bgp techniques for internet service providers 2011 1 bgp techniques for internet service providers...
Post on 01-Apr-2018
221 Views
Preview:
TRANSCRIPT
AfNOG 2011 1
BGP Techniques for Internet ServiceProviders
Philip Smith <pfs@cisco.com>AfNOG 2011Dar Es Salaam, Tanzania5 June 2011
AfNOG 2011 2
Presentation Slides
Will be available onftp://ftp-eng.cisco.com/pfs/seminars/AfNOG2011-BGP-Techniques.pdfAnd on the AfNOG2011 website
Feel free to ask questions any time
AfNOG 2011 3
Deploying BGP in an ISP Network
We’ve learned about BGP in SI-E/F and AR-E… What now?
AfNOG 2011 4
Deploying BGP
The role of IGPs and iBGP
Aggregation
Receiving Prefixes
Configuration Tips
Deploying 4-byte ASNs
AfNOG 2011 5
The role of IGP and iBGP
Ships in the night?OrGood foundations?
AfNOG 2011 6
BGP versus OSPF/ISIS
Internal Routing Protocols (IGPs)examples are ISIS and OSPFused for carrying infrastructure addressesNOT used for carrying Internet prefixes or customer prefixesdesign goal is to minimise number of prefixes in IGP to aidscalability and rapid convergence
AfNOG 2011 7
BGP versus OSPF/ISIS
BGP used internally (iBGP) and externally (eBGP) iBGP used to carry
some/all Internet prefixes across backbonecustomer prefixes
eBGP used toexchange prefixes with other ASesimplement routing policy
AfNOG 2011 8
BGP/IGP model used in ISP networks
Model representation
IGP
iBGP
IGP
iBGP
IGP
iBGP
IGP
iBGP
eBGP eBGP eBGP
AS1 AS2 AS3 AS4
AfNOG 2011 9
BGP versus OSPF/ISIS
DO NOT:distribute BGP prefixes into an IGPdistribute IGP routes into BGPuse an IGP to carry customer prefixes
YOUR NETWORK WILL NOT SCALE
AfNOG 2011 10
Injecting prefixes into iBGP
Use iBGP to carry customer prefixesDon’t ever use IGP
Point static route to customer interface Enter network into BGP process
Ensure that implementation options are used so that the prefixalways remains in iBGP, regardless of state of interfacei.e. avoid iBGP flaps caused by interface flaps
AfNOG 2011 11
Aggregation
Quality or Quantity?
AfNOG 2011 12
Aggregation
Aggregation means announcing the address blockreceived from the RIR to the other ASes connected toyour network
Subprefixes of this aggregate may be:Used internally in the ISP networkAnnounced to other ASes to aid with multihoming
Unfortunately too many people are still thinking aboutclass Cs, resulting in a proliferation of /24s in theInternet routing table
AfNOG 2011 13
Aggregation
Address block should be announced to the Internet asan aggregate
Subprefixes of address block should NOT beannounced to Internet unless for traffic engineeringpurposes
(see BGP Multihoming Tutorial)
Aggregate should be generated internallyNot on the network borders!
AfNOG 2011 14
Announcing an Aggregate
ISPs who don’t and won’t aggregate are held in poorregard by community
Registries publish their minimum allocation sizeAnything from a /20 to a /24 depending on RIRDifferent sizes for different address blocksThere are currently >185000 /24s!
APNIC changed (Oct 2010) its minimum allocation sizeon all blocks to /24
IPv4 run-out is starting to have an impact
AfNOG 2011 15
AS100customer
100.10.10.0/23Internet
100.10.10.0/23100.10.0.0/24100.10.4.0/22…
Aggregation – Example
Customer has /23 network assigned from AS100’s /19 address block
AS100 announces customers’ individual networks to the Internet
AfNOG 2011 16
Customer link returnsTheir /23 network is nowvisible to their ISPTheir /23 network is re-advertised to peersStarts rippling through InternetLoad on Internet backbonerouters as network isreinserted into routing tableSome ISP’s suppress the flapsInternet may take 10-20 min orlonger to be visibleWhere is the Quality ofService???
Customer link goes downTheir /23 network becomesunreachable/23 is withdrawn from AS100’siBGP
Their ISP doesn’t aggregate its/19 network block
/23 network withdrawalannounced to peersstarts rippling through theInternetadded load on all Internetbackbone routers as networkis removed from routing table
Aggregation – Bad Example
AfNOG 2011 17
AS100customer
100.10.10.0/23
100.10.0.0/19aggregate
Internet
100.10.0.0/19
Aggregation – Example
Customer has /23 network assigned from AS100’s /19 address block
AS100 announced /19 aggregate to the Internet
AfNOG 2011 18
Aggregation – Good Example
Customer link goes downtheir /23 network becomesunreachable/23 is withdrawn from AS100’siBGP
/19 aggregate is still beingannounced
no BGP hold down problemsno BGP propagation delaysno damping by other ISPs
Customer link returns
Their /23 network is visibleagain
The /23 is re-injected intoAS100’s iBGP
The whole Internet becomesvisible immediately
Customer has Quality ofService perception
AfNOG 2011 19
Aggregation – Summary
Good example is what everyone should do!Adds to Internet stabilityReduces size of routing tableReduces routing churnImproves Internet QoS for everyone
Bad example is what too many still do!Why? Lack of knowledge?Laziness?
AfNOG 2011 20
Separation of iBGP and eBGP
Many ISPs do not understand the importance ofseparating iBGP and eBGP
iBGP is where all customer prefixes are carriedeBGP is used for announcing aggregate to Internet and forTraffic Engineering
Do NOT do traffic engineering with customer originatediBGP prefixes
Leads to instability similar to that mentioned in the earlier badexampleEven though aggregate is announced, a flapping subprefix willlead to instability for the customer concerned
Generate traffic engineering prefixes on the BorderRouter
AfNOG 2011 21
The Internet Today (1st June 2011)
Current Internet Routing Table StatisticsBGP Routing Table Entries 358603Prefixes after maximum aggregation 162337Unique prefixes in Internet 178173Prefixes smaller than registry alloc 149545/24s announced 186667ASes in use 37758
AfNOG 2011 22
“The New Swamp”
Swamp space is name used for areas of pooraggregation
The original swamp was 192.0.0.0/8 from the former class Cblock
Name given just after the deployment of CIDR
The new swamp is creeping across all parts of the InternetNot just RIR space, but “legacy” space too
AfNOG 2011 23
“The New Swamp”RIR Space – February 1999
RIR blocks contribute 88% of the Internet Routing TableBlock Networks118/8 0119/8 0120/8 0121/8 0122/8 0123/8 0124/8 0125/8 0126/8 0173/8 0174/8 0186/8 0187/8 0189/8 0190/8 0192/8 6275193/8 2390194/8 2932195/8 1338196/8 513198/8 4034199/8 3495200/8 1348
Block Networks201/8 0202/8 2276203/8 3622204/8 3792205/8 2584206/8 3127207/8 2723208/8 2817209/8 2574210/8 617211/8 0212/8 717213/8 1216/8 943217/8 0218/8 0219/8 0220/8 0221/8 0222/8 0
Block Networks24/8 16541/8 058/8 059/8 060/8 061/8 362/8 8763/8 2064/8 065/8 066/8 067/8 068/8 069/8 070/8 071/8 072/8 073/8 074/8 075/8 076/8 077/8 078/8 0
Block Networks79/8 080/8 081/8 082/8 083/8 084/8 085/8 086/8 087/8 088/8 089/8 090/8 091/8 096/8 097/8 098/8 099/8 0112/8 0113/8 0114/8 0115/8 0116/8 0117/8 0
AfNOG 2011 24
“The New Swamp”RIR Space – February 2010
Block Networks118/8 1349119/8 1694120/8 531121/8 1756122/8 2687123/8 2400124/8 2259125/8 2514126/8 106173/8 1994174/8 1089186/8 1223187/8 1501189/8 3063190/8 6945192/8 6952193/8 6820194/8 5177195/8 5325196/8 1857198/8 4504199/8 4372 200/8 8884
Block Networks201/8 4136202/8 11354203/8 11677204/8 5744205/8 3037206/8 3951207/8 4635208/8 6498209/8 5536210/8 4977211/8 3130212/8 3550213/8 3442216/8 7645217/8 3136218/8 1512219/8 1303220/8 2108221/8 980222/8 1058
Block Networks24/8 332841/8 344858/8 167559/8 157560/8 88861/8 289062/8 241863/8 311464/8 660165/8 396666/8 778267/8 377168/8 322169/8 528070/8 200871/8 132772/8 405073/8 474/8 507475/8 116476/8 103477/8 196478/8 1397
Block Networks79/8 111980/8 233581/8 170982/8 135883/8 135784/8 134185/8 249286/8 78087/8 146688/8 106889/8 316890/8 37791/8 455596/8 77897/8 72598/8 131299/8 288112/8 883113/8 890114/8 996115/8 1616116/8 1755117/8 1611
RIR blocks contribute about 87% of the Internet Routing Table
AfNOG 2011 25
“The New Swamp”Summary
RIR space shows creeping deaggregationIt seems that an RIR /8 block averages around 5000 prefixes(and upwards) once fully allocated
Food for thought:The 120 RIR /8s combined will cause:635000 prefixes with 5000 prefixes per /8 density762000 prefixes with 6000 prefixes per /8 densityPlus 12% due to “non RIR space deaggregation”→ Routing Table size of 853440 prefixes
AfNOG 2011 26
“The New Swamp”Summary
Rest of address space is showing similar deaggregationtoo
What are the reasons?Main justification is traffic engineering
Real reasons are:Lack of knowledgeLazinessDeliberate & knowing actions
AfNOG 2011 27
Efforts to improve aggregation
The CIDR ReportInitiated and operated for many years by Tony BatesNow combined with Geoff Huston’s routing analysis
www.cidr-report.orgResults e-mailed on a weekly basis to most operations listsaround the worldLists the top 30 service providers who could do better ataggregating
RIPE Routing WG aggregation recommendationRIPE-399 — http://www.ripe.net/ripe/docs/ripe-399.html
AfNOG 2011 28
Efforts to Improve AggregationThe CIDR Report
Also computes the size of the routing table assumingISPs performed optimal aggregation
Website allows searches and computations ofaggregation to be made on a per AS basis
Flexible and powerful tool to aid ISPsIntended to show how greater efficiency in terms of BGP tablesize can be obtained without loss of routing and policyinformationShows what forms of origin AS aggregation could be performedand the potential benefit of such actions to the total table sizeVery effectively challenges the traffic engineering excuse
AfNOG 2011 29
AfNOG 2011 30
AfNOG 2011 31
AfNOG 2011 32
AfNOG 2011 33
AfNOG 2011 34
AfNOG 2011 35
Importance of Aggregation
Size of routing tableRouter Memory is not so much of a problem as it was in the1990sRouters can be specified to carry 1 million+ prefixes
Convergence of the Routing SystemThis is a problemBigger table takes longer for CPU to processBGP updates take longer to deal withBGP Instability Report tracks routing system update activityhttp://bgpupdates.potaroo.net/instability/bgpupd.html
AfNOG 2011 36
AfNOG 2011 37
AfNOG 2011 38
Aggregation Potential(source: bgp.potaroo.net/as2.0/)
AS Path
AS Origin
AfNOG 2011 39
AggregationSummary
Aggregation on the Internet could be MUCH better35% saving on Internet routing table size is quite feasibleTools are available
Commands on the routers are not hardCIDR-Report webpage
AfNOG 2011 40
Receiving Prefixes
AfNOG 2011 41
Receiving Prefixes
There are three scenarios for receiving prefixes fromother ASNs
Customer talking BGPPeer talking BGPUpstream/Transit talking BGP
Each has different filtering requirements and need to beconsidered separately
AfNOG 2011 42
Receiving Prefixes:From Customers
ISPs should only accept prefixes which have beenassigned or allocated to their downstream customer
If ISP has assigned address space to its customer, thenthe customer IS entitled to announce it back to his ISP
If the ISP has NOT assigned address space to itscustomer, then:
Check the five RIR databases to see if this address space reallyhas been assigned to the customerThe tool: whois
AfNOG 2011 43
Example use of whois to check if customer is entitled to announceaddress space:
$ whois -h whois.apnic.net 202.12.29.0inetnum: 202.12.28.0 - 202.12.29.255netname: APNIC-APdescr: Asia Pacific Network Information Centredescr: Regional Internet Registry for the Asia-Pacificdescr: 6 Cordelia Streetdescr: South Brisbane, QLD 4101descr: Australiacountry: AUadmin-c: AIC1-APtech-c: NO4-APmnt-by: APNIC-HMmnt-irt: IRT-APNIC-APchanged: hm-changed@apnic.netstatus: ASSIGNED PORTABLEchanged: hm-changed@apnic.net 20110309source: APNIC
Portable – means its an assignmentto the customer, the customer canannounce it to you
Receiving Prefixes:From Customers
AfNOG 2011 44
Example use of whois to check if customer is entitled to announce addressspace:
$ whois -h whois.ripe.net 193.128.0.0inetnum: 193.128.0.0 - 193.133.255.255netname: UK-PIPEX-193-128-133descr: Verizon UK Limitedcountry: GBorg: ORG-UA24-RIPEadmin-c: WERT1-RIPEtech-c: UPHM1-RIPEstatus: ALLOCATED UNSPECIFIEDremarks: Please send abuse notification to abuse@uk.uu.netmnt-by: RIPE-NCC-HM-MNTmnt-lower: AS1849-MNTmnt-routes: AS1849-MNTmnt-routes: WCOM-EMEA-RICE-MNTmnt-irt: IRT-MCI-GBsource: RIPE # Filtered
ALLOCATED – means that this isProvider Aggregatable address spaceand can only be announced by the ISPholding the allocation (in this caseVerizon UK)
Receiving Prefixes:From Customers
AfNOG 2011 45
Receiving Prefixes:From Peers
A peer is an ISP with whom you agree to exchangeprefixes you originate into the Internet routing table
Prefixes you accept from a peer are only those they haveindicated they will announcePrefixes you announce to your peer are only those you haveindicated you will announce
AfNOG 2011 46
Receiving Prefixes:From Peers
Agreeing what each will announce to the other:Exchange of e-mail documentation as part of the peeringagreement, and then ongoing updates
ORUse of the Internet Routing Registry and configuration toolssuch as the IRRToolSet
www.isc.org/sw/IRRToolSet/
AfNOG 2011 47
Receiving Prefixes:From Upstream/Transit Provider
Upstream/Transit Provider is an ISP who you pay togive you transit to the WHOLE Internet
Receiving prefixes from them is not desirable unlessreally necessary
Traffic Engineering – see BGP Multihoming Tutorial
Ask upstream/transit provider to either:originate a default-route
ORannounce one prefix you can use as default
AfNOG 2011 48
Receiving Prefixes:From Upstream/Transit Provider
If necessary to receive prefixes from any provider, careis required.
Don’t accept default (unless you need it)Don’t accept your own prefixes
For IPv4:Don’t accept private (RFC1918) and certain special useprefixes:
http://www.rfc-editor.org/rfc/rfc5735.txtDon’t accept prefixes longer than /24 (?)
For IPv6:Don’t accept certain special use prefixes:
http://www.rfc-editor.org/rfc/rfc5156.txtDon’t accept prefixes longer than /48 (?)
AfNOG 2011 49
Receiving Prefixes:From Upstream/Transit Provider
Check Team Cymru’s list of “bogons”www.team-cymru.org/Services/Bogons/http.html
For IPv4 also consult:datatracker.ietf.org/doc/draft-vegoda-no-more-unallocated-slash8s
For IPv6 also consult:www.space.net/~gert/RIPE/ipv6-filters.html
Bogon Route Server:www.team-cymru.org/Services/Bogons/routeserver.htmlSupplies a BGP feed (IPv4 and/or IPv6) of address blockswhich should not appear in the BGP table
AfNOG 2011 50
Receiving Prefixes
Paying attention to prefixes received from customers,peers and transit providers assists with:
The integrity of the local networkThe integrity of the Internet
Responsibility of all ISPs to be good Internet citizens
AfNOG 2011 51
Configuration Tips
Of passwords, tricks and templates
AfNOG 2011 52
iBGP and IGPsReminder!
Make sure loopback is configured on routeriBGP between loopbacks, NOT real interfaces
Make sure IGP carries loopback /32 address
Consider the DMZ nets:Use unnumbered interfaces?Use next-hop-self on iBGP neighboursOr carry the DMZ /30s in the iBGPBasically keep the DMZ nets out of the IGP!
AfNOG 2011 53
iBGP: Next-hop-self
BGP speaker announces external network to iBGPpeers using router’s local address (loopback) as next-hop
Used by many ISPs on edge routersPreferable to carrying DMZ /30 addresses in the IGPReduces size of IGP to just core infrastructureAlternative to using unnumbered interfacesHelps scale networkMany ISPs consider this “best practice”
AfNOG 2011 54
Limiting AS Path Length
Some BGP implementations have problems with longAS_PATHS
Memory corruptionMemory fragmentation
Even using AS_PATH prepends, it is not normal to seemore than 20 ASes in a typical AS_PATH in theInternet today
The Internet is around 5 ASes deep on averageLargest AS_PATH is usually 16-20 ASNs
AfNOG 2011 55
Limiting AS Path Length
Some announcements have ridiculous lengths of AS-paths:
*> 3FFE:1600::/24 22 11537 145 12199 1031810566 13193 1930 2200 3425 293 5609 5430 13285 693914277 1849 33 15589 25336 6830 8002 2042 7610 i
This example is an error in one IPv6 implementation
*> 96.27.246.0/24 2497 1239 12026 12026 1202612026 12026 12026 12026 12026 12026 12026 1202612026 12026 12026 12026 12026 12026 12026 1202612026 12026 12026 i
This example shows 21 prepends (for no obvious reason)
If your implementation supports it, consider limiting themaximum AS-path length you will accept
AfNOG 2011 56
BGP TTL “hack”
Implement RFC5082 on BGP peerings(Generalised TTL Security Mechanism)Neighbour sets TTL to 255Local router expects TTL of incoming BGP packets to be 254No one apart from directly attached devices can send BGPpackets which arrive with TTL of 254, so any possible attack bya remote miscreant is dropped due to TTL mismatch
ISP AS 100Attacker
TTL 254
TTL 253 TTL 254R1 R2
AfNOG 2011 57
BGP TTL “hack”
TTL Hack:Both neighbours must agree to use the featureTTL check is much easier to perform than MD5(Called BTSH – BGP TTL Security Hack)
Provides “security” for BGP sessionsIn addition to packet filters of courseMD5 should still be used for messages which slip through theTTL hackSee www.nanog.org/mtg-0302/hack.html for more details
AfNOG 2011 58
Templates
Good practice to configure templates for everythingVendor defaults tend not to be optimal or even very useful forISPsISPs create their own defaults by using configuration templates
eBGP and iBGP examples followAlso see Team Cymru’s BGP templates
http://www.team-cymru.org/ReadingRoom/Documents/
AfNOG 2011 59
iBGP TemplateExample
iBGP between loopbacks!
Next-hop-selfKeep DMZ and external point-to-point out of IGP
Always send communities in iBGPOtherwise accidents will happen
Hardwire BGP to version 4Yes, this is being paranoid!
AfNOG 2011 60
iBGP TemplateExample continued
Use passwords on iBGP sessionNot being paranoid, VERY necessaryIt’s a secret shared between you and your peerIf arriving packets don’t have the correct MD5 hash, they areignoredHelps defeat miscreants who wish to attack BGP sessions
Powerful preventative tool, especially when combinedwith filters and the TTL “hack”
AfNOG 2011 61
eBGP TemplateExample
BGP dampingDo NOT use it unless you understand the impactDo NOT use the vendor defaults without thinking
Remove private ASes from announcementsCommon omission today
Use extensive filters, with “backup”Use as-path filters to backup prefix filtersKeep policy language for implementing policy, rather than basicfiltering
Use password agreed between you and peer on eBGPsession
AfNOG 2011 62
eBGP TemplateExample continued
Use maximum-prefix trackingRouter will warn you if there are sudden increases in BGP tablesize, bringing down eBGP if desired
Limit maximum as-path length inbound
Log changes of neighbour state…and monitor those logs!
Make BGP admin distance higher than that of any IGPOtherwise prefixes heard from outside your network couldoverride your IGP!!
AfNOG 2011 63
Summary
Use configuration templates
Standardise the configuration
Be aware of standard “tricks” to avoid compromise ofthe BGP session
Anything to make your life easier, network less prone toerrors, network more likely to scale
It’s all about scaling – if your network won’t scale, thenit won’t be successful
AfNOG 2011 64
Deploying 32-bit ASNs
How to support customers using the extended ASN range
AfNOG 2011 65
32-bit ASNs
Standards documentsDescription of 32-bit ASNs
www.rfc-editor.org/rfc/rfc4893.txtTextual representation
www.rfc-editor.org/rfc/rfc5396.txtNew extended community
www.rfc-editor.org/rfc/rfc5668.txt
AS 23456 is reserved as interface between 16-bit and32-bit ASN world
AfNOG 2011 66
32-bit ASNs – terminology
16-bit ASNsRefers to the range 0 to 65535
32-bit ASNsRefers to the range 65536 to 4294967295(or the extended range)
32-bit ASN poolRefers to the range 0 to 4294967295
AfNOG 2011 67
Getting a 32-bit ASN
Sample RIR policywww.apnic.net/docs/policy/asn-policy.html
From 1st January 200732-bit ASNs were available on request
From 1st January 200932-bit ASNs were assigned by default16-bit ASNs were only available on request
From 1st January 2010No distinction – ASNs assigned from the 32-bit pool
AfNOG 2011 68
Representation
Representation of 0-4294967295 ASN rangeMost operators favour traditional format (asplain)A few prefer dot notation (X.Y):
asdot for 65536-4294967295, e.g 2.4asdot+ for 0-4294967295, e.g 0.64513
But regular expressions will have to be completely rewritten forasdot and asdot+ !!!
For example:^[0-9]+$ matches any ASN (16-bit and asplain)This and equivalents extensively used in BGP multihomingconfigurations for traffic engineering
Equivalent regexp for asdot is: ^([0-9]+)|([0-9]+\.[0-9]+)$
Equivalent regexp for asdot+ is: ^[0-9]+\.[0-9]+$
AfNOG 2011 69
Changes
32-bit ASNs are backward compatible with 16-bit ASNs There is no flag day You do NOT need to:
Throw out your old routersReplace your 16-bit ASN with a 32-bit ASN
You do need to be aware that:Your customers will come with 32-bit ASNsASN 23456 is not a bogon!You will need a router supporting 32-bit ASNs to use a 32-bitASN locally
If you have a proper BGP implementation, 32-bit ASNswill be transported silently across your network
AfNOG 2011 70
How does it work?
If local router and remote router supports configurationof 32-bit ASNs
BGP peering is configured as normal using the 32-bit ASN
If local router and remote router does not supportconfiguration of 32-bit ASNs
BGP peering can only use a 16-bit ASN
If local router only supports 16-bit ASN and remoterouter/network has a 32-bit ASN
Compatibility mode is initiated…
AfNOG 2011 71
Compatibility Mode:
Local router only supports 16-bit ASN and remote router uses 32-bit ASN
BGP peering initiated:Remote asks local if 32-bit supported (BGP capability negotiation)When local says “no”, remote then presents AS23456Local needs to be configured to peer with remote using AS23456
BGP peering initiated (cont):BGP session established using AS2345632-bit ASN included in a new BGP attribute called AS4_PATH
(as opposed to AS_PATH for 16-bit ASNs)
Result:16-bit ASN world sees 16-bit ASNs and 23456 standing in for 32-bitASNs32-bit ASN world sees 16 and 32-bit ASNs
AfNOG 2011 72
180.10.0.0/16 123 23456 23456170.10.0.0/16 123 23456
AS 80000
AS 123
AS 70000
AS 90000
AS 321
170.10.0.0/16 180.10.0.0/16
150.10.0.0/16
180.10.0.0/16 123 70000 80000170.10.0.0/16 123 70000150.10.0.0/16 123 321
Example:
Internet with 32-bit and16-bit ASNs
AS-PATH lengthmaintained
AfNOG 2011 73
What has changed?
Two new BGP attributes:AS4_PATH
Carries 32-bit ASN path infoAS4_AGGREGATOR
Carries 32-bit ASN aggregator infoWell-behaved BGP implementations will simply pass thesealong if they don’t understand them
AS23456 (AS_TRANS)
AfNOG 2011 74
asdotformat
asplainformat
What do they look like?
IPv4 prefix originated by AS196613as4-7200#sh ip bgp 145.125.0.0/20BGP routing table entry for 145.125.0.0/20, version 58734Paths: (1 available, best #1, table default) 131072 12654 196613 204.69.200.25 from 204.69.200.25 (204.69.200.25) Origin IGP, localpref 100, valid, internal, best
IPv4 prefix originated by AS3.5as4-7200#sh ip bgp 145.125.0.0/20BGP routing table entry for 145.125.0.0/20, version 58734Paths: (1 available, best #1, table default) 2.0 12654 3.5 204.69.200.25 from 204.69.200.25 (204.69.200.25) Origin IGP, localpref 100, valid, internal, best
AfNOG 2011 75
TransitionAS
What do they look like?
IPv4 prefix originated by AS196613But 16-bit AS world view:
BGP-view1>sh ip bgp 145.125.0.0/20
BGP routing table entry for 145.125.0.0/20, version 113382
Paths: (1 available, best #1, table Default-IP-Routing-Table)
23456 12654 23456
204.69.200.25 from 204.69.200.25 (204.69.200.25)
Origin IGP, localpref 100, valid, external, best
AfNOG 2011 76
If 32-bit ASN not supported:
Inability to distinguish between peer ASes using 32-bit ASNsThey will all be represented by AS23456Could be problematic for transit provider’s policy
Inability to distinguish prefix’s origin ASHow to tell whether origin is real or fake?The real and fake both represented by AS23456(There should be a better solution here!)
Incorrect NetFlow summaries:Prefixes from 32-bit ASNs will all be summarised under AS23456Traffic statistics need to be measured per prefix and aggregatedMakes it hard to determine peerability of a neighbouring network
AfNOG 2011 77
iBGP Deployment (1)
Typical ISP design is thus:ISIS/OSPF for IGP, carrying loopback and point to point linkaddressesiBGP mesh (full/RR/Confederation) to carry customer andInternet prefixes
All routers support 4-byte ASNs:Proceed with iBGP design as normal
Not all routers support 4-byte ASNs:Three viable options
AfNOG 2011 78
iBGP Deployment (2)
1. Return 4-byte ASN to the RIR and request 2-byteASN instead:Works if RIR is willing to do soWorks as long as there are 2-byte ASNs remaining
2. Partial iBGP mesh:Routers which support 4-byte ASNs run iBGP meshRouters which do not support 4-byte ASNs either:
Run in private ASN (as a pseudo-customer) orDo not run BGP at all
3. Use a BGP Confederation(see AR-E Workshop)
AfNOG 2011 79
Implementations (May 2011)
Cisco IOS-XR 3.4 onwards
Cisco IOS-XE 2.3 onwards
Cisco IOS 12.0(32)S12, 12.4(24)T, 12.2SRE, 12.2(33)SXI1 onwards
Cisco NX-OS 4.0(1) onwards
Quagga 0.99.10 (patches for 0.99.6)
OpenBGPd 4.2 (patches for 3.9 & 4.0)
Juniper JunOSe 4.1.0 & JunOS 9.1 onwards
Redback SEOS
Force10 FTOS7.7.1 onwards
http://as4.cluepon.net/index.php/Software_Support for a complete list
AfNOG 2011 80
Cisco Routers Supporting 4-byte ASNs
CRSIOS-XR 3.4 onwards
GSRIOS-XR 3.4 onwardsIOS 12.0(32)S12, 12.0(33)S and 12.0(32)SY8 onwards
ASR1000IOS-XE 2.3 onwards
Nexus SwitchesNX-OS 4.0(1) onwards
AfNOG 2011 81
Cisco Routers Supporting 4-byte ASNs
Catalyst 6500IOS 12.2(33)SXI1 onwards
7600IOS 12.2(33)SRE1 onwards
7200 seriesIOS 12.0(32)S12, 12.0(33)S, 12.2(33)SRE1, 12.4(24)T, 15.0onwards
7301IOS 12.2(33)SRE1, 12.4(24)T, 15.0 onwards
AfNOG 2011 82
Cisco Routers Supporting 4-byte ASNs
3900/2900/1900 seriesIOS 15.0 onwards
3800/2800/1800/800 seriesIOS 12.4(24)T and IOS 15.0 onwards
3745/3725IOS 12.4(24)T
AS5350/5400IOS 12.4(24)T and IOS 15.0 onwards
AfNOG 2011 83
Cisco Routers NOT supporting 4-byteASNs
Routers which will never support 4-byte ASNs include:1700 series2500 series2600 series3600 seriesAS53007304
AfNOG 2011 84
Summary
Deploying 32-bit ASNs is simpleYour network can talk to a network which is using a 32-bit ASNYou have options with iBGP if not all your routers supportconfiguration of 32-bit ASN
Vendor support should be much betterRecent software support only, meaning older hardware will beproblematic
AfNOG 2011 85
Summary
AfNOG 2011 86
Summary
Tutorial has examined BGP deployment techniques:The role of IGPs and iBGPAggregationReceiving PrefixesConfiguration TipsDeploying 4-byte ASNs
AfNOG 2011 87
BGP Techniques for Internet ServiceProviders
The End!
top related