[c-nsp] Advice - c7200VXR with 2 bgp tables and peering fabric

Eric A Louie elouie at yahoo.com
Thu Jun 6 12:02:03 EDT 2013


Erik, I noticed that typo after I sent the message away.

Thank you for the bandwidth warning - I'm only about halfway there.  From the  old information at http://www.cisco.com/en/US/prod/collateral/routers/ps341/prod_qas0900aecd80471791.html it appears the NPE-G2 doubles the packet forwarding rate and the RAM capacity.  I'm not adverse to upgrading the 7206VXR with the NPE-G2




>________________________________
> From: Erik Versaevel <erik.versaevel at aspider.com>
>To: Eric A Louie <elouie at yahoo.com>; Cisco NSP <cisco-nsp at puck.nether.net> 
>Sent: Wednesday, June 5, 2013 10:38 PM
>Subject: RE: [c-nsp] Advice - c7200VXR with 2 bgp tables and peering fabric
> 
>
>Your milage  may vary but I never got more then +- 350/400 Mbps out  of an 7200vxr (npe-1g), above that CPU starts spending so much time on traffic forwarding that routing protocols get a bit neglected. (and I think you mean 150Mbps, not Gbps ;))
>
>
>
>
>
>
>
>
>-----Oorspronkelijk bericht-----
>Van: cisco-nsp [mailto:cisco-nsp-bounces at puck.nether.net] Namens Eric A Louie
>Verzonden: donderdag 6 juni 2013 0:50
>Aan: Cisco NSP
>Onderwerp: [c-nsp] Advice - c7200VXR with 2 bgp tables and peering fabric
>
>I have a c7204VXR NPE-G1 1GB RAM 6 GigE (3 on the NP, 1 on the I/O, and 2 PA-GE).  Passing about 150Gbps of traffic.  It's taking a full eBGP feed (470k routes), and connected to a peering fabric (30k routes so far).
>
>When I turned up the peering fabric, I spiked the cpu for about 5 minutes and it settled down nicely.  CPU utilization now is 25% max on core1, 20% on core2.
>
>I need to enable iBGP on it.  It's peer will be another c7204 (NPE-G1, 1GB, 3 GigE) with a full BGP feed (450k routes).  I have about 250Mbps backhaul link between the routers.  
>
>Right now, the memory utilization is about 40% on both routers which leads me to believe I could easily hold 2 full tables on both routers.  I may end up with a 3rd iBGP peer with another full feed too, so if this ends up being linear, I'm figuring I can do simple math on the RAM utilization.
>
>Can I get some confirmation from anyone who either has 2 BGP feeds on their 7200VXR or has iBGP with full eBGP tables on both routers, and tell me that I'm not going to melt down the two routers when I turn up iBGP?  My gut says I'll be just fine.
>
>thanks
>Eric
>
>Stats below
>
>core1#sh ip bgp sum
>BGP router identifier 67.xxx.xxx.100, local AS number 2xxxx BGP table version is 17685401, main routing table version
>17685401
>447841 network entries using 52397397 bytes of memory
>504542 path entries using 26236184 bytes of memory
>80432/75327 BGP path/bestpath attribute entries using 9973568 bytes of memory
>66127 BGP AS-PATH entries using 1820802 bytes of memory
>4894 BGP community entries using 290336 bytes of memory
>1 BGP extended community entries using 24 bytes of memory
>0 BGP route-map cache entries using 0 bytes of memory
>0 BGP filter-list cache entries using 0 bytes of memory BGP using 90718311total bytes of memory (about 100 Mbytes)
>1258 received paths for inbound soft reconfiguration BGP activity 3808783/3360941 prefixes, 4450994/3946452 paths, scan interval 60 secs
> 
>Neighbor
>V    AS MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down  State/PfxRcd
>64.211.193.209  4  3549 4587262   41246
>17685384    0    0
>1w5d       446643
>66.xxx.xxx.10   4 4xxx6   41288 3175949
>17685384    0    0
>4w0d            1
>66.xxx.xxx.82   4 3xxx1  378453  371596
>17685401    0    0
>3w1d            1
>66.209.124.90   4 5xxx7   41003
>41292 17685401    0    0
>4w0d            0
>206.223.143.252 4 19996    9026
>1035 17685384    0    0
>17:04:00    28207
>206.223.143.253 4 19996    9015
>1035 17685384    0    0
>17:04:00    28417
> 
>core1#show memory
>               
>Head    Total(b)
>Used(b)     Free(b)   Lowest(b)  Largest(b) Processor   63ADD410   928131852   362097708   566034144
>564822488   374483724 (about 350Mbytes)
>      I/O
> C000000    67108864
>12448104    54660760
>54490928    54262236
>Transient   7B000000
>16777216       17436
>16759780    16367684    16752164
>utilization is 39%
> 
>core2#sh ip bgp sum
>BGP router identifier 66.xxx.xxx.252, local AS number 2xxxx
>BGP table version is 788940103, main routing table version
>788940103
>447789 network entries using 52391313 bytes of memory
>447789 path entries using 23285028 bytes of memory
>114097/114095 BGP path/bestpath attribute entries using 14148028
>bytes of memory
>114070 BGP AS-PATH entries using 3217852 bytes of memory
>5 BGP community entries using 120 bytes of memory
>0 BGP route-map cache entries using 0 bytes of memory
>0 BGP filter-list cache entries using 0 bytes of memory
>BGP using 93042341total bytes of memory (about 100Mbytes
>RAM)
>17 received paths for inbound soft reconfiguration
>BGP activity 11660283/11212493 prefixes, 13183481/12735692
>paths, scan interval 60 secs
> 
>Neighbor       
>V    AS MsgRcvd MsgSent   TblVer  InQ OutQ
>Up/Down  State/PfxRcd
>64.94.111.13    4 10912 733746106 1128729
>788940103    0    0
>25w5d      447757
> 
>core2#sh mem
>               
>Head    Total(b)    
>Used(b)     Free(b)   Lowest(b)  Largest(b)
>Processor   63AE23D0   928111436   369270884  
>558840552   557289036   374463308 (350Mbytes used)
>      I/O   
>C000000    67108864    
>8678976    58429888    58311184   
>58154780
>Transient   7B000000   
>16777216       14944   
>16762272    16525620    16762272
>utilization is 40%
>_______________________________________________
>cisco-nsp mailing list  cisco-nsp at puck.nether.net
>https://puck.nether.net/mailman/listinfo/cisco-nsp
>archive at http://puck.nether.net/pipermail/cisco-nsp/
>
>
>


More information about the cisco-nsp mailing list