HP Load Balancing Module Appendix Protocol Reference Part number: 5998-4222 Software version: Feature 3221 Document version: 6PW100-20130326
Legal and notice information © Copyright 2013 Hewlett-Packard Development Company, L.P. No part of this documentation may be reproduced or transmitted in any form or by any means without prior written consent of Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.
Contents IP routing basics ··························································································································································· 1 Routing table ······································································································································································ 1 Dynamic routing protocols ······················································································································
Settlements for problems in large-scale BGP networks······························································································· 24 MP-BGP ··········································································································································································· 27 MP-BGP extended attributes ································································································································· 27 Address family ····
IP routing basics IP routing directs IP packet forwarding on routers based on a routing table. This book focuses on unicast routing protocols. The term "router" in this document refers to both routers and LB modules. Routing table A router maintains at least two routing tables: a global routing table and a FIB. The FIB table contains only the optimal routes, and the global routing table contains all routes. The router uses the FIB table to forward packets.
• Cost—If multiple routes to a destination have the same preference, the one with the smallest cost becomes the optimal route. • NextHop—Next hop. • Interface—Output interface. Dynamic routing protocols Static routes work well in small stable networks. They are easy to configure and require fewer system resources. However, in networks where topology changes occur frequently, a typical practice is to configure a dynamic routing protocol.
Route type Preference RIP 100 OSPF ASE 150 OSPF NSSA 150 IBGP 255 EBGP 255 Unknown (route from an untrusted source) 256 Load sharing A routing protocol may find multiple optimal equal-cost routes to the same destination. You can use these routes to implement equal-cost multi-path (ECMP) load sharing. Static routing, IPv6 static routing, RIP, RIPng, OSPF, OSPFv3, BGP, and IPv6 BGP support ECMP load sharing. The maximum number of ECMP routes for load sharing is 8.
Static routing Static routes are manually configured. If a network's topology is simple, you only need to configure static routes for the network to work properly. Static routes cannot adapt to network topology changes. If a fault or a topological change occurs in the network, the network administrator must modify the static routes manually.
Default route A default route is used to forward packets that match no entry in the routing table. Without a default route, a packet that does not match any routing entries is discarded. A default route can be configured in either of the following ways: • The network administrator can configure a default route with both destination and mask being 0.0.0.0. For more information, see Network Management Configuration Guide. • Some dynamic routing protocols, such as OSPF and RIP, can generate a default route.
RIP The term "router" in this document refers to both routers and LB modules. Routing Information Protocol (RIP) is a distance-vector simple interior gateway protocol suited to small-sized networks. It employs UDP to exchange route information through port 520. Overview RIP uses a hop count to measure the distance to a destination. The hop count from a router to a directly connected network is 0. The hop count from a router to a directly connected router is 1.
Routing loop prevention RIP uses the following mechanisms to prevent routing loops: • Counting to infinity—A destination with a metric value of 16 is considered unreachable. When a routing loop occurs, the metric value of a route will increment to 16 to avoid endless looping. • Split horizon—Disables RIP from sending routing information on the interface from which the information was learned to prevent routing loops and save bandwidth.
RIP message format A RIP message consists of a header and up to 25 route entries. (A RIPv2 authentication message uses the first route entry as the authentication entry, so it has up to 24 route entries.) RIPv1 message format Figure 1 RIPv1 message format 0 Header 7 Command 15 Version 31 Must be zero AFI Must be zero IP address Route entries Must be zero Must be zero Metric • Command—Type of message.
• IP address—Destination IP address. It can be a natural network address, subnet address or host address. • Subnet mask—Mask of the destination address. • Next hop—If set to 0.0.0.0, it indicates that the originator of the route is the best next hop; otherwise it indicates a next hop better than the originator of the route. RIPv2 authentication message format RIPv2 sets the AFI field of the first route entry to 0xFFFF to identify authentication information.
OSPF The term "router" in this document refers to both routers and LB modules. Open Shortest Path First (OSPF) is a link state IGP developed by the OSPF working group of the IETF. OSPF version 2 is used for IPv4. Unless otherwise stated, OSPF refers to OSPFv2 throughout this chapter. Overview OSPF has the following features: • Wide scope—Supports various network sizes and up to several hundred routers in an OSPF routing domain.
LSA types OSPF advertises routing information in Link State Advertisements (LSAs). The following describes some commonly used LSAs: • Router LSA—Type-1 LSA, originated by all routers and flooded throughout a single area only. This LSA describes the collected states of the router's interfaces to an area. • Network LSA—Type-2 LSA, originated for broadcast and NBMA networks by the designated router, and flooded throughout a single area only. This LSA contains the list of routers connected to the network.
Figure 4 Area based OSPF network partition Backbone area and virtual links Each AS has a backbone area that distributes routing information between non-backbone areas. Routing information between non-backbone areas must be forwarded by the backbone area. OSPF requires the following: • All non-backbone areas must maintain connectivity to the backbone area. • The backbone area must maintain connectivity within itself. In practice, the requirements might not be satisfied due to lack of physical links.
Figure 6 Virtual link application 2 The virtual link between the two ABRs acts as a point-to-point connection. You can configure interface parameters, such as hello interval, on the virtual link as they are configured on a physical interface. The two ABRs on the virtual link unicast OSPF packets to each other, and the OSPF routers in between convey these OSPF packets as normal IP packets.
Figure 7 NSSA area Router types OSPF classifies routers into the following types based on their positions in the AS: • Internal router—All interfaces on an internal router belong to one OSPF area. • Area Border Router (ABR)—Belongs to more than two areas, one of which must be the backbone area. An ABR connects the backbone area to a non-backbone area. An ABR and the backbone area can be connected through a physical or logical link.
• Type-1 external route • Type-2 external route The intra-area and inter-area routes describe the network topology of the AS. The external routes describe routes to external ASs. A Type-1 external route has high credibility. The cost from a router to the destination of a Type-1 external route = the cost from the router to the corresponding ASBR + the cost from the ASBR to the destination of the external route. A Type-2 external route has low credibility.
DR and BDR On a broadcast or NBMA network, any two routers must establish an adjacency to exchange routing information with each other. If n routers are present on the network, n(n-1)/2 adjacencies are established. Any topology change on the network results in an increase in traffic for route synchronization, consuming many system and bandwidth resources. The DR and BDR mechanisms can solve this problem. • DR—Elected to advertise routing information among other routers.
If a router with a higher router priority is added to the network after DR and BDR election, the router cannot become the DR or BDR immediately because no DR election is performed for it. Therefore, the DR of a network might not be the router with the highest priority, and the BDR might not be the router with the second highest priority.
BGP The term "router" in this document refers to both routers and LB modules. Overview Border Gateway Protocol (BGP) is an exterior gateway protocol. It is called internal BGP (IBGP) when it runs within an AS and called external BGP (EBGP) when it runs between ASs. The current version in use is BGP-4 (RFC 4271). Unless otherwise stated, BGP refers to BGP-4 in this document. BGP has the following characteristics: • Focuses on route control and the selection rather than route discovery and calculation.
• Notification—BGP sends a Notification message upon detecting an error and immediately closes the connection. BGP path attributes BGP uses the following path attributes in update messages for route filtering and selection: • ORIGIN The ORIGIN attribute identifies the origin of routing information (how a route became a BGP route). This attribute has the following types: { IGP—Has the highest priority. Routes generated in the local AS have the IGP attribute. { EGP—Has the second highest priority.
passing AS 40 for sending data to the destination 8.0.0.0. In some applications, you can apply a routing policy to control BGP route selection by modifying the AS_PATH length. { • Filter routes—By configuring an AS path filtering list, you can filter routes based on AS numbers contained in the AS_PATH attribute. For more information about routing policies and AS path filtering lists, see Network Management Configuration Guide.
Figure 12 MED attribute MED = 0 Router B 2.1.1.1 D = 9.0.0.0 Next_hop = 2.1.1.1 MED = 0 EBGP IBGP 9.0.0.0 IBGP Router A D = 9.0.0.0 Next_hop = 3.1.1.1 MED = 100 AS 10 EBGP Router D IBGP 3.1.1.1 Router C AS 20 MED = 100 Generally, BGP only compares MEDs of routes received from the same AS. You can also use the compare-different-as-med command to force BGP to compare MED values of routes received from different ASs.
Figure 13 LOCAL_PREF attribute • COMMUNITY The COMMUNITY attribute identifies the community of BGP routes. A BGP community is a group of routes with the same characteristics. It has no geographical boundaries. Routes of different ASs can belong to the same community. A route can carry one or more COMMUNITY attribute values (each of which is represented by a four-byte integer).
BGP route selection BGP discards routes with unreachable NEXT_HOPs. If multiple routes to the same destination are available, BGP selects the best route in the following sequence: 1. Highest Preferred_value 2. Highest LOCAL_PREF 3. Summary route 4. Shortest AS_PATH 5. IGP, EGP, or INCOMPLETE route in turn 6. Lowest MED value 7. Learned from EBGP, confederation, or IBGP in turn 8. Smallest next hop metric 9. Shortest CLUSTER_LIST 10. Smallest ORIGINATOR_ID 11.
the same number of next hops to forward packets. BGP load balancing based on route recursion is always enabled by the system rather than configured by using commands. • BGP load balancing through route selection BGP differs from IGP in the implementation of load balancing in the following ways: { { IGP routing protocols, such as RIP and OSPF, compute metrics of routes, and then implement load balancing over routes with the same metric and to the same destination. The route selection criterion is metric.
Route summarization can reduce the BGP routing table size by advertising summary routes rather than more specific routes. The system supports both manual and automatic route summarization. Manual route summarization allows you to determine the attribute of a summary route and whether to advertise more specific routes. • Route dampening BGP route dampening solves the issue of route instability such as route flaps—a route comes up and disappears in the routing table frequently.
• Route reflector IBGP peers must be fully meshed to maintain connectivity. If n routers exist in an AS, the number of IBGP connections is n(n-1)/2. If a large number of IBGP peers exist, large amounts of network and CPU resources are consumed to maintain sessions. Using route reflectors can solve this issue. In an AS, a router acts as a route reflector, and other routers act as clients connecting to the route reflector.
After route reflection is disabled between clients, routes can still be reflected between a client and a non-client. • Confederation Confederation is another method to manage growing IBGP connections in an AS. It splits an AS into multiple sub-ASs. In each sub-AS, IBGP peers are fully meshed. As shown in Figure 18, intra-confederation EBGP connections are established between sub-Ass in AS 200.
To support multiple network layer protocols, MP-BGP defines the following path attributes: • MP_REACH_NLRI—Multiprotocol Reachable NLRI, for carrying prefixes of feasible routes and next hops for multiple network layer protocols. Such routes can then be advertised. • MP_UNREACH_NLRI—Multiprotocol Unreachable NLRI, for carrying prefixes of unfeasible routes for multiple network layer protocols. Such routes can then be withdrawn.
• RFC 1771, A Border Gateway Protocol 4 (BGP-4) • RFC 2858, Multiprotocol Extensions for BGP-4 • RFC 3392, Capabilities Advertisement with BGP-4 • RFC 2918, Route Refresh Capability for BGP-4 • RFC 2439, BGP Route Flap Damping • RFC 1997, BGP Communities Attribute • RFC 2796, BGP Route Reflection • RFC 3065, Autonomous System Confederations for BGP • RFC 4271, A Border Gateway Protocol 4 (BGP-4) • RFC 4360, BGP Extended Communities Attribute • RFC 4760, Multiprotocol Extensions for BGP-
IPv6 static routing The term "router" in this document refers to both routers and LB modules. Static routes are manually configured. If a network's topology is simple, you only need to configure static routes for the network to work properly. Static routes cannot adapt to network topology changes. If a fault or a topological change occurs in the network, the network administrator has to modify the static routes manually.
IPv6 default route An IPv6 default route is used to forward packets that match no entry in the routing table. An IPv6 default route can be configured in either of the following ways: • The network administrator can configure a default route with a destination prefix of ::/0. For more information, see Network Management Configuration Guide. • Some dynamic routing protocols, such as OSPFv3 and RIPng, can generate an IPv6 default route.
RIPng The term "router" in this document refers to both routers and LB modules. Overview RIP next generation (RIPng) is an extension of RIP-2 for IPv4. Most RIP concepts are applicable in RIPng. RIPng for IPv6 has the following basic differences from RIP: • UDP port number—RIPng uses UDP port 521 for sending and receiving routing information. • Multicast address—RIPng uses FF02::9 as the link-local-router multicast address. • Destination Prefix—128-bit destination address prefix.
RIPng packet format Basic format A RIPng packet consists of a header and multiple route table entries (RTEs). The maximum number of RTEs in a packet depends on the IPv6 MTU of the sending interface. Figure 19 RIPng basic packet format Packet header description: • Command—Type of message. 0x01 indicates Request, 0x02 indicates Response. • Version—Version of RIPng. It can only be 0x01. • RTE—Route table entry. It is 20 bytes for each entry.
• Prefix length—Length of the IPv6 address prefix • Metric—Cost of a route RIPng packet processing procedure Request packet When a RIPng router first starts or must update entries in its routing table, it usually sends a multicast request packet to ask for needed routes from neighbors. The receiving RIPng router processes RTEs in the request.
OSPFv3 The term "router" in this document refers to both routers and LB modules. Overview Open Shortest Path First version 3 (OSPFv3) supports IPv6 and complies with RFC 2740 (OSPF for IPv6).
LSA types OSPFv3 sends routing information in LSAs, which, as defined in RFC 2740, have the following types: • Router-LSA—Originated by all routers. This LSA describes the collected states of the router's interfaces to an area, and is flooded throughout a single area only. • Network-LSA—Originated for broadcast and NBMA networks by the Designated Router. This LSA contains the list of routers connected to the network, and is flooded throughout a single area only.
LSA delay timer Each LSA has an age in the local link state database (LSDB) (incremented by one per second), but an LSA does not age on transmission. You must add an LSA delay time into the age time before transmission, which is important for low-speed networks. SPF timer Whenever the LSDB changes, an SPF calculation happens. If recalculations become frequent, a large amount of resources are occupied.
IPv6 BGP This chapter describes only configuration for IPv6 BGP. For BGP-related information, see Network Management Configuration Guide. The term "router" in this document refers to both routers and LB modules. IPv6 BGP overview BGP-4 can only carry IPv4 routing information. To support multiple network layer protocols, IETF extended BGP-4 by introducing Multiprotocol Border Gateway Protocol (MP-BGP). MP-BGP for IPv6 is called "IPv6 BGP" for short.
Support and other resources Contacting HP For worldwide technical support information, see the HP support website: http://www.hp.
Conventions This section describes the conventions used in this documentation set. Command conventions Convention Description Boldface Bold text represents commands and keywords that you enter literally as shown. Italic Italic text represents arguments that you replace with actual values. [] Square brackets enclose syntax choices (keywords or arguments) that are optional. { x | y | ... } Braces enclose a set of required syntax choices separated by vertical bars, from which you select one.
Network topology icons Represents a generic network device, such as a router, switch, or firewall. Represents a routing-capable device, such as a router or Layer 3 switch. Represents a generic switch, such as a Layer 2 or Layer 3 switch, or a router that supports Layer 2 forwarding and other Layer 2 features. Represents a security product, such as a firewall, a UTM, or a load-balancing or security card that is installed in a device.
Index BCDILMOPRST B P BGP configuration views,28 Packets,35 BGP load balancing,23 Protocols and standards,28 BGP message types,18 Protocols and standards,9 BGP path attributes,19 Protocols and standards,17 BGP route advertisement rules,23 Protocols and standards,34 BGP route selection,23 Protocols and standards,37 BGP speaker and BGP peer,18 R C Related information,39 Contacting HP,39 RIP message format,8 Conventions,40 RIP operation,7 D RIP route entries,6 RIP timers,6 DR and BDR,16
