Presentation is loading. Please wait.

Presentation is loading. Please wait.

Alcatel-Lucent Routing Protocols

Similar presentations


Presentation on theme: "Alcatel-Lucent Routing Protocols"— Presentation transcript:

1 Alcatel-Lucent Routing Protocols
Module 1 — Introduction Module 2 — Static Routing and Default Routes Module 3 — Routing Information Protocol Module 4 – Link-State Protocols Module 5 — Open Shortest Path First Module 6 — Intermediate System–to–Intermediate System Module 7 — Border Gateway Protocol

2 Alcatel-Lucent Routing Protocols
Module 1 — Introduction

3 IP Addressing — Basic Subnetting
Subnetting allows a network to be subdivided into smaller networks with routing between them. With basic subnetting, each segment uses the same subnet mask. Potential for wasting IP addresses on links that do not require high client density Easiest to implement Required for classful routing protocols VLSM allows the use of different subnet masks for different parts of the network. Basic subnetting allows for each segment to have an identical subnet mask. This means that both the topology’s Ethernet segments and the point-to-point WAN segments must use the same subnet mask. Using this method, administrators tend to waste many usable IP addresses as they are lost on the WAN interfaces. Basic subnetting is not a recommended implementation of the allocation of IP networks in a topology. However, if the routing protocol in question is classful, there is no choice but to use basic subnetting. An example is on the following page.

4 Different subnet masks per network
IP Addressing — VLSM Different subnet masks per network Routing protocols must advertise the subnet mask with updates More efficient use of IP addressing than basic subnetting Requires a good understanding of subnetting RFC 1878 defines VLSM Routing protocols that support VLSM are: RIPv2 OSPF IS-IS BGP VLSM allows for each network in a major network to use a different subnet mask. As an update is sent by a routing protocol, it includes the subnet mask with the network advertisement. This allows the receiving router to forward traffic based on the longest-matching IP network entry compared to the destination IP address in the packet being forwarded. When a VLSM network is being designed, it is important to keep in mind that the high-order bits cannot be reused after they are allocated. This requires that the network administrator have a good understanding of IP subnetting.

5 IP addresses are broken into classes: A, B, C, and D
IP Addressing Review IP addresses are broken into classes: A, B, C, and D Class A: or /8 Network Host Class B: or /16 Network Host Class C: or /24 Network Host An IP address is 32 bits long and has two parts: Network number Host number The address format is known as dotted-decimal notation. Example address: The minimum value for an octet is 0; it contains all 0s. The maximum value for an octet is 255; it contains all 1s. IP addressing is divided into five ranges. Classes A, B, and C are unicast IP addressing ranges. This means that they are used to identify only a single device. Class D addressing is used for multicast traffic (from one source to multiple destination devices, but not all devices). Class E addressing is used for research and development. An easy way to remember how many bytes (octets) are used to identify the classful network portion of an IP address is the following. Because “A” is the first letter in the alphabet, only the first byte (high-order byte) is used to identify the network. The remaining 3 bytes are used to identify the clients. Because “B” is the second letter in the alphabet, the first two bytes identify the network and the remaining two bytes identify the clients in the network. “C” is the third letter in the alphabet; therefore, the first three bytes identify the network and the fourth byte identifies the hosts in the networks. “D” is the fourth letter in the alphabet; this means that all four bytes (all 32 bits) are used to identify the multicast stream of data being sent. Class D: or /32 Multicast

6 Introduction to IP routing
Section Objectives Introduction to IP routing Review of IP forwarding Control plane vs. data plane functions Common layer 3 routing protocols Distance vector Link state Classful and classless addressing Variable length subnet masking Classless interdomain routing Private IP addresses Network address translation (NAT/PAT) Section 2 — Introduction to IP Routing This section provides an overall review of Networking Fundamentals. Within this section students will review the OSI model, basic Network Flow, and flow of a packet through a 7750 SR chassis. It will also review concepts relating to Classful and Classless protocols and the differences between Distance Vector and link-state protocols. These concepts are a necessary prerequisite to the discussion on routing protocols that will follow.

7 Movement of Data 1.1.1.2 2.2.2.2 (MAC address = A) (MAC address = D)
(MAC address = C) (MAC address = B) Data Source Dest. S D A B F C The basic flow of a packet of data through a network is as follows: Device A ( ) wants to send data to server D ( ). Because device A is not located on the same segment as that of device D, it must use the default gateway for the segment. This default gateway is seen as IP address in the slide above. Device A will ARP the address to learn the MAC address of the gateway. The router responds with the MAC address of “B”. Device A is now able to encapsulate the data, as shown in the top block diagram. Note that the source and destination IP addresses identify the overall source and destination devices whereas the frame source and destination addresses identify the path across the Ethernet segment only. When the data arrives at the left router (router B), the router removes the L2 header and trailer, checks its routing table, and determines that the data needs to be sent to the right router (router C). To accomplish this, router B encapsulates the data in a PPP frame of data and forwards it. Router C removes the PPP frame and consults its routing table. Seeing that the destination IP network is directly connected to its Ethernet port, router C creates the frame of data and forwards it to device D. Note that the IP addressing did not change throughout this movement of data. However, the L2 framing changed over each segment that the packet traversed. It can be surmised that the IP address identifies a device in the entire network topology whereas the L2 address identifies a device on that segment only. Data Source Dest. WAN PPP F C S Data Source Dest. S D C F

8 When a router receives a packet, it:
Packet Forwarding When a router receives a packet, it: Compares the destination IP address of the packet to the FIB Looks for the longest (most specific) match If no match is found, the packet is dropped. If the packet is to be forwarded, the next hop and egress interface must be known. If a match is found, the packet is sent to the next-hop address via the interface specified in the FIB. The next-hop is the next router in the path toward the destination. The egress interface is required for encapsulation. When a router receives a packet that is to be forwarded, the basic forwarding rules are as follows. First, the router compares the destination IP address in the received packet to the entries in the FIB, commonly called the routing table. The router performs a bit-by-bit comparison of the destination IP address to these entries, searching for the entry that matches with the most common bits. This is called longest-match or most-specific-match routing. If no entries match any bits, the packet is discarded, and in most cases an ICMP-unreachable message is generated back to the source. When a matching entry is located, the router requires 2 elements from the matching entry in the FIB to forward the packet. The egress or outbound interface for the packet must be known, so the router may encapsulate the packet in the appropriate frame for forwarding toward the destination. Secondly, the next-hop IP address must be known so that the frame may be addressed and forwarded to the next router (unless the packet destination is local to the router, in which case the frame is addresses to the destination itself). The egress interface is usually listed in the FIB associated with any route entry, based on the interface on which the route was learned. The IP next hop may also be located in the FIB or may be stored in the router’s memory. The next hop is usually derived from the source IP address of the routing update received because neighbor routers in the IGP world are directly connected. This process traditionally occurred for each packet received by the router, independently from the next packet. In today's advanced hardware, there are many optimizations designed into the router hardware and software to streamline the lookup procedure.

9 Common IP Routing Protocols
Legacy routing protocols: RIP version 1 RIP version 2 Modern routing protocols: OSPF IS-IS BGP Over the lifetime of IP there have been multiple routing protocols that have come and gone. Today there are a handful of accepted protocols that are supported and commonly accepted. The oldest of these is RIP. There are two versions of RIP, the older RIPv1 and the slightly more modern RIPv2. Typically, RIP is called a legacy protocol because it has been around so long and because it uses older methods for calculating where networks are and how best to access them. RIP is covered in more depth in a later module. Modern IP routing protocols include OSPF, IS-IS, and BGP. OSPF and IS-IS are used in large network topologies, have very fast convergence and, compared to RIP, use more modern methods for calculating where networks are and how best to access them. These protocols are covered individually in later modules. BGP is the routing protocol of the Internet. It is a very stable and robust routing protocol. For an in-depth understanding of BGP, it is recommended that you take the BGP course.

10 Distance Vector Protocols
Distance = How far away Vector = What direction (interface) RIPv1, RIPv2, and BGP are distance vector protocols Int 1/1/2 IP – Int 1/1/2 IP – Int 1/1/1 Int 1/1/1 IP – Distance vector routing algorithms (Bellman-Ford) pass periodic copies of a routing table from router to router. Regular (timed-interval) updates between routers communicate topology changes. With distance vector routing protocols, no routing table is transmitted beyond the immediate neighbor. The distance vector algorithm allows network metrics to accumulate and maintains a table showing the next hop for all destinations listed in the table. To begin the distance vector routing process, the router must be configured with two items of information: first, the IP addresses of all directly connected networks that the router needs to communicate about. These addresses become permanent entries in the routing table. Second, the router must be configured with the name of the routing protocol being used to send and receive updates (RIPv1 or RIPv2). The basic operation is as follows: Each router sends its entire routing table to each of its adjacent neighbors. This table includes reachable addresses, a value that represents the distance metric, and the IP address of the first router on the path to each network that it knows about. As each router receives an update from its neighbor, it calculates a new routing table and transmits that to each of its neighbors at the next timed interval. In a very large network with many routers, this can take quite a while. Convergence is the state that occurs when all routers in an internetwork have been updated after a topology change and normal routing operations have resumed. The time to reach convergence is a major factor in internetwork performance. Fast convergence is preferred and is a prime criterion in evaluating a routing protocol. IP – Routing Table: – Direct 1/1/2 – Direct 1/1/1 – 1 hop via 1/1/1 Routing Table: – Direct 1/1/2 – Direct 1/1/1 – 1 hop via 1/1/1

11 State = Active or inactive interface
Link-State Protocols Link = An interface State = Active or inactive interface OSPF and IS-IS are link-state protocols More complex than distance vector Faster convergence Triggered updates Three databases: Adjacency — Neighbor database Topology — Link-state database Routing — Forwarding database Link-state, also known as SPF, maintains a complex database of topology information. Whereas distance vector has nonspecific information about distant networks and no knowledge of distant routers, link-state maintains full knowledge of distant routers and how they interconnect. OSPF and IS-IS are examples of link-state routing protocols. LSPs are used to transmit the information necessary to build a topological database, which is used by the SPF algorithm to construct a SPF tree, and finally, a routing table of paths and ports to each network. When a link-state topology changes, the routers must become aware of the change and send information to other routers or to a designated router that all other routers can use for updates. This involves the propagation of common routing information to all routers in the network. To achieve convergence, each router does the following: Keeps track of its neighbors Constructs an LSP that lists neighbor router names and link metrics (cost). This includes new neighbors, changes metrics, and links to neighbors that have gone down. Sends out this LSP so that all routers receive it When it receives an LSP, records the LSP in its database so that it can store the most recent LSP received Using accumulated LSP data to construct a complete network topology, proceeds from the common starting point for the SPF algorithm and compute routes to every network Each time an LSP causes a change to the link-state database, the link-state algorithm recalculates the best paths and updates the routing table.

12 Link-State Protocols (continued)
Adjacency database Link-state database Forwarding database RTR - C Network /24 1/1/2 RTR - A RTR - B 1/1/1 Adjacency Database RTR-B – on 1/1/1 RTR-C – on 1/1/2 Link-state protocols keep three databases in the routers: The Adjacency database, sometimes called the neighbor database, keeps track of all other routers that are directly attached and passing link-state routing information. The LSDB has all the learned paths to all the destination networks. It is this database that is used to create the SPF tree that ultimately creates the routing table. The routing table, sometimes called the forwarding database, is used by the router to accurately forward IP packets to the destination network. /24 – via 1/1/1 cost 20 – via 1/1/2 cost 40 LSDB Routing Table: /24 – via 1/1/1

13 Routing Table Management
Each routing protocol populates its routes into its RIB. Each protocol independently selects its best routes based on the lowest metric. The best routes from each protocol are sent to the RTM. OSPF RIB RIP RIB RTM When a routing protocol learns routes from neighbors, it populates these routes into its RIB. The protocol’s RIB is where each protocol stores the routes it has learned from its neighbors. For each destination in the RIB, the routing protocol selects the best route based on the lowest metric. These best routes are sent to the RTM. Multiple routes for the same destination can be learned by the router. If these routes are learned from the same routing protocol, the metric for the protocol is used as a selection criterion. The route with the lowest metric is selected as the best route and is sent to the RTM. If there are multiple routing protocols in use, each protocol independently selects its best route based on the lowest metric from its RIB. There are now multiple best routes (one from each protocol), and each protocol sends its best route to the RTM. The RTM can select only one of these best route as there can only be one best route in the routing table for each destination.

14 The RTM may have a best route from multiple protocols.
Preference The RTM may have a best route from multiple protocols. Selection is based on lowest preference value. The RTM sends its best route to the FIB. This route is the active route and is used for forwarding. RIP RIB FIB Because metrics from different protocols are not comparable, the RTM uses preference to select from all the best routes it receives. The lower the protocols preference, the more likely that the best or active route will be selected from that protocol. Different protocols should not be configured with the same preference. If this occurs, the tie-breaker is based on the default preference table, shown on the next page. If multiple routes are learned with an identical preference, using the same protocol and with equal metrics, the best-route decision is determined by the configuration of ECMP in the config>router context. The best routes from the RTM are placed in the FIB, also commonly referred to as the routing table. The FIB is distributed to the various line cards on the 7750 SR. OSPF RIB OSPF RTM BGP RIB BGP OSPF

15 Default Preference Table
Route type Preference Configurable Direct attached No Static 5 Yes OSPF internal 10 IS-IS Level 1 internal 15 IS-IS Level 2 internal 18 RIP 100 OSPF external 150 IS-IS Level 1 external 160 IS-IS Level 2 external 165 BGP 170 The slide above shows the default preference values assigned to each routing protocol on the 7750 SR. All preference values except direct attached are configurable, so preferences other than the default may be used.

16 IP Addressing — Classful and Classless
/16 /24 Routing Table: – direct 1/1/2 – direct 1/1/1 – 1 hop via 1/1/1 /24 /24 Classless /24 /24 /16 /24 /24 Classful routing protocols such as RIPv1 allow only FLSMs. When a mask is selected to support a bit boundary, no other mask can be used for that network number. These routing protocols do not support manual summarization but always summarize to a classful A, B, or C boundary. Inefficient use of the address space results from this type of addressing scheme. Classless routing protocols such as OSPF and IS-IS may use VLSMs and also support manual summarization. RIPv2 also allows VLSM. This results in more efficient use of the IP addresses. Routing is also more efficient with classless routing protocols because a mask is always sent along with the routing update to avoid confusion. This also allows for manual route summarization to take place based on the network administrator’s needs and not based on IP address network boundaries. Routing Table: /16 – direct 1/1/2 /24 – direct 1/1/1 /24 – 2 hops via 1/1/1 /24 – 1 hop via 1/1/1 /24 /24

17 Different subnet masks per network
IP Addressing — VLSM Different subnet masks per network Routing protocols must advertise the subnet mask with updates. High-order bits are not reusable. Routing decisions are made based on the longest match. A more efficient use of IP addressing than basic subnetting Requires a good understanding of subnetting RFC 1878 defines VLSM. Routing protocols that support VLSM are: RIPv2 OSPF IS-IS BGP VLSM allows each network in a major network to use a different subnet mask. As an update is sent by a routing protocol, it includes the subnet mask with the network advertisement. This allows the receiving router to forward traffic based on the longest-matching IP network entry compared to the destination IP address in the packet that is being forwarded. When a VLSM network is designed, it is important to keep in mind that the high-order bits cannot be reused once they are allocated. This requires that the network administrator have a good understanding of IP subnetting.

18 IP Addressing — VLSM Example
– – Reserved for WAN segments – hhhhhhhh – First Ethernet segment …. – hhhhhhhh – Last Ethernet segment – – Ethernet mask – hh – First WAN segment – hh – Last WAN segment – – WAN mask In the example above, classful network has been assigned for use in the network topology. It has been determined that each Ethernet segment must support up to 250 clients. In addition, all serial links are point-to-point; therefore they only require IP addressing to support 2 hosts per WAN segment. Considering that the number of bits required to identify 250 host addresses is 8 bits, it can be determined that the entire last byte must be allocated for host identification on all Ethernet networks in the topology. Considering that the given address is a Class B address and that only 8 bits must be used for the host field, it can be determined that the entire third byte can be used for identifying the subnets. Keep in mind that there is also the requirement to support the WAN links. The first available subnet is that of /24. This subnet could easily be confused with the overall classful network of /16, so this first subnet is reserved for future use with the WAN interfaces. That leaves networks /24 to /24 for allocation to the Ethernet segments. Each of these subnets can support up to 254 host addresses. Keep in mind that the routers and switches connected to these subnets typically require an IP address too. The /24 subnet that was reserved now comes into play. Considering that each WAN link only requires 2 hosts, there is only a need to have 2 host bits. This means that the subnet mask can be further expanded from a /24 mask to a /30 mask. The WAN subnet ranges are from /30 (hosts 5 and 6, directed broadcast 7) to /30 (hosts 253 and 254, directed broadcast 255). Keep in mind that with the WAN subnetting, the higher-order bits were not reused. All that was accomplished was the borrowing of more bits from the host field. The maximum number of bits that can be borrowed is 30 because there must always be at least 2 bits remaining to support the hosts on a segment.

19 Alcatel-Lucent Routing Protocols
Module 2 — Static Routing and Default Routes

20 What a Router Needs to Know
/24 /24 R1 R2 /30 Routing Table: /24 – Direct /30 – Direct /24 – static via Routing Table: /24 – Direct /30 – Direct /24 – static via A router’s primary purpose is to forward data from one location to another. Routers do this by learning the location of networks in a topology. There are two ways this can be accomplished. First, a router can communicate with other routers via a dynamic routing protocol. The second is to have an administrator program in a static route. Static routes are manually configured entries in the router that define a network and the path to take to access the network. Alcatel-Lucent routers support static routes. The command structure is simple and easily understood. In the example above, both the R1 and R2 routers have routing tables. Both routers automatically know about the networks they are directly connected too. This is accomplished by configuring the IP address and mask on the interfaces. In addition, both have been configured with static routes to the remote Ethernet segments. The routing table reflects this configuration with route entries. Routers need to know where networks are located and how best to access them. This can be accomplished statically with administrative commands.

21 Static Routes — Basic Static Routes
static-route /0 next-hop /24 R1 R2 Corporate Headquarters static-route /24 next-hop In the figure above, the corporate router has an explicit static route that points to the branch site’s stub Ethernet segment. The branch site has a default static route that points to the corporate headquarters. With this configuration, there is no need to have a dynamic routing protocol running between the two routers. This reduces overhead and configuration requirements in the two routers. The branch router will have only three entries in its routing table: the two directly connected networks and the static route. The corporate router will have all routes in the corporate topology as well as the static route that points to the branch site’s stub Ethernet. Configuration of static routes between stub networks and corporate locations

22 Static Routes — Configuration Example
/24 R1 R2 Corporate Headquarters In the figure above, the corporate router has an explicit static route that points to the branch site’s stub Ethernet segment. The branch site has a default static route that points to the corporate headquarters. With this configuration, there is no need to have a dynamic routing protocol running between the two routers. This reduces overhead and configuration requirements in the two routers. The branch router will have only three entries in its routing table: the two directly connected networks and the static route. The corporate router will have all routes in the corporate topology as well as the static route that points to the branch site’s stub Ethernet. config>router> static-route /24 next-hop config>router> static-route /0 next-hop

23 Default Routes — Basic Default Route
/24 R1 R2 Corporate Headquarters R2# show router route-table ============================================================================ Route Table Dest Address Next Hop Type Protocol Age Metric Pref /24 System Local Local 01d02h 0 0 /24 System Local Local 08d03h / Remote Static 01d02h The above show command from the R2 router lists the entries for the directly connected networks and the static default route. When traffic arrives on the branch Ethernet interface with a destination unknown to the router, the router forwards the traffic to the corporate router because that is defined as the default route. The routing table shows the two directly connected networks ( and ) as well as the default static route that points to corporate headquarters.

24 Static Routes — Floating Static Routes
Backup /24 R1 R2 Corporate Headquarters Primary path config>router> static-route /24 next-hop config>router> static-route /24 next-hop preference 200 In the example above, the R1 router has two static routes configured. The top static route points all traffic destined to the branch site’s Ethernet segment out the primary path. Because there is no preference setting explicitly stated on this line of the configuration, the default preference value of 5 is used. On the next line, the R1 router has configured a second static route to the branch site’s Ethernet segment. However, in this case the preference value is set to 200. The static route with the lowest preference value is the preferred route. As a result, all traffic will use the primary path as long as it is operational. Should that link fail, the static route that points out the interface will be removed and the static route across the backup link will be added to the routing table and become operational. This secondary path, when configured with static routes, is commonly called a floating static route. For this to work, there needs to be a route for the opposite direction. Therefore, the branch router also needs the following static routes configured: config>router> static-route /0 next-hop config>router> static-route /0 next-hop preference 200 Note: If the interface connections are connected to an intermediate third device such as a switch or hub, one router interface could fail while the router at the other end still has an active interface. When that happens, traffic may flow in one direction only, making it impossible to establish a connection over this link. Configuration of a floating static route between stub networks and corporate locations

25 Static Route Verification — Show Command
The command below shows static routes configured in the routing table. Context: show>router> Syntax: static-route [[ip-prefix [/mask]] | [preference preference] | [next-hop ip-addr] | tag tag Example: R1# show router route-table protocol static ============================================================================== Route Table (Router: Base) Dest Address Next Hop Type Proto Age Metric Pref / Remote Static 00h01m34s / Remote Static 00h01m15s No. of Routes: 1 Syntax [no] static-route [[ip-prefix [/mask]] | [preference preference] | [next-hop ip-addr] | tag tag ] Context show>router Description This command lists the static entries in the routing table. If no options are present, all static routes are displayed, sorted by prefix. Parameters I p-prefix[/mask] — Lists only static routes that match the specified ip-prefix and optional mask. preference — Lists only static routes with the specified route preference. The range is from 0 to Next-hop ip-addr — Lists only static routes with the specified next-hop IP address.

26 Static Route Verification — Show Command (continued)
/24 Corporate Headquarters R1 R2 R1# show router route-table /24 ============================================================================== Route Table (Router: Base) =============================================================================== Dest Address Next Hop Type Proto Age Metric Pref / Remote Static 00h02m54s No. of Routes: 1 The example above shows the use of the “show router static-route” command as applied to the corporate router. Only the static routes are shown in the example. Note that the command specified the static route to network , so no other networks are shown.

27 Static Routes — Ping Command
/24 Corporate Headquarters R1# ping detail PING : 56 data bytes 64 bytes from via fei0: icmp_seq=0 ttl=64 time=0.000 ms. 64 bytes from via fei0: icmp_seq=1 ttl=64 time=0.000 ms. 64 bytes from via fei0: icmp_seq=2 ttl=64 time=0.000 ms. 64 bytes from via fei0: icmp_seq=3 ttl=64 time=0.000 ms. 64 bytes from via fei0: icmp_seq=4 ttl=64 time=0.000 ms. PING Statistics ---- 5 packets transmitted, 5 packets received, 0.00% packet loss round-trip min/avg/max/stddev = 0.000/0.000/0.000/0.000 ms R1# Syntax ping {ip-address | dns-name} [rapid | detail] [ttl time-to-live] [tos type-of-service] [size bytes] [pattern pattern] [source ip-address] [interval seconds] [{next-hop ip-address} | {interface interface-name} | bypass-routing] [count requests] [do-not-fragment] [router [router-name | service-id] Context <GLOBAL> Description This command is the TCP/IP utility to verify IP reachability. Parameters ip-address | dns-name — The remote host to ping. ttl time-to-live — The IP TTL value to include in the ping request. The range is from 1 to 128. tos type-of-service — The ToS bits in the IP header of the packets. The range is from 0 to 255. size bytes — The size of the ping request packets in bytes. The default is 56 bytes. pattern pattern — A 16-bit pattern string to include in the ping packet, expressed as a decimal integer. source ip-address — The source IP address to use in the ping requests. The default is the egress interface. interval seconds — The interval in seconds between consecutive pings requests. The default is 1 second. next-hop ip-address — This option disregards the routing table and will send this packet to the specified next-hop address. interface interface-name — Specifies the interface name. bypass-routing — Send the ping request to a host in a directly attached network, bypassing routing. count requests — The number of pings sent to the remote host. The default is 5. do-not-fragment — Specifies that the request frame should not be fragmented. router router-name — Specifies the alphanumeric character string up to 32 characters. router service-id — Specifies a numeric string that identifies the service. Note: The source IP address used for the ping command is the interface IP address when pinging a directly connected interface. The source is the system interface when pinging IP addresses that are multiple hops away. You can specify the source address when using the ping command to change this behavior.

28 Static Routes — Traceroute Command
/24 R1 R2 Corporate Headquarters R1# traceroute traceroute to , 30 hops max, 40 byte packets <10 ms <10 ms <10 ms <10 ms <10 ms <10 ms Syntax traceroute {ip-address | dns-name} [ttl ttl] [wait milliseconds] [no-dns] [source ip-address] [tos type-of-service] Context <GLOBAL> Description This command is the TCP/IP traceroute utility to determine the route to a destination address. Note that cancelling a traceroute with the <Ctrl-C> command could require issuing a second <Ctrl-C> command before the prompt is returned. Parameters ip-address | dns-name — The remote address to traceroute. The IP address or DNS name (if DNS name resolution is configured) can be specified. ttl ttl — The maximum TTL value in the traceroute request. The range is from 1 to 255. wait milliseconds — The time to wait for a response to a probe, in milliseconds. The range is from 1 to no-dns — When specified, a DNS lookup for the specified host name is not performed. source ip-address — The source IP address to use as the source of the probe packets. tos type-of-service — The ToS bits in the IP header of the probe packets.

29 Learning Assessment Do static routes have a higher or lower preference value than dynamic routes? What is the command syntax to create a static route in the 7750 SR? A router has a default route, a static route to /24, and a route to /14 learned from RIP. Which route is used for a packet with destination address ?

30 Alcatel-Lucent Routing Protocols
Module 3 — Routing Information Protocol

31 Distance vector overview
Section Objectives Distance vector overview Split horizon Route poisoning Poison reverse Hold-down timers Section 1 — Distance Vector Overview This section discusses the attributes of a distance vector routing protocol. Specifically, the topics covered are: Split horizon Route poisoning Poison reverse Hold-down timers

32 Distance Vector Overview
Routers send periodic updates to physically adjacent neighbors Updates contain the distance (how far) and vectors (direction) for networks RTR-B RTR-A 100 Mb/s 1 Gb/s Distance vector routing algorithms (Bellman-Ford) pass periodic copies of a routing table from router to router. Regular (timed-interval) updates between routers communicate topology changes. Each router receives a routing table from its direct neighbor. In the figure above, RTR-B receives information from RTR-A. RTR-B uses the information received from RTR-A to recalculate its routing table. RTR-B then sends its routing table to RTR-D. This same step-by-step process occurs in all directions between neighboring routers. Note: With distance vector, no routing table is transmitted beyond the immediate neighbor. In the example, RTR-D will never see a routing update sent directly from RTR-A. The distance vector algorithm allows network metrics to accumulate and maintains a table showing the next hop for all destinations listed in the table. 1 Gb/s 1 Gb/s RTR-C RTR-D

33 Distance Vector Overview (continued)
The router processes and compares the information contained in the routing update received with what is in its routing table. Process and compare with routing table Periodic update Sent to neighbor routers The figure above shows the step-by-step process with distance vector for updating all routers in an internet when a topology change occurs. Each router sends its entire routing table to each of its adjacent neighbors. This table includes reachable addresses, values that represent the distance metric, and the IP address of the first router on the path to each network that it knows about. As each router receives an update from its neighbor, it calculates a new routing table and transmits it to each of its neighbors at the next timed interval. In a very large network with many routers, this can take quite a while. Update from neighbor

34 Split Horizon An adjacent router does not advertise networks back to the source of the network information. – 2 hops – 1 hop X RTR-A RTR-B RTR-C Split horizon is a loop-avoidance technique for physically adjacent devices. In simplistic terms, split horizon states that an adjacent router will not re-advertise learned networks to the router that originally advertised the network. Without this policy, routers would be susceptible to routing loops. If RTR-C lost network and if RTR-B did not block re-advertisements to RTR-C, RTR-C might think that network was accessible via RTR-B. This would cause a loop and a major disruption in traffic flow. To ensure that does not happen, all routers that run a distance vector protocol support split horizon. Routing Table: – 2 hops via 1/1/1 Routing Table: – 1 hop via 1/1/1 Routing Table: – 0 hops via 1/1/1

35 Route Poisoning When a network goes away, the sourcing router sets the hop value to infinity and sends a triggered update to its neighbors. – 16 hops – 16 hops X RTR-A RTR-B RTR-C Routing Table: – 16 hops via 1/1/1 Routing Table: – 16 hops via 1/1/1 Routing Table: – 16 hops via 1/1/1 Route poisoning is used to speed up convergence. When used in conjunction with triggered updates, the network converges more quickly. Route poisoning is performed by the router directly connected to the network that goes away. When it determines that the network is not accessible, the router sets the hop count to infinity (16 hops for RIP) and sends a message to all directly attached neighbors. These neighbors change their routing tables and forward the message to their neighbors on all other links. Keep in mind that split horizon still applies when a route poison advertisement is forwarded. In the example above, RTR-C’s Ethernet fails. RTR-C sets its routing entry to infinity and sends an update to RTR-B. RTR-B changes its routing table entry and forwards the change to RTR-A. This ensures that all routers learn of the topology change, and by keeping the route in the routing table, decreases the possibility of creating a false path to network Routing Table: – 2 hops via 1/1/1 Routing Table: – 1 hop via 1/1/1 Routing Table: – 0 hops via 1/1/1

36 Poison Reverse Poison reverse is the only time that split horizon is violated. This helps to avoid loop creation when a network fails. — 16 hops Poison reverse — 16 hops Poison reverse — 16 hops — 16 hops X RTR-A RTR-B RTR-C Poison reverse is the only time that split horizon is violated in a distance vector routing protocol environment. The idea of poison reverse is to confirm to the preceding device that the update about a network going away has been recorded. This response to the originator also ensures that a loop-free topology is created. In the example above, RTR-C’s link to network fails. RTR-C sets its routing entry to infinity and sends an update to RTR-B. RTR-B changes its routing table entry and forwards the change to RTR-A. RTR-B also sends a poison reverse message back out the interface that RTR-C’s message came in on. This ensures a loop-free topology. When RTR-A gets the route poisoning message from RTR-B, it too sends a poison reverse message back on the interface that the message was received on. Routing Table: — 16 hops via 1/1/1 Routing Table: — 16 hops via 1/1/1 Routing Table: — 16 hops via 1/1/1 Routing Table: — 2 hops via 1/1/1 Routing Table: — 1 hop via 1/1/1 Routing Table: — 0 hops via 1/1/1

37 Hold-Down Timers Hold-down timers provide time for other routers to converge and reduce loops from being created when a network fails. — 16 hops — 16 hops X RTR-A RTR-B RTR-C Hold-down timers keep the failed network in the routing table, with the hop count set to infinity, for a predetermined interval (usually 180 seconds for RIP). This allows time for the other routers in the network to receive the topology change update without causing loops. In the example above, RTR-C’s link to network fails. RTR-C sets its routing entry to infinity and sends an update to RTR-B. RTR-B changes its routing table entry by changing the metric to infinity, and it starts its hold-down timer. It does not remove the route until the hold-down timer has expired. This ensures that all routers learn of the topology change without causing a loop during convergence. Keeping the route in the routing table decreases the possibility of creating a false path to network Routing Table: – 16 hop – Via 1/1/1 Routing Table: — 2 hops via 1/1/1 Routing Table: – 16 hop – Via 1/1/0 Routing Table: — 1 hop via 1/1/1 Routing Table: — 0 hops via 1/1/1 Routing Table: – 16 hop – Via 1/1/1 Hold-down timer 180 seconds Hold-down timer 180 seconds Hold-down timer 180 seconds

38 Combined Loop Avoidance Techniques
Combined, all attributes function as follows: — 16 hops Poison reverse — 16 hops Poison reverse — 16 hops — 16 hops X RTR-A RTR-B RTR-C When combined, the mixture of route poisoning, poison reverse, triggered updates, and hold-down timers provides a robust loop avoidance technique for when routes fail in a network. In the example above, RTR-C’s link to network fails. RTR-C sets its routing entry to infinity, sets the hold-down timer, and sends an update to RTR-B. RTR-B changes its routing table entry, sets the hold-down timer, and forwards the change to RTR-A. RTR-B also sends a poison-reverse message back out the interface that RTR-C’s message came in on. When RTR-A gets the route-poisoning message from RTR-B, it also sends a poison-reverse message back on the interface that message was received on. In addition, RTR-A modifies the routing entry by setting it to infinity and invokes its hold-down timer. Routing Table: – 16 hop – Via 1/1/0 Routing Table: — 2 hops via 1/1/1 Routing Table: – 16 hop – Via 1/1/1 Routing Table: — 1 hop via 1/1/1 Routing Table: — 0 hops via 1/1/1 Routing Table: – 16 hop – Via 1/1/0 Hold-down timer 180 seconds Hold-down timer 180 seconds Hold-down timer 180 seconds

39 Uses a hop-count metric
RIP Overview Uses a hop-count metric Sends updates of the routing table to neighbors Maximum of 15 hops; 16 hops equals infinity 30-second advertisement interval by default Authentication is available in RIPv2 VLSM is supported by RIPv2 RIPv1 was originally outlined in June 1988 and is defined in RFC 1058. RIP is an interior gateway protocol that uses a distance vector algorithm to determine the best route to a destination, using hop count as the metric. A hop is a network layer device, such as a router. For the protocol to provide complete information about routing, every router in the domain must participate in the protocol. RIP is a routing protocol based on a distance vector (Bellman-Ford) algorithm, which advertises network reachability by advertising the prefix/mask and the metric (also known as hop count or cost). RIPv1 uses broadcast updates to advertise the networks. In the updates, the maximum number of networks that can be advertised per packet is 25. Therefore, if a router needs to advertise 30 networks to its peers, it sends 2 packets every 30 seconds. The first contains 25 network entries and the second contains the remaining 5 network entries. Alcatel-Lucent supports modification of this parameter to a maximum of 255 network entries per packet. By default, RIP advertises all RIP routes to each peer every 30 seconds. In RIP, the hop metric is limited to a maximum value of 15 hops. Another way to look at this is that networks can be no more than 15 routers away. To indicate that a network is unreachable, the hop value is set to 16, which equates to infinity for RIP. Each router along the path increments the hop-count value by 1. The maximum number of hops in a path is 15. If a router receives a routing update with a metric of 15 and that contains a new or modified entry, increasing the metric value by 1 will cause the metric to increment to 16 (infinity), resulting in the destination being considered unreachable. The 7750 SR implementation of RIP uses split horizon with poison reverse to avoid such problems as “counting to infinity”. Split horizon with poison reverse means that routes learned from a neighbor through a given interface are advertised in updates out of the same interface but with a metric of 16 (infinity). RIPv1 does not support any security or authentication mechanism. However, RIPv2 does support authentication. The 7750 SR software supports RIPv1 and RIPv2. RIPv1 is a classful routing protocol. It assumes the following netmask information for non-local routes, based on the class the route belongs to: Class A – 8-bit mask Class B – 16-bit mask Class C – 24-bit mask

40 RIP Overview (continued)
100 Mb/s RTR-B RTR-A 1 Gb/s 1 Gb/s RIP always chooses the path that has the fewest hops. Recall that a hop is a network-layer device. RIP does not take into account the bandwidth of the links or delays within the network-layer equipment. In the example above, RTR-A will choose the path across the 100 Mb/s link because the number of hops is lower than taking the path via RTR-C. Although the path chosen is dramatically slower than the path via RTR-C, RIP is only concerned with the fewest hops to a destination network. If the link between RTR-A and RTR-B fails, the alternate link via RTR-C would then be taken. RIP, a UDP-based protocol, sends updates to its directly attached neighbors, update their neighbors, and so on. Any router that runs RIP has a routing process that sends and receives RIP updates periodically on UDP port 520. A RIPv1 update can contain a maximum of 25 route advertisements and RIPv2 supports up to 255 routes per update. The format of RIPv1 and RIPv2 updates is slightly different. Additionally, RIPv1 updates are sent to a broadcast address whereas RIPv2 updates can be sent to either a broadcast or multicast address ( ). RIP is supported on all IP interfaces, including both network and access interfaces. 1 Gb/s RTR-C RTR-D

41 RIPv1 vs. RIPv2 RIPv1 RIPv2 Defined in RFC 1058
Defined in RFCs 1721, 1722, and 2453 Classful routing protocol Classless routing protocol No subnet mask in updates Sends subnet mask in updates Does not support VLSM Supports VLSM and CIDR No manual route summarization Manual route summarization Does not support authentication Supports authentication Broadcast updates Multicast or broadcast updates RIPv2 is defined in RFCs 1721, 1722, and 2453. Like RIPv1, RIPv2 is an IGP that uses a distance vector algorithm to determine the best route to a destination, using hop count as the metric. RIP is a routing protocol, based on a distance vector (Bellman-Ford) algorithm, that advertises network reachability by advertising the prefix/mask and the metric (also known as hop count or cost). RIPv2 was written after CIDR was developed and transmits netmask information with every route. Because of the support for CIDR routes and other enhancements in RIPv2, such as triggered updates, multicast advertisements, and authentication, most production networks now use RIPv2. RIPv2 supports subnet masks, a feature that was not available in RIPv1. A network address of is considered a default route. A default route is used when it is not convenient to list every possible network in the RIP updates, and when one or more closely connected gateways in the system are prepared to handle traffic to the networks that are not listed explicitly. These gateways create RIP entries for the address as if it were a network to which they are connected. RIPv2 updates can be sent to a broadcast or multicast address ( ). The 7750 SR defaults to using the broadcast address. RIPv2 supports the use of a secure authentication mechanism to obtain table updates. The 7750 SR OS implementation enables the use of a simple password (plain-text) or MD5 authentication. The 7750 SR OS allows you to specify the RIP version that is sent to RIP neighbors and RIP updates that are accepted and processed.

42 RIP – Major Component Configuration
Router Interface (assumed to be already complete) Route policies RIP Group Neighbor RIP configuration commands have three primary configuration levels: rip for global configuration, group group-name for RIP group configuration, and neighbor ip-int-name for RIP neighbor configuration. Within the different levels, the configuration commands are identical. For the repeated commands, the command that is most specific to the neighboring router is in effect (i.e., neighbor settings have precedence over group settings, which have precedence over RIP global settings). For a router to accept RIP updates, in the config>router>rip context, you must define at least one group and one neighbor. The 7750 SR ignores updates received from routers on interfaces that are not configured for RIP. Configuring other RIP commands and parameters is optional. By default, the local router imports all routes from this neighbor and does not advertise routes. The router receives both RIPv1 and RIPv2 update messages, with 25 to 255 route entries per message. This section provides information about configuring RIP and examples of common configuration tasks. Unlike other IGP protocols supported in the 7750 SR series, RIP does not automatically redistribute routing information to its neighbors. To ensure that the local IP addresses are in the local database, you must configure a routing policy and apply it as a RIP export policy. However, if the intent is to pass RIP updates through a router and not to advertise the directly connected networks, policy configuration can be bypassed. Interface — A logical IP routing interface. After this interface has been created, attributes such as IP address, port, link aggregation group, or system can be associated with the IP interface. RIP — The context for configuring a RIP protocol instance. When a RIP instance is created, the protocol is enabled by default. Group — RIP groups are a way of logically associating RIP neighbor interfaces to facilitate a common configuration for RIP interfaces. Neighbor — Specify an existing interface name of a neighbor to configure a RIP neighbor interface. By default, the local router imports all routes from the neighbor and does not advertise routes. This command can be issued multiple times to create multiple neighbor associations.

43 Alcatel-Lucent Routing Protocols
Module 4 – Link-State Protocols

44 Distance Vector vs. Link State
Views the network topology from the neighbor’s perspective Adds distance vectors from router to router Frequent, periodic updates: slow convergence Passes copies of the routing table to neighbor routers Has a common view of the entire network topology Calculates the shortest path to other routers Event-triggered updates: faster convergence Passes link-state routing updates to other routers Link state and distance vector can be compared in several key areas: Distance vector views and learns everything as "next hop“. Link state obtains a wide view of the entire internetwork topology by accumulating information about all links in the network. Distance vector determines the best path by adding to the metric value that it receives as tables move from router to router. With link state, each router calculates its own shortest path to destinations. Distance vector is a daisy chain of tables that are passed using periodic table updates. This leads to slow convergence, particularly in large networks. With link state, updates are triggered by topology changes. Relatively small LSPs are passed to all other routers or to a multicast group of routers, which usually results in faster time to converge.

45 Link State Overview Link state-driven updates, periodic hellos
Classless routing protocol Sends subnet mask in update Supports VLSM, CIDR, and manual route summarization Supports authentication Link-state protocols have the following common attributes: Link-state protocols trigger an update when a link (interface) changes state. The router connected to the link initiates a triggered update to its neighbors, notifying them of the topology change. If the network is stable and no changes in links are detected, the routers send periodic hello messages to maintain connectivity without having to consume excessive bandwidth. A common attribute shared by link-state protocols is that they are classless and support all the common attributes of a classless routing protocol. Because link-state protocols are classless, updates contain the subnet mask of each network being advertised. This allows for more optimum network design and accurate path selection. VLSM and CIDR are supported in all link-state protocols. Due to the classless aspects of link-state protocols, manual summarization is actively supported. This allows network administrators to have much more control of where and how the summarization takes place. All modern link-state protocols support authentication of the updates that are sent between the routers. This ensures that accurate network topologies are created without false information or errors. Link-state protocols maintain three common databases: topology (link-state DB), neighbor (adjacency DB), and routing table (forwarding DB). Modern link-state protocols use a multicast address to convey updates and hellos to their neighbor link-state routing peers. This reduces processing on devices in the network that are not running the link-state protocol. Maintains multiple databases Sends updates using multicast addressing

46 Link State Overview (continued)
Link = An interface State = Active or inactive interface, cost IS-IS and OSPF are link-state protocols More complex than distance vector Faster convergence Triggered updates Three databases: Adjacency – neighbor database Topology – link-state database Routing – forwarding database Link-state, also known as SPF, maintains a complex database of topology information. Whereas distance vector has nonspecific information about distant networks and no knowledge of distant routers, link state maintains full knowledge of distant routers and how they interconnect. OSPF and IS-IS are examples of link-state routing protocols. LSPs are used to transmit the information necessary to build a topological database, which is used by the SPF algorithm to construct an SPF tree, as well as a routing table of paths and ports for each network. When a link-state topology changes, the routers must become aware of the change and send information to other routers or to a designated router that all other routers can use for updates. This involves the propagation of common routing information to all routers in the network. To achieve convergence, each router does the following: Keeps track of its neighbors Constructs an LSP that lists neighbor router names and link metrics (cost) (this includes new neighbors, change metrics, and links to neighbors that have gone down) Sends out the LSP so that all routers receive it When it receives an LSP, records the LSP in its database so that it can store the most recent LSP received Using the accumulated LSP data to construct a complete network topology, proceeds from the common starting point for the SPF algorithm and compute routes to every network Each time an LSP causes a change to the link-state database, the link-state algorithm recalculates the best paths and updates the routing table.

47 Link State Overview (continued)
Adjacency database Link-state database Forwarding database RTR - C Network /24 1/1/1 RTR - A 1/1/2 RTR - B Adjacency database RTR-B – on 1/1/2 RTR-C – on 1/1/1 Link-state protocols keep three databases in the routers: The adjacency database, sometimes called the neighbor database, keeps track of all other routers that are directly attached and passing link-state routing information. The adjacency database is maintained with periodic hello messages. The LSDB has all learned paths to all the destination networks. This database is used to create the SPF tree that ultimately creates the routing table. The routing table, sometimes called the forwarding database, is used by the router to accurately forward IP packets to the destination network. /24 via 1/1/2 cost 20 via 1/1/1 cost 40 LSDB Routing table /24 via 1/1/2

48 Link State Overview (continued)
/30 /8 .1 .2 .2 Step 1 – Updates received from peers .1 /30 Routing table /8 via Step 2 – Topology database created /8 Via Cost 10 Via Cost 20 Step 3 – SPF algorithm determines the best path to destination networks A link-state protocol, also known as SPF, maintains a complex database of topology information. A distance-vector protocol has nonspecific information about distant networks and no knowledge of distant routers, whereas a link-state protocol maintains full knowledge of distant routers and how they interconnect. Step 4 – Routing table created /8 Via Cost 10 – BEST Via Cost 20

49 Exchanging Link-State Information
B C D R1 Link-state packet A 10 B R2 Link-state packet B 10 C R3 Link-state packet C 10 D Network discovery for link-state routing uses the following processes: Routers exchange LSPs with each other. Each router begins with directly connected networks for which it has direct link-state information. The router floods its link-state information to other routers in the network. Routers exchange LSPs with each other. Each begins with directly connected networks for which it has direct link-state information.

50 Building a Topological Database
R1 R2 R3 A B C D R1 Link-state packet A 10 B R2 Link-state packet C R3 Link-state packet D R1 Link-state packet A 10 B R1 Link-state packet A 10 B R2 Link-state packet B 10 C R2 Link-state packet B 10 C Network discovery for link-state routing uses the following processes (continued): Each router constructs a topological database that consists of the link-state information from the other routers in the network. R3 Link-state packet C 10 D R3 Link-state packet C 10 D

51 Calculating the SPF Tree and Populating the Routing Table
R1 Link-state packet A 10 B R2 Link-state packet C R3 Link-state packet D SPF 1 2 Network discovery for link-state routing uses the following processes (continued): 1.The SPF algorithm computes network reachability by determining the shortest path to the other networks in the link-state network. 2.The router constructs this logical topology of shortest paths as a tree with itself as root. 3.The router lists its best paths and the ports for these destination networks in the routing table. It also maintains additional topology elements and status information. When these processes are complete, normal routing of packets can begin. SPF tree 3 R1 Routing table

52 SPF Algorithm R1 LSDB R3 R1 R2 R1, R2, 5 R1, R3, 10 R2, R1, 5
R3, net1, 0 10 /8 (net1) 100 R1 5 R2 In an OSPF network such as the one above, all routers exchange link-state information and collect the information about every router’s links in the link-state database. From this database, the SPF calculation is made to populate the routing table.

53 SPF Algorithm (continued)
Step Candidate Cost to root SPF tree 1 R1, R1, 0 2 R1, R2, 5 R1, R3, 10 5 10 3 4 R2, R3, 100 105 R3, net1, 0 6 R3 10 /8 (net1) 100 R1 5 R2 From the link-state database shown on the previous slide, the SPF calculation is used to construct the SPF tree, which contains the shortest path to each destination in the network. The creation of the routing table from the SPF tree is straightforward. The steps for calculating the SPF tree on R1 are as follows: R1 puts itself as the root of the SPF tree. Neighbors of R1 that are not already in the SPF tree are added to the candidate list. The candidate neighbor {R1, R2, 5} with the lowest cost to the root (R2) is added to the SPF tree. Neighbors of R2 that are not in the SPF tree are added to the candidate list. {R2, R3, 100} is removed from the candidate list because the cost to this destination is higher than the cost for another candidate in the list. The candidate neighbor {R1, R3, 10} with the lowest cost is moved to the SPF tree. All this candidate’s neighbors are moved to the list. The lowest-cost candidate is moved to the SPF tree. Because there are no more candidates in the list, the algorithm is terminated. The SPF tree is complete.

54 Link State – Topology Change
Link-state updates are driven by topology changes. Run SPF Update routing table Topology change Run SPF Update routing table When a router recognizes a topology change (link down, neighbor down, new link, or new neighbor), it must notify its neighbors. To do this, each link-state router does the following: The router that recognizes the change sends out new link-state information that reflects the change. When a router receives new link-state information, it populates the information in its topological database and passes it on to its neighbors. The SPF algorithm is run against the new topological database to update the routing table with the new information. Every time there is a topology change that causes an update to the topological database, the SPF algorithm must be run. Run SPF Update routing table Link-state information

55 Sequence numbers must be included in the link-state information.
Without sequence numbers, the link-state information could be flooded indefinitely. The sequence number remains the same, router-to-router, during the flooding process. In a link-state environment, routers use the sequence numbers for the following decisions when they receive link-state updates: If the sequence number is lower than the one in the database, the link-state information is discarded. If the sequence number is the same as the one in the database, an ACK is sent. The link-state information is then discarded. If the sequence number is higher, the link-state information is populated in the topological database, an ACK is sent, and the link-state information is forwarded to its neighbors.

56 Sequence Numbers (continued)
A B C D R1 Link-state packet R1 Link-state packet R1 Link-state packet Seq=2 Seq=1 Seq=1 R1 R2 R3 A B C D R1 generates new link-state information for network A. It increments the sequence number and sends the link-state information to its neighbor. On receiving the link-state information, R2 checks the sequence number and sees that it is newer. R2 populates is topological database and floods the link-state information to its neighbor. R1 Link-state packet R1 Link-state packet R1 Link-state packet Seq=2 Seq=2 Seq=1

57 Sequence Numbers (continued)
R1 receives 2 copies of the link-state information for network Z. R1 must decide what to do with the second copy of the link-state information it receives. R1 R2 B C R3 Cost 10 Cost 10 R1 receives link-state information via R2 first. It populates its topological database with the newly received link-state information. Link-state information is then received from R6. R1 must compare the link-state information with the information it already has in its database. R1 can see that the sequence numbers are the same. It therefore discards the link-state information and does not forward it to R2. This process stops the link-state information from being flooded indefinitely. In the same example, if network Z comes up immediately after it goes down, the sequence number is incremented again. For some reason, the link-state information about network Z going down, with a sequence number of 2, is delayed via R4-R3-R2-R1. The link-state information for network Z being available, with a sequence number of 3, arrives at R1 via R4-R5-R6-R1 first. When the delayed link-state information with a sequence number of 2 arrives, R1 compares it with the link-state information it has in its topological database. R1 determines that the link-state information is older and therefore discards it. A Cost 10 Cost 10 D F E Z Cost 20 Cost 20 R6 R5 R4

58 Link-State Information Aging
Link-state information includes an age field. The age of newly created link-state information is set to 0 for OSPF and 1200 for IS-IS. It is incremented by every hop during the flooding procedure for OSPF and is decremented for IS-IS. The link-state age is also incremented for OSPF and decremented for IS-IS as it is held in the topological database. Maximum age When the link-state information reaches its maximum age, it is no longer used for routing. The link-state information is flooded to the neighbors with the maximum age, and the link-state information is removed from the topological database.

59 IS-IS – Packet Processing
A router deals with topology changes as follows: Sequence No. same? Is entry in LSDB? Yes Yes LSU/LSA Ignore No No Add to LSDB Is sequence number higher than one in LSDB? Send ACK Yes No Flood LSA Send LSU back with newer information Run SPF End

60 Hierarchy in Link-State Networks
Scalability issues exist for link-state networks: The size of the link-state database increases exponentially with the size of the network. The complexity of the SPF calculation also increases exponentially. A topology change requires complete recalculation of the forwarding table on every router. Hierarchy allows a large routing domain to be split into several smaller routing domains. IS-IS and OSPF both implement hierarchy but use different techniques. Hierarchy results in suboptimal routing. Hierarchy is less common than in the past due to the increased capacity of routers. Scalability issues exist for link-state networks: The size of the link-state database increases exponentially with the size of the network. Every router must add and keep track of any new destination reachable in the network. A large database increases the consumption of router resources. The complexity of the SPF calculation also increases exponentially. A topology change requires complete recalculation of the forwarding table on every router. The increased overhead in calculating new routing information can overwhelm a router if it has insufficient resources. Hierarchy allows a large routing domain to be split into several smaller routing domains. Routing happens within the smaller routing domains and between the domains, simplifying the SPF calculation. IS-IS and OSPF both implement hierarchy but use different techniques. They both define areas and route within and between areas. Hierarchy results in suboptimal routing. The best path to leave the area may not be the best route to the final destination. Hierarchy is less common than in the past due to the increased capacity of routers. Many large networks are now configured as a single area, which simplifies configuration and optimizes routing. Modern routers have the ability to handle hundreds of nodes.

61 IS-IS – Hierarchical View
Integrated IS-IS Network L1 L2 Area 2 L1/L2 L1 L1/L2 Area 3 The IS-IS hierarchy is organized as follows: ISO standard 10589, subsequently RFC 1142 Link-state Highly scalable (1000 routers per area) Areas are connected by level 2 routers in a mesh. The network between level 2 routers must be highly available. L1/L2 Backbone (Level 2) links Level 1 links L Level 1 L Level 2 L1/L2 Level 1/Level 2 L1 Area 1

62 OSPF – Hierarchical View (continued)
OSPF Hierarchical Routing Area The OSPF environment is organized using two primary elements: Area – An area is a grouping of OSPF routers. OSPF areas are logical subdivisions of the OSPF autonomous systems. The topology of each area is invisible to entities in other areas, and each area maintains its own topological database. Autonomous System – A group of networks and network equipment under common administration For each autonomous system, the central area is defined as area 0 and is called the backbone area. All other areas must connect to area 0. The backbone area distributes routing information between areas. All other areas communicate through the backbone area. Area Area

63 Alcatel-Lucent Routing Protocols
Module 5 — Open Shortest Path First

64 OSPF — RFC History OSPF workgroup formed OSPF v1 RFC 1131 defined 1987
1998 1997 1994 1991 1989 Present 1999 OSPF v2 RFC 1247 defined OSPF v2 Updated RFC 1583 OSPF v2 Updated RFC 2178 OSPF v2 Updated RFC 2328 Over the course of OSPF’s existence, multiple RFCs have been created and commonly accepted. The slide above lists the RFCs that explicitly define the characteristics of basic OSPF. In 1987, a work group headed by John Moy started outlining and creating the first RFC for OSPF. RFC 1131 is the first RFC for OSPF. It defined the basic operation and common characteristics of OSPF. This was the OSPFv1 specification that was first released for public use. In 1991, RFC 1247 was released. This RFC incorporated additional features and modifications to the original RFC. During the next seven years, OSPF was updated to support a variety of features that the industry wanted. This included multicast OSPF, support for CIDR, packet authentication for updates, and point-to-multipoint network support. The currently accepted RFC is 2328, and it supersedes all other RFCs mentioned on this page. In December 1999, RFC 2740 was released. This was an update to the OSPF standard to support IPv6. Some of the other RFCs related to OSPF are listed below: RFC 1131: Original version 1 spec, Oct. 1989 RFC 1247: Original version 2 spec, July 1991 RFC 1584: Multicast extensions, March 1994 RFC 1586: OSPF over frame relay, March 1994 RFC 1587: NSSA extension, March 1994 RFC 1745: OSPF/BGP interaction, Dec. 1994 RFC 1793: Demand circuits, April 1995 RFC 1850: OSPF MIB definition, Nov. 1995 RFC 2328: Latest version 2 update, April 1998 OSPF work in progress OSPF for IPv6 RFC 2740

65 OSPF — Protocol Overview
Link state-driven updates, periodic hellos Classless routing protocol Subnet mask sent in update Support for VLSM, CIDR, and manual route summarization Support for authentication Link-state protocols have the following common attributes: Link-state protocols trigger an update when a link (interface) changes state. The router connected to the link initiates a triggered update to its neighbors, notifying them of the change in the topology. If the network is stable and no changes in links are detected, the routers send periodic hello messages to maintain connectivity without consuming excessive bandwidth. A common attribute shared by link-state protocols is that they are classless and support all the common attributes of a classless routing protocol. Because link-state protocols are classless, updates contain the subnet mask of each network being advertised. This allows for more optimal network design and accurate path selection. VLSM and CIDR are supported in all link-state protocols. Due to the classless aspects of link-state protocols, manual summarization is actively supported. This allows network administrators to have much more control of where and how the summarization takes place. All modern link-state protocols support authentication of the updates that are sent between routers. This ensures that accurate network topologies are created without false information or errors and protects the network against DoS attacks. Link-state protocols maintain three common databases: topology (link-state DB), neighbor (adjacency DB), and routing table (forwarding DB). Modern link-state protocols use a multicast address to convey updates and hellos to their neighbor link-state routing peers. This reduces processing on devices in the network that are not running the link-state protocol. Maintenance of multiple databases Multicast addressing – and

66 OSPF — Key Features Key OSPF features are: Backbone areas Stub areas
NSSAs Virtual links Authentication Support for VLSM and CIDR Route redistribution Routing interface parameters OSPF-TE extensions OSPF is a hierarchical link-state protocol. OSPF is an IGP used in large autonomous systems (ASs). OSPF routers exchange state, cost, and other relevant interface information with neighbors. The information exchange enables all participating routers to establish a network topology map. Each router applies the Dijkstra (SPF) algorithm to calculate the shortest path to each destination in the network. The resulting OSPF forwarding table is submitted to the RTM to calculate the routing table. When a router is started with OSPF configured, OSPF, along with the routing-protocol data structures, is initialized and waits for indications from lower-layer protocols that its interfaces are functional. Alcatel’s implementation of OSPF conforms to the OSPFv2 specifications presented in RFC 2328, OSPF Version 2. Routers that run OSPF can be enabled with minimal configuration. All default and command parameters can be modified.

67 OSPF — Protocol Comparison
Feature RIPv2 IS-IS OSPF Updates Periodic Incremental Incremental Update type Broadcast/Multicast L2 Multicast L3 Multicast Transport UDP Layer 2 IP Authentication Simple and MD5 Simple and MD5 Simple and MD5 Metric Hops Cost Cost Metric type Distance vector Link-state Link-state The table above shows the differences and similarities of RIPv2, OSPF, and IS-IS. All three protocols are supported on the 7750 SR. RIPv1 is not listed as it is not the default implementation of RIP when configured on the 7750 SR. OSPF and IS-IS are very similar in ability and operation. The major difference is in how they are configured and optimized. Both support the IP protocol; however, OSPF was designed around IP and IS-IS was adapted to support IP. OSPF updates are encapsulated in an IP header. IS-IS does not use IP to exchange its updates, it uses the data-link layer directly. OSPF does not use an upper-layer protocol, such as TCP or UDP. Instead, OSPF has a reserved IP-header protocol number 89 that identifies the received traffic as destined for the OSPF process in the router. This is unlike RIP, which uses UDP and port 520. VLSM / CIDR support Yes Yes Yes Topology size Small/Medium Large Large Convergence Slow Fast Fast

68 OSPF — Link-State Protocol Comparison
Feature IS-IS OSPF Updates Incremental Incremental Multicast layer Layer 2 Layer 3 Authentication Simple and MD5 Simple and MD5 Metric Default: all ports cost 10 Auto-calculation on interface Metric type Link-state Link-state LSA types L1 and L2 Multiple types The table above diagram shows some common features of link-state protocol and how IS-IS and OSPF have implemented them. Updates — Both IS-IS and OSPF use incremental updates. This means that if a link changes state, only that change is conveyed to the respective neighbors. The entire database is not flooded. In addition, periodic updates to ensure LSDB consistency are sent (IS-IS every 20 minutes, OSPF every 30 minutes). Multicast layer — IS-IS and OSPF use multicast updates to communicate with peers. IS-IS uses layer 2 multicast addresses and OSPF uses layer 3 multicast addresses. Authentication — Secure communication is supported by both IS-IS and OSPF. By default, neither runs any authentication; but, both support simple and MD5 authentication methods. Metric — IS-IS uses a default cost of 10 for all segments, regardless of the actual speed of the link. OSPF defaults to an automatic calculation based on a set metric value divided by the bandwidth of the link. Metric type — Both IS-IS and OSPF use Dijkstra’s SPF link-state algorithm for best-path calculation. LSA types — IS-IS has two major types of link-state advertisements (L1: intra-area, L2: inter-area), and OSPF has multiple types of LSAs. Area hierarchy — OSFP requires a backbone area, and all other areas must directly attach to the backbone. IS-IS does not have such a requirement. All areas can directly connect to all other areas. Area boundaries — IS-IS area boundaries are connected on segments. OSPF areas are defined on router interfaces. Convergence — because both IS-IS and OSPF use the same Dijkstra algorithm, convergence time is identical. Area hierarchy Not required Backbone area Area boundaries On segment At interface Convergence Fast Fast

69 OSPF — Path Determination
OSPF uses SPF for path determination. SPF uses cost values to determine the best path to a destination. RTR-C Cost 125 Cost 125 Cost 0 Cost 10 RTR-A RTR-B Cost 125 Metrics In OSPF, all interfaces have a cost value or routing metric used in the OSPF link-state calculation. A metric value is configured based on bandwidth to compare different paths through an AS. OSPF uses cost values to determine the best path to a particular destination: the lower the cost value, the more likely the interface will be used to forward data traffic. To calculate the cost of a link a reference bandwidth is set. The reference bandwidth is referenced in kilobits per second and provides a reference for the default costing of interfaces based on their underlying link speed. The default interface cost is calculated as follows: The default reference-bandwidth is kb/s or 100 Gb/s, so the default auto-cost metrics for various link speeds are as as follows: 10-Mb/s link default cost of 100-Mb/s link default cost of 1000 1-Gb/s link default cost of 100 10-Gb/s link default cost of 10 The reference-bandwidth command assigns a default cost to the interface based on the interface speed. To override this default cost on a particular interface, use the metric metric command in the config>router>ospf>area>interface ip-int-name context. RTR-A – Cost 260 via RTR C * – Cost 135 via RTR B * = Best path

70 Cost = reference-bandwidth ÷ bandwidth
Calculating Link Cost Cost = reference-bandwidth ÷ bandwidth The default reference-bandwidth is kb/s or 100 Gb/s. The default auto-cost metrics for various link speeds are as follows: 10-Mb/s link default cost of 100-Mb/s link default cost of 1000 1-Gb/s link default cost of 100 10-Gb/s link default cost of 10 The cost is configurable.

71 Interfaces must be configured in an OSPF area.
Configuration Basics Interfaces must be configured in an OSPF area. By default, interfaces in an area are advertised by OSPF. Routes received through OSPF are advertised by OSPF. No other routes are advertised by default. Verify that adjacencies are formed with neighbors. Verify that routes are in the routing table.

72 OSPF — Multicast Addressing
OSPF uses class D multicast addresses in the range to Specially reserved addresses for OSPF: : All routers that speak OSPF on the segment : All DR/BDRs on the segment IP multicast addresses use the lower 23 bits of the IP address as the low-order bits of the MAC multicast address E-XX-XX-XX. = MAC E = MAC E OSPF uses IP multicast addressing to communicate with routing peers. This reduces the overhead of other devices on the same segment that are not running OSPF. OSPF has two reserved multicast IP addresses. The first is and is used to communicate with all OSPF speakers. The second is and is used in multi-access broadcast topologies in which a DR/BDR is required for proper OSPF operations. When an OSPF update is sent on an Ethernet topology, the destination MAC address is modified to use the reserved multicast range. The range has the first 24 bits of the MAC address, normally reserved for the manufacturer code, set to E. The remaining 24 bits of the MAC address are the lower 24 bits of the IP multicast address. With OSPF, the relationship between the IP multicast address and the MAC address is as follows: and E : Any OSPF speaker and E : Any DR/BDR

73 OSPF packets use protocol number 89 in the IP header.
OSPF — Generic Packet OSPF packets use protocol number 89 in the IP header. OSPF is its own transport layer. Link header IP header OSPF packet types Link trailer OSPF updates are sent using the IP header at the network layer. However, unlike RIP, OSPF does not use a transport-layer protocol. Instead, all OSPF updates are sent directly from the IP layer to the OSPF process. To accomplish this, reserved protocol number 89 in the IP header is allocated to identify OSPF traffic. IP header protocol ID 89 = OSPF

74 OSPF database descriptor OSPF link-state request
OPSF — Packet Types OSPF hello OSPF database descriptor OSPF link-state request OSPF link-state update OSPF link-state ACK OSPF uses 5 different types of packets to establish and maintain router connectivity and network convergence. Hello packet — This packet is used to establish adjacencies with other routers that speak OSPF. It is also used to maintain neighbor connectivity by being propagated periodically, typically every 10 seconds. However, this value can be modified from 0 to seconds. Database description — This packet conveys a summary of all networks in the router’s database. Typically this is the classless network, the router’s cost to access, and the sequence number associated with the network entry. Link-state request — When a neighbor router receives a database description packet, it compares the entry in its current link-state database with the information received. If a received network is not in the database or if the sequence number for a network is higher, the router generates a link-state request for more information about the network. Link-state update — When it receives a link-state request, the router responds with the complete link-state database entry. To accomplish this, the router generates a type 4 (link-state update) packet and forwards it back to the requesting router. Link-state ACK — Each newly received LSA must be acknowledged. This is usually done by sending link-state ACK packets. Many ACKs may be grouped together in a single link-state ACK packet.

75 OSPF — Link Topology Types
Multi-access Point-to-point OSPF supports several types of network topology: Multi-access (broadcast) — This topology is commonly an Ethernet segment in which multiple routers are connected and are actively exchanging OSPF updates. A multi-access network, a DR and BDR must be elected. This type of network will be described in upcoming sections. Point-to-point — RFC 2328 defines a point-to-point connection as a connection in which the neighbor router is fully adjacent. This typically means that there are only two devices on the link. Point-to-multipoint — Also defined in RFC 2328, this configuration is typically used in an NBMA environment such as frame-relay or ATM, in which a full mesh is not physically configured. Non-broadcast multi-access — A typical implementation is in an NBMA environment that has full-mesh connectivity at the physical layer.

76 OSPF — Router ID Each router must have a router ID, the ID by which the router is known to OSPF. The default RID is the last 32 bits of the chassis MAC address. Configuring a system interface overrides the default. Using a system interface is easier to document. The RID uniquely identifies the router in the routing domain. Unless defined otherwise, it is set to the system-interface address. A RID can be derived by: Defining the value in the config>router router-id context Defining the system interface in the config>router>interface ip-int-name context (used if the router ID is not specified in the config>router router-id context) Inheriting the last four bytes of the MAC address When a new RID is configured, protocols are not automatically restarted with the new RID. The next time a protocol is initialized or reinitialized, the new RID is used. A delay can occur when different protocols use different RIDs. To force the new RID, issue the shutdown and no shutdown commands for each protocol that uses the RID, or restart the entire router. By default, the system uses the system-interface address, which is also the loopback address. If a system-interface address is not configured, the last 32 bits of the chassis MAC address are used.

77 OSPF — Point-to-Point Segments
On point-to-point links, there is no need for a DR or BDR. All packets are sent via IP multicast address Usually a leased-line (i.e., HDLC, PPP) segment Can be configured on point-to-point Ethernets RTR - C Network /24 When the connection between two OSPF routers is a point-to-point link, there is no need for a DR or BDR. All packets are sent using the IP multicast address. This implementation is typically used on serial interfaces; however, it can also be configured on point-to-point Ethernet segments, in which only two routers are connected. RTR - A RTR - B

78 OSPF — LAN Communication
Election of the DR and BDR in multi-access networks: A B RTR-A Has the highest RID, so it will be the DR RTR-B Has the second highest RID, so it will be the BDR C D E A router uses the OSPF hello protocol to discover neighbors. A neighbor is a router that is configured with an interface to a common network. The router sends hello packets to a multicast address and receives hello packets in return. In broadcast networks, a DR and a BDR are elected. The DR is responsible for sending LSAs that describe the network, which reduces the amount of network traffic. The routers attempt to form adjacencies. An adjacency is a relationship that is formed between a router and the DR or BDR. For point-to-point networks, no DR or BDR is elected. An adjacency must be formed with the neighbor. To significantly improve adjacency formation and network convergence, a network should be configured as point-to-point if only two routers are connected, even if the network is a broadcast media such as Ethernet. In the example above, RTR-A is the DR and RTR-B is the BDR. Routers C, D, and E will only form adjacencies with RTR-A and RTR-B, not with each other. Not every neighboring router forms an adjacency. Routing protocol updates are only sent to and received from adjacencies. Routers that do not become fully adjacent remain in the 2-way neighbor state. Each router sends hellos. The router with the highest priority is the DR. If all priorities are the same, the DR is the router with the highest RID.

79 OSPF — Exchanging Updates in a LAN
Election of the DR and BDR in multi-access networks: RTR-A (DR) RTR-B (BDR) RTR-C sends update to All DRs using IP address RTR-A sends update to All OSPF routers using IP address RTR-C D E In the example above, RTR-C has a topology change that needs to be conveyed. The following steps occur: RTR-C sends its update to the DRs using IP multicast address Both DRs receive the update. The BDR monitors to see if the DR sends out updates to all other routers, including the BDR. The DR takes the update from RTR-C and floods the change to all other routers on the segment, using IP multicast address Note: DR and BDR election is not required in point-to-point networks. Routers use the IP address to send updates to the DRs. The BDR monitors the DR to ensure that it sends updates. The DR uses to send updates to all OSPF routers.

80 Alcatel-Lucent Routing Protocols
Module 6 — Intermediate System–to–Intermediate System

81 IS-IS — Protocol Overview
Development began prior to that of OSPF. The U.S. government required ISPs to use IS-IS for early stages of the Internet. IS-IS supports IPv6. Many large enterprise networks and ISPs use IS-IS due to the scalability and stability of the protocol. Development of IS-IS was initiated prior to OSPF, and IS-IS was one of the first routing protocols to scale to the size required to support ISPs. Use of IS-IS became a requirement for ISPs in the early stages of Internet development. This requirement was subsequently dropped as OSPF became a more commonly deployed IGP. However, multiple ISPs still use IS-IS as their backbone routing protocol to support their BGP implementations. IS-IS is an incredibly stable protocol that also provides very rapid convergence. These features, in combination with its scalability, are prime reasons that some large ISPs use IS-IS as the backbone routing protocol in their networks. The only real downfall of IS-IS is the lack of engineering support. Although many engineers understand OSPF, very few truly understand IS-IS at a level sufficient to implement, optimize, and troubleshoot quickly and accurately. Because IS-IS was developed independently from IPv4, modification to IPv6 is easy and fairly seamless. This cannot be said for OSPF or other common routing protocols.

82 IS-IS — RFC History RFC 1142 Original RFC 1990 RFC 1195 TCP/IP support
ISO 10589 released 1992 1994 RFC 1629 NSAP and Internet ….. Other IS-IS RFCs released Over the course of IS-IS’s existence, multiple RFCs have been created and commonly accepted. The slide above lists the RFCs that explicitly define the characteristics of basic IS-IS. In February 1990, RFC 1142 was released. It was based upon the ISO’s soon-to-be-released standard The IS-IS standard is closely modeled after DecNet v5 and ISO standard 8473. In December 1990, RFC 1195 was released. This RFC outlined how to support TCP/IP and OSI systems simultaneously using IS-IS. It is commonly known as Integrated IS-IS. In 1992, ISO standard 10589:1992 was released. This is the commonly accepted standard for IS-IS, and there are a few minor updates that the ISO has released. However, this document is the key document in understanding IS-IS. In May 1994, RFC 1629 was released. This document outlined the NSAP addressing to be used in the Internet environment. Throughout the remainder of the 1990s and into the 2000s, several minor RFCs that dealt with IS-IS were released. Only the major RFCs are highlighted in the slide above. The most recent major RFC release occurred in August 2002: RFC This RFC provides a standardized definition of TLV code points and their interpretation in IS-IS. 2002 Present RFC 33509 TLV code points IS-IS work in progress

83 IS-IS — Protocol Overview (continued)
Link-state driven updates, periodic hellos Classless routing protocol Subnet mask sent in update Support for VLSM, CIDR, and manual route summarization Support for authentication The IS-IS and OSPF link-state protocols have the following common attributes: Link-state protocols trigger an update when a link (interface) changes state. The router connected to the link initiates a triggered update to its neighbors, notifying them of the changed topology. If the network is stable and no changes in links are detected, the routers send periodic hello messages to maintain connectivity without consuming excessive bandwidth. A common attribute shared by the OSPF and IS-IS link-state protocols is that they are classless and support all the common attributes of a classless routing protocol. Because IS-IS is classless, the updates contain the subnet mask of each network being advertised. This allows for more optimal network design and accurate path selection. VLSM and CIDR are supported in both OSPF and IS-IS. The classless nature of these protocols enables the manual summarization of networks. This allows network administrators to have much more control of where and how the summarization takes place. Both protocols support authentication of the updates that are sent between routers. This ensures that accurate network topologies are created without false information or errors. Link-state protocols maintain three common databases: topology (link-state DB), neighbor (adjacency DB), and routing table (forwarding DB). Modern link-state protocols use a multicast address to convey updates and hellos to their neighbor link-state routing peers. This reduces processing on devices in the network that are not running the link-state protocol. Layer 2 multicast addressing for IS-IS is as follows: L1 updates use C L2 updates use C Maintenance of multiple databases Layer 2 multicast addressing

84 Key IS-IS features are:
IS-IS — Key Features Key IS-IS features are: Area hierarchy Authentication Support for VLSM and CIDR Route redistribution Routing interface parameters IS-IS TE extensions IS-IS is a link-state IGP that uses the SPF algorithm to determine routes. Routing decisions are made using the link-state information. IS-IS evaluates topology changes and, if necessary, performs SPF recalculations. Entities in IS-IS include networks, intermediate systems, and end systems. In IS-IS, a network is an AS (routing domain), with end systems and intermediate systems. Intermediate systems send, receive, and forward PDUs. A router, such as the 7750 SR, is an intermediate system. End-system and intermediate-system protocols allow routers and nodes to identify each other. IS-IS periodically sends out link-state updates throughout the network so that each router can maintain current network topology information. IS-IS supports large ASs by using a two-level hierarchy. A large AS can be administratively divided into smaller, more manageable areas. A system only belongs to one area. Level 1 routing is performed within a single area. Level 2 routing is performed between areas. The 7750 SR can be configured as level 1, level 2, or level 1/2.

85 IS-IS — Protocol Comparison
Feature RIPv2 OSPF IS-IS Updates Periodic Incremental Incremental Update type Broadcast/Multicast L3 Multicast L2 Multicast Authentication Simple and MD5 Simple and MD5 Simple and MD5 Metric Hops Cost Cost Metric type Distance vector Link-state Link-state VLSM / CIDR support Yes Yes Yes The table above shows the differences and similarities of RIPv2, OSPF, and IS-IS. All three protocols are supported on the 7750 SR. RIPv1 is not listed as it is not the default version and is seldom configured on the 7750 SR. IS-IS and OSPF are very similar in ability and operation. Both support the IP protocol; however, OSPF was designed around IP and IS-IS was adapted to support IP. OSPF messages are encapsulated in an IP header whereas IS-IS encapsulates its messages directly in the data link layer. Neither OSPF nor IS-IS uses an upper-layer protocol, such as TCP or UDP. This is unlike RIP, which uses UDP port 520. Topology size Small Very large Very large Summarization Manual Manual Manual Convergence Slow Fast Fast

86 IS-IS — Link-State Protocol Comparison
Feature IS-IS OSPF Updates Incremental Incremental Multicast layer Layer 2 Layer 3 Authentication Simple and MD5 Simple and MD5 Metric Default: all ports cost 10 Auto-calculation on interface Metric type Link-state Link-state Update types L1 and L2 Multiple types The table above shows some common characteristics of a link-state protocol and how IS-IS and OSPF implement these features. Updates — Both IS-IS and OSPF use incremental updates. This means that if a link changes state, only that change is conveyed to the respective neighbors. The entire database is not flooded. In addition, periodic updates to ensure LSDB consistency are sent (IS-IS every 20 minutes, OSPF every 30 minutes). Multicast layer — IS-IS and OSPF use multicast updates to communicate with peers. IS-IS uses layer 2 multicast addresses and OSPF uses layer 3 multicast addresses. Authentication — Secure communication is supported by both IS-IS and OSPF. By default, neither runs any authentication, but both support simple and MD5 authentication methods. Metric — IS-IS uses a default cost of 10 for all segments, regardless of the actual capacity of the link. OSPF defaults to an automatic calculation based on a set metric value divided by the bandwidth of the link. Metric type — Both IS-IS and OSPF use Dijkstra’s SPF link-state algorithm for best-path calculation. Update types — IS-IS has two major types of link-state updates (L1: intra-area, L2: inter-area). OSPF has multiple types of link-state updates. Area hierarchy — OSFP requires a backbone area, and all other areas must directly attach to the backbone. In IS-IS, all level 2 routers must be connected in a continuous link, and areas can directly connect to any other area. Area boundaries — IS-IS area boundaries are on segments. OSPF areas are defined on router interfaces. Convergence — Because both IS-IS and OSPF use the same Dijkstra algorithm, convergence times are identical. Area hierarchy Not required Backbone area Area boundaries On segment At interface Convergence Fast Fast

87 IS-IS — Frequently Used Terms
Area — Corresponds to the level 1 subdomain End system — Typically a computer, printer, or other attached device Intermediate system — Router in an IS-IS network Neighbor — A physically adjacent router Adjacency — A separate adjacency is created for each neighbor on a circuit and for each level of routing (level 1 and level 2) on a broadcast circuit. Circuit — A single locally attached network Link — The communication path between 2 neighbors CSNP — Complete sequence number PDU PSNP — Partial sequence number PDU PDU — Protocol data unit Area — An area is a routing subdomain that maintains detailed routing information about its own internal topology and also maintains routing information that allows it to reach other routing subdomains. Areas correspond to the level 1 subdomain. End system — End systems send NPDUs to other systems and receive NPDUs from other systems, but do not relay NPDUs. This international standard does not specify additional end-system functions beyond those supplied by ISO 8473 and ISO OSI end systems do not exist in an IS-IS network that is used only for routing TCP/IP. Intermediate system — An intermediate system sends NPDUs to other systems and receives NPDUs from other systems, as well as relaying NPDUs. Intermediate systems are the routers of an IS-IS network. Neighbor — An adjacent system that is reachable by the traversing of a single subnetwork by a PDU Adjacency – An adjacency is a portion of the local routing information that pertains to the reachability of a single neighboring end system or intermediate system over a single circuit. Adjacencies are used as input to the decision process to form paths through the routing domain. A separate adjacency is created for each neighbor on a circuit and for each level of routing (level 1 and level 2) on a broadcast circuit. Circuit — The subset of the local routing information base that is pertinent to a single local subnetwork point of attachment Link — The communication path between two neighbors. A link is up when communication is possible between the two SNPAs. CNSP — A sequence number that provides an index to the contents of a router’s link-state database. Used to determine which routing updates need to be exchanged to bring the database up-to-date. PSNP — A sequence number used between routers to request a specific routing update or to acknowledge the receipt of an update PDU — OSI term for the data exchanged at a specific level of the OSI network model. For example, layer 3 PDUs are known as packets in an IP network.

88 IS-IS — Frequently Used Terms (continued)
Designated IS — The intermediate system in a LAN that is designated to generate updates on behalf of the nodes in the LAN Pseudo node — When a broadcast subnetwork has n connected intermediate systems, the broadcast subnetwork itself is considered to be a pseudo node. Broadcast subnetwork — A multi-access subnetwork (such as Ethernet) that supports the capability of addressing a group of attached systems with a single PDU General topology subnetwork — A topology that is modeled as a set of point-to-point links, each of which connects 2 systems Routing subdomain — A set of intermediate systems and end systems that are located within the same routing domain Level 2 subdomain — The set of all level 2 intermediate systems in a routing domain Designated IS — The intermediate system in a LAN that is designated to perform additional duties. In particular, the designated IS generates link-state PDUs on behalf of the LAN, treating the LAN as a pseudo node. Pseudo node — When a broadcast subnetwork has n connected ISs, the broadcast subnetwork itself is considered to be a pseudo node. The pseudo node has links to each of the n ISs, and each of the ISs has a single link to the pseudo node (rather than n-1 links to each of the other ISs). Link-state PDUs are generated on behalf of the pseudo node by the designated IS. Broadcast subnetwork — A multi-access subnetwork (such as Ethernet) that supports the capability of addressing a group of attached systems with a single PDU General topology subnetwork — A topology that is modeled as a set of point-to-point links, each of which connects two systems. There are several generic types of general topology subnetworks: multipoint links, permanent point-to-point links, and dynamic and static point-to-point links. Routing subdomain — A set of ISs and end systems located within the same routing domain Level 2 subdomain — The set of all level 2 in a routing domain

89 IS-IS — Protocol Overview
IS-IS uses SPF for path determination. SPF uses cost values to determine the best path to a destination. RTR-C Cost: 10 Cost: 10 Cost: 10 Cost: 10 RTR-A RTR-B Cost: 10 Metrics To calculate the lowest cost to reach a given destination, each configured level on each interface must have a cost. The costs for each level on an interface may be different. In IS-IS, if the metric is not configured, default cost 10 is used. IS-IS does not use a reference-bandwidth as in OSPF. Each link has a set cost value of 10 unless it is manually changed to another value in the range 1 to RTR-A : cost 30 via RTR-C * : cost 20 via RTR-B * = Best path Packet flow

90 IS-IS — ISO Network Addressing
IS-IS uses unique addressing (OSI NSAP addresses) compared to that of other IP routing protocols. Each address identifies the area, system, and sector. Routers with common area addresses form L1 adjacencies. Routers with different area addresses form L2 adjacencies, if capable. 2-layer hierarchy: Level 1: Builds the local area topology and forwards traffic to other areas through the nearest L1/L2 router Level 2: Exchanges prefix information and forwards traffic between areas IS-IS uses ISO network addresses. Each address identifies a point of connection to the network, such as a router interface, and is called a network service access point. An end system can have multiple NSAP addresses, in which case the addresses differ only by the last byte (called the n-selector). Each NSAP represents a service that is available at that node. In addition to having multiple services, a single node can belong to multiple areas. Each network entity has a special network address called a network entity title. Structurally, an NET is identical to an NSAP address but has an n-selector of 00. Most end systems have one NET. Intermediate systems can have up to three area IDs (area addresses). NSAP addresses are divided into three parts. Only the area ID portion is configurable. Area ID — A variable-length field between 1 and 13 bytes. This includes the AFI as the most significant byte, and the area ID. System ID — A 6-byte system ID. This value is not configurable. The system ID is derived from the system or router ID. Selector ID — A 1-byte selector ID that must contain zeros when a NET is configured. This value is not configurable. The selector ID is always 00. Of the total 20 bytes that comprise the NET, only the first 13 bytes, the area ID portion, can be manually configured. As few as one byte can be entered or, at most, 13 bytes. If fewer than 13 bytes are entered, the rest of the area ID is padded with zeros. Routers with common area addresses form level 1 adjacencies. Routers with no common NET addresses form level 2 adjacencies, if they are capable.

91 IS-IS — ISO Network Addressing (continued)
Layer 2 multicast addressing is implemented to support IS-IS. On Ethernet, the following multicast addresses are reserved: L1 updates use C L2 updates use C

92 IS-IS — Link-State Overview
Area L1/L2 L1 L1/L2 Area IS-IS: ISO standard 10589, subsequently RFC 1142 Link-state Highly scalable (1000 routers per area) Areas are connected by level 2 routers in a mesh. The network between level 2 routers must be highly available. The routing protocol engine is almost identical to OSPF, except that area boundaries are on links between routers rather than through a border router (an IS-IS router is always in one area only) All routers in an IS-IS topology are identified as level 1, level 2, or level 1/2. Level 1 routers exchange topology information for the local area. Level 2 routers exchange topology information between the different areas. Level 1/2 routers exchange information between level 1 and level 2 routing domains. Backbone (level 2) link Level 1 link L Level 1 L Level 2 L1/L Level 1/level 2 L1/L2 L1 Area

93 IS-IS — NSAP Addressing
IDP DSP AFI IDI High Order-DSP System ID SEL variable 6 1 Area ID System Address NSEL NSAP — Network service access point IDP — Initial domain part DSP — Domain specific part AFI — Authority and format indicator IDI — Initial domain identifier (e.g., 49 is local assigned, binary) High Order-DSP — High Order Domain Specific Part SEL — N-selector (NSEL) IS-IS uses ISO network addresses. Each address identifies a point of connection to the network, such as a router interface, and is called an NSAP. An end system can have multiple NSAP addresses, in which case the addresses differ only by the last byte (called the n-selector). Each NSAP represents a service that is available at that node. In addition to having multiple services, a single node can belong to multiple areas. Each network entity has a special network address called an NET. Structurally, an NET is identical to an NSAP address but has an n-selector of 00. Most end systems have one NET. Intermediate systems can have up to three area IDs (area addresses). NSAP addresses are divided into three parts as follows. Only the area ID portion is configurable: Area ID — A variable-length field from 1 to 13 bytes. This includes the AFI as the most significant byte, and the area ID. An AFI of 49 specifies that the areas have been assigned by a local authority instead of being assigned by a national or international authority. System ID — A 6-byte system ID. This value is not configurable. The system ID is derived from the system ID or router ID. Typically, this ID is the MAC address of the device or the RID of the device. Selector ID — A 1-byte selector ID that must contain zeros when a NET is configured. This value is not configurable. The selector ID is always 00. This is also referred to as the NSEL in some documentation. This value could change on NBMA and multi-access networks when a DIS is created. Of the total 20 bytes that comprise the NET, only the first 13 bytes (the area ID portion) can be manually configured. As few as 1 byte can be entered, or, at most, 13 bytes. If fewer than 13 bytes are entered, the rest of the area ID is padded with zeros. Routers with common area addresses form level 1 adjacencies. Routers with no common NET addresses form level 2 adjacencies.

94 IS-IS — Protocol Characteristics
Item Value Maximum metric value assignable to a link Maximum metric value for a path All L1 IS multicast address C All L2 IS multicast address C SAP for IS-IS on LANs FE Protocol discriminator for IS-IS 83 NSAP selector for IS-IS 00 Sequence modulus 232 Size of LSP, which all IS routers must be able to handle 1492 Maximum age 1200 Zero life age 60 Maximum number of area addresses in a single area 3 Originally, the maximum metric for a link was limited to 63 (6 bits), with a total path metric of This was considered to be not granular enough for modern networks, especially with traffic engineering, so a new “wide metric” was defined. This uses 24 bits to support a link metric of and a total path metric of The maximum metric value (narrow metric) for a link is limited to a cost of 63. This is because 6 bits are allocated per link for cost (decimal values from 0 to 63). For the maximum metric value for a path (narrow metric), the total number of bits allocated for a path is 10, with a range from 0 to This value is the total cumulative cost to a destination network within the topology. A path that has a greater cost is considered unreachable. The L1 multicast MAC address is a unique address for all L1 packets sent on an Ethernet. The L2 multicast MAC address is a unique address for all L2 packets sent on an Ethernet The LLC DSAP and SSAP ID for IS-IS packets is set to FE. In an frame with LLC, the value is FE:FE. The network-layer header contains protocol ID 83, as assigned by ISO (not used for TCP/IP). The NSAP defaults to “00” when IS-IS is used for IP network updates. The sequence modulus has a maximum value of 232. An LSP must support an MTU of 1492. The maximum age for a network entry in IS-IS without an update is 1200 seconds (20 minutes). The zero life age is the time, in seconds, that a network will remain in the LSDB when the maximum life age has expired. There can be a total of 3 area IDs in a single area. This is useful when changing the topology of your areas, based on network requirements.

95 IS-IS packets use layer 2 encapsulation of the media.
IS-IS — Packet Format IS-IS packets use layer 2 encapsulation of the media. The Ethernet type field is set to 0xFEFE to denote an IS-IS packet instead of an IP packet. The TLV identifies the type of information in the IS-IS packet. IS-IS packets are called PDUs. Ethernet header Type = 0xFEFE IS-IS header IS-IS TLV Link trailer The basic IS-IS update packet is shown above. Note that IS-IS does not use IP at the network layer for exchanging messages. IS-IS uses the data link layer directly for framing and transmission over network segments. The IS-IS header is always the same, yet the TLV varies depending on the type of update being sent.

96 IS-IS — Packet Format Details
Ethernet destination address: 01-80-C – L1 updates 01-80-C – L2 updates Ethernet source address: source router interface MAC address 802.3 LLC DSAP and SSAP = FE:FE Layer 3 protocol discriminator: 83 Ethernet header Type = 0xFEFE IS-IS header IS-IS TLV Link trailer The IS-IS destination MAC addressing varies depending on the type of information being conveyed. L1 and L2 updates use different MAC multicast addressing. Routers only look for the respective MAC addresses based on their configuration. L1/L2 routers listen for both MAC addresses. The source address in all IS-IS updates is always the interface MAC address of the sender. In the standard version of Ethernet (802.3), when LLC (802.2) is used, the DSAP and SSAP are set to FE:FE to specify that the frame data is IS-IS traffic.

97 IS-IS — Packet Format Details (continued)
IS-IS sends PDUs. PDUs are encapsulated directly into the layer 2 frame. There are 4 types of PDUs: Hello (ESH, ISH, and IIH) — Maintain adjacencies LSP (link-state packet) — Information about neighbors and links, generated by all L1 and L2 routers PSNP (Partial Sequence Number PDU) — Specific requests and responses about links, generated by all L1 and L2 routers CSNP — Complete list of LSPs exchanged to maintain database consistency There are four types of PDUs: Hello PDU — Generated by all devices running IS-IS. The actual format of the hello varies depending on the level the device is configured for. An L1 device sends L1 hellos, an L2 device sends L2 hellos, and an L1/L2 device sends both L1 and L2 hello PDUs, independent of each type. LSP PDU — Generated by L1 and L2 devices to convey information about neighbors and links. LSPs are used to create the topological database for the L1 and L2 devices in an area. PSNP PDU — Used to request specific information about a network. A PSNP can contain a subset of LSPs in the database or can be used to acknowledge one or more LSPs in point-to-point subnetworks. In addition, a PSNP can be used to request transmission of a specific LSPs (seq = 0) on broadcast subnetworks. CSNP PDU — Lists every LSP in the database. CSNPs are multicast periodically by IS-IS routers to maintain database consistency. Several CSNPs can be sent at once when the database is large.

98 Alcatel-Lucent Routing Protocols
Module 7 — Border Gateway Protocol

99 Enables the implementation of administrative policies
BGP Scope Enables the exchange of routing information between autonomous systems (AS) An AS is a collection of routers that are under a single administration, which presents a consistent routing policy. Enables the implementation of administrative policies BGP has already scaled to: Large number of ASs Large number of neighbors Large volume of table entries High rate of change BGPv4, defined in RFC 1771, provides reachability information to foreign networks (outside the AS) by enabling the exchange of routing information between ASs to allow data flow between them. This course uses the term AS throughout. The classic definition of an AS is a set of routers under a single technical administration, which uses an interior gateway protocol and common metrics to route packets within the AS and an exterior gateway protocol to route packets to other ASs. Since this classic definition was developed, it has become common for a single AS to use several interior gateway protocols and sometimes several sets of metrics. The use of the term AS here stresses that, even when multiple IGPs and metrics are used, the administration of an AS appears to other ASs to have a single coherent interior routing plan and presents a consistent picture of which destinations are reachable through it. When the exchange has been enabled, of equal or greater concern is the application of administrative policy onto the traffic flows. Policy implementation is a key strength of BGP and allows the administration to manipulate traffic based on virtually any policy. BGP has proven scalability. It is the protocol of choice for ISPs to run on their Internet routers. BGP is the fundamental building block of the Internet and is used by every ISP in the world for ISP interoperability. BGP is the most feature-rich and scalable routing protocol in use today. It supports the current requirements of the Internet, and with extended capabilities, such as multiple protocol families and extended AS numbers, is well-positioned for the future.

100 Autonomous Systems in BGP
AS-65002 AS-65003 An AS is a group of networks and network equipment under a common administration. IGP protocols such as OSPF, IS-IS, and RIP run in an AS. BGP is used to connect ASs. AS-65001

101 Autonomous Systems in BGP (continued)
Public autonomous systems: Assigned by ARIN or another authority Must be used when connecting to other ASs on the Internet. Range from 0 to Private autonomous systems: Assigned by ISPs (for some clients) and local administrators Not allowed to be advertised to other ISPs or on the Internet Range from to

102 Similar behavior as other TCP/IP applications
BGP Features Path vector protocol: Neighbor is any reachable device Unicast exchange of information Reliability using TCP Uses well-known TCP port 179 Periodic keepalive for session management Event-driven Robust metrics Authentication Similar behavior as other TCP/IP applications Because BGP peers are not always directly connected, BGP relies on IGP to route between peers. Although BGP is an enhanced distance vector protocol, it is specifically called a path vector protocol. Neighbor relationships in BGP are somewhat different to what is normal in the IGP world. Traditionally, a neighbor is always a directly connected router. With BGP, this is no longer the case. Neighbors may be directly connected, but it is not required. Because of this, BGP relies on an IGP to route between peers that are not directly connected. BGP uses unicast TCP/IP for neighbor establishment. It is possible for neighbor relationships to be established with any device that is IP-reachable. There is no guarantee that the neighbor relationship will succeed because factors such as firewalls or access control lists may prevent certain types of traffic from passing, but the relationship is possible and likely to occur. At the application layer, BGP functions similarly to other TCP/IP applications, such as Telnet, FTP, and HTTP. BGP may be viewed as an application because it uses registered port number 179 in the TCP/IP model. Generic TCP/IP applications use a 3-way handshake for session establishment, and when this is complete, a TCP/IP session is formed. After the session, the applications exchange or negotiate a set of parameters for the session. In Telnet, for example, parameters such as terminal types and passwords are typically negotiated. If application-level parameters are also acceptable, a session is established at the application layer and data is exchanged. Periodic user data keeps the session alive, and when the session is to be terminated, either user input or an inactivity timeout causes the application session to be torn down. TCP/IP initiates the 4-way session teardown.

103 2 types of BGP sessions are possible.
eBGP vs. iBGP Overview 2 types of BGP sessions are possible. The routers may be in different ASs: Called external BGP or eBGP Typically directly connected, but not mandatory Different administrations The routers may be in the same AS: Called internal BGP or iBGP Typically remote, but could be directly connected Same administration Two types of BGP neighbor relationships are possible. Regardless of the type, a BGP session between 2 devices is alternatively referred to as a neighbor or peer session. A BGP router is also referred to as a BGP speaker. A session between 2 devices in different ASs is referred to as an eBGP session. It is typical for devices that are having an eBGP session to be directly connected, sharing a common data link, but it is not mandatory. Because the devices are in different ASs, the administration of each device is typically handled separately. Care must be taken to ensure that the configuration parameters match so that the peering will succeed. A session between 2 devices in the same AS is referred to as an iBGP session. It is typical for devices that are having an iBGP session to not be directly connected. Because the devices are in the same AS, the administration of each device is typically handled by the same organization. Care must still be taken to ensure that the configuration parameters match so that the peering will succeed. With the devices locally controlled, this is often an easier task.

104 3HE-02770-AAAA-WBZZA Edition 01
3HE AAAA-WBZZA Edition 01


Download ppt "Alcatel-Lucent Routing Protocols"

Similar presentations


Ads by Google