Copy(Packet)
switch(adjacency)
case forward_connected(interface,neighbor,DNR):
if(DNR)
PacketCopy->BitString |= 2<<(Index-1)