Versions Compared

Old Version 33

changes.mady.by.user Māris B.

Saved on

compared with

New Version Current

changes.mady.by.user Māris B.

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Table of Contents
Feature Status

N/A - Feature not yet available

...

  • Yellow - partially working
  • Green - Working
  • Red - Not working at the moment

Feature

v7.1v7.2v7.
0Beta7
3v7.
1Beta1
6v7.
1Beta2
10v7.
1Beta3
12v7.
1Beta4
14

v7.

1Beta7

15

Winbox        
BGP support







OSPF support







RIP support







Router ID support







Routing filter support
















Generic        
/31 address supportN/A
Convert route rules

Routed traffic does not work to odd address.




Convert route rules after upgrade from v6.x
N/AOK








Static IPv6 upgrade from ROS v6
NOKOK








IPv4 Route Rules
OK








IPv6 Route Rules
NOK
OK








ECMP flags
OK








dst@table
OK








gateway@table
OK








gateway%interface
OK








recursive route over ipv6 LL address
OK








3 level recursive gateway with ECMP 
OK








IPV6 ECMP
OK








IPv6 connected ECMP

N/A

OK








Addresses from same subnet to multiple interfaces
NOK
N/A






Show time when route was last updatedN/A






Check Gateway
Ping OK, ARP NOK (works as ping), BFD NOK (not implemented)
BFD not ready






Scope and target scope
OK








IPv4 Mangle routing-mark
OK








IPv6 Mangle routing-mark
NOK (Added but does not work at the moment)OK








Packet SRC address

NOK (wrong src is picked for locally originated packets)

Does not work correctly with /32 addresses






Routing-table parameter for ping and telnet
NOK








Show if route is hardware accelerated
N/A
Shows if route is candidate for HW acceleration






Custom route selection policy 
N/A








IPv4 with IPv6 nexthops
and RFC5549N/AOK
for RFC5549
















Routing
Instance
id
OKRenamed to Router-IDVRFNeeds more testing








VRF







Management services support for VRFs telnet, ssh, api, www services can be set to listen on specific VRF






Some kind of mechanism to import/export routes from one vrf to another within same routerN/A






BFDN/A


Initial support











OSPF        
Convert OSPF config from v6 to v7 after upgrade
N/AOSPF neighbours in in NSSA AreaCannot establish adjacency

Known conversion problems:

  • NBMA neighbors place in backbone

  • ospf-v2 networks + interface may have issues

  • dynamic interfaces may have issues

  • MPLS PE CE features are not converted








OSPF neighbors in NSSA Area







OSPF in broadcast network
Does not work with more than 2 neighboursOK








OSPF with routing filters
OK








OSPF Virtual Link
OK








OPSF input filtering







HMAC-SHA auth RFC5709
BGP and OSPF SNMP monitoring
N/A
BGP


Initial support



OSPF SNMP monitoringN/A






BGP SNMP monitoring



For ipv4 sessions











IS-IS







IPv4




Initial support

IPv6







Traffic Engineering
















BGP        
Convert BGP config from v6 to v7 after upgrade
N/A
OK








BGP Templates and dynamic peers
OK








BGP connect listen on a network
OK








BGP guess remote.as
OKShow








Show from which peer route receivedOK ( /routing/route/print detail --> belongs-to)






BGP Address Families
Currently only IPv4, IPv6








BGP input.accept-*
Does not work right after peer established, need to resend routes








eBGP nexthop self
OK








Input Filter
OK








Output Filter
OK








BGP Local address auto selection
OK








BGP route reflect
OK








BGP route server
OK








BGP Roles
https://datatracker.ietf.org/doc/draft-ietf-idr-bgp-open-policy/?include_text=1
OK
rfc roles not fully implemented






BGP
peer-cache
session uptime in "established" state
N/A








BGP session last established time
















BGP Flow SpecFlow spec attributes are forwarded






BGP Selection
OK

















BGP Selection (Multipath)N/A






BGP Confederation
OK








BGP AggregationN/A






BGP ORFN/A






Discard prefix RTBH 
rfc
RFC 6666N/A






AS-wide Unique BGP Identifier RFC 6286N/A
MPLSStatic label mappingLDPN/ALDP signaled VPLSTEBGP signaled VPLSRPKI sessionOKRPKI possibility to view received info of specific prefixN/AOKRPKI show connection statusN/AFiltersConvert routing filters after upgrade from v6.xN/ARouting filter chain drop by default without rulesOKRouting filter prefix match

OK

Routing filter protocol matchMissing other protocols like modem DHCP etc (these routes are not matched by "static")Routing filter append communitiesOKRouting filter append large communityOKRouting filter set weight"sub, add" does not work if attribute has default unset value, if attribute is not set then default value should be assumed.







Exported PDU PCAP saver







Exported PDU PCAP loader







BGP Advertisement monitoring
Initial implementation by dumping to pcap
Advertisements rework



BGP Prefix limit

Initial support




BGP advertise IPv4 prefix with IPv6 nexthop (RFC5549)







BGP VPNv6 support



Prerequisites are made, need to add actual BGP Afi











MPLS        
Static label mapping







Static mapping upgrade from v6







LDP IPv4 mapping







LDP IPv6 mapping







LDP signaled VPLS







LDP config upgrade from v6







LDP Dual Stack







TE







TE Config upgrade from v6







VPLS Encap to TE







BGP signaled VPLS







VPLS config upgrade from v6







Fast reroute







MPLS ECMP







One label per VRF







Ability to use MPLS EXP-bit in QueuesN/A






MPLS Fast-PathN/A















RPKI session







RPKI possibility to view received info of specific prefix







RPKI show connection status
















Filters        
Convert routing filters after upgrade from v6.x







Syntax completion







Routing filter chain drop by default without rules







Routing filter prefix match







Routing filter protocol match







Routing filter append communities







Routing filter append large community







Routing filter set weight







Routing filter set local pref







Routing filter set MED







Routing filter set origin







Routing filter set igp metric from OSPF cost
Routing filter set local pref"sub, add" does not work if attribute has default unset value, default value should be 100 if not set.Routing filter set MED"sub,add" should assume default value. Routing filter set originOKRouting filter set igp metric from OSPF cost Can be set with set-num-prop=bgp-igp-metric<add>ospf-ext-metric or ospf-metric, but what to do if it is not possible to guess which metric to use?








Routing filter match prefix with address list
OK








Routing filter match community/large community lists
OK








Routing filter add a prefix to address listN/A






Routing filter validate
prefix with RPKIOK
prefix with RPKI
















Multicast        
IGMP-Proxy







PIM-SMInitial support








Performance Status


Used hardware:

  • CCR1036, 16GB RAM (tile)
  • CCR2004(arm64)
  • CCR1100AHx4(arm)
  • Intel(R) Core(TM) i7-4790 CPU @ 3.60GHz 32GB RAM (as a host for CHRs)


Simulated The simulated upstream peer is a CHR router running ROSv6 with a copy of the global IPv4 routing table (585K routes loaded from MRT dump).

One Peer Receive Only

DUT establishes a connection to simulated upstream peerpeers, receives routes, and installs them in FIB.


v6.44v7.1beta3v7.1rc7
CCR0:40 - 2:120:46
RB1100x4 1.4GHz0:32-0:380:23
CCR20040:320:18
x86 (CHR)0:20

RB450G (in/out affinity=alone)after trying for 9min - ran out of memory at 558K routes2:02 (121MB free)
RB450G (in/out affinity=main)-1:54
RB450G (affinity in=alone out=input)-2:12


Two Peers Receive Only

DUT establishes a connection to two simulated upstream peers, receives routes, picks the best route, and installs in FIB. On ROSv7 affinity settings are set to "alone".


v6.44FRRv7.1beta3v7.1rc7 (846k routes per peer)
CCR1:01 - 2:45
1:11
RB1100x4 1.4GHz0:51
0:30
CCR20040:51
0:290:33
router x


0:40
x86 (CHR)0:25


x86 (virtual)
0:26(4cores)



0:46(2cores)



0:30(2cores no LDP)


Multi-homing Sim

Two DUT devices establishes establish eBGP sessions to simulated x86 upstream routerrouters. Both DUTs are interconnected with the iBGP session. Each DUT receives routes from upstream and readvertise readvertises routes over iBGP. On ROSv7 affinity, settings are set to "alone" and early-cut disabled.

...

  • Route Provider: CHR (ROSv6)
  • DUT_1: CCR1032CCR1036
  • DUT_2: CCR1032CCR1036
v7.1beta31:11
v7.1beta21:29
v6.xx1:02 - 8:30

...

  • Route Provider: CHR (ROSv6)
  • DUT_1: CCR2004
  • DUT_2: RB1100AHx2
v7.1beta30:36
v6.xx0:59


Memory Usage:

Code Block
languagetext
Columns: TASKS, PRIVATE-MEM-BLOCKS, SHARED-MEM-BLOCKS, PSS, RSS, VMS, RETIRED, ID, PID, RPID, PROCESS-TIME, KERNEL-TIME, CUR-BUSY, MAX-BU>
   #  TASKS                         PRIVATE-M  SHARED-M  P  R  V  RE  ID       PID  R  PROCESS-  KERNEL-  CUR  MAX-BUS  CUR  MAX-CALC
   0  routing tables                12.0MiB    30.2MiB   0  0  0  12  main     111  0  8s980ms   2s60ms   0ms  1s320ms  0ms  10s700ms
      rib                                                                                                                            
      connected networks                                                                                                             
   1  fib                           2816.0KiB  0         0  0  0      fib      130  1  3s        4s660ms       7s220ms       7s220ms 
   2  ospf                          512.0KiB   256.0KiB  0  0  0      ospf     137  1  1s220ms   130ms         980ms         1s40ms  
      connected networks                                                                                                             
   3  fantasy                       256.0KiB   0         0  0  0      fantasy  138  1  60ms      80ms          40ms          40ms    
   4  configuration and reporting   3840.0KiB  512.0KiB  0  0  0      static   139  1  1s270ms   110ms         260ms         260ms   
   5  rip                           512.0KiB   0         0  0  0      rip      136  1  120ms     70ms          60ms          120ms   
      connected networks                                                                                                             
   6  routing policy configuration  768.0KiB   768.0KiB  0  0  0      policy   133  1  2s290ms   3s170ms       80ms          80ms    
   7  BGP service                   768.0KiB   0         0  0  0      bgp      134  1  2s760ms   5s480ms       20ms          60ms    
      connected networks                                                                                                             
   8  BFD service                   512.0KiB   0         0  0  0      12       135  1  100ms     90ms          40ms          120ms   
      connected networks                                                                                                             
   9  BGP Input 10.155.101.186      3072.0KiB  6.2MiB    0  0  0      20       183  1  1s350ms   1s190ms       20ms          20ms    
  10  BGP Output 10.155.101.186     5.5MiB     0         0  0  0      21       184  1  5s400ms   500ms         3s880ms       3s880ms 
  11  BGP Input 10.155.101.232      3072.0KiB  6.2MiB    0  0  0      22       187  1  970ms     740ms         20ms          20ms    
  12  BGP Output 10.155.101.232     8.2MiB     0         0  0  0      23       188  1  10s830ms  960ms         7s            7s      
  13  Global memory                            256.0KiB               global     0  0

...