CCNP Enterprise ENSLD (300-420): Designing BGP Routing

2. IBGP Scalability, Route Reflectors and Split Horizon

Multiple BGP can provide a controlled interconnection between multiple routing domains that are running different internal gate protocols, and it can also support internal multi-protocol label switching. By design, IBGP requires a full mesh of BGP pierced. Now, BGP can be used to interconnect multiple routing domains or multiple internal gay protocols or IGPs, and it also supports MPLS multi-protein label switching at the ends or virtual private networks. So, by design, IBGP requires a full mesh of its peers. Now, in our example here on the left hand side, it means that I have four routers, and then the equation is n minus one, for a total of three sessions per router and a total of six sessions. Even with five routers, it's still a manageable situation as far as the full mesh of the peers is concerned. But down at the bottom on the right, you see that I've got 10 routers, and then it becomes unmanageable and hard to troubleshoot. By design, IBGP requires full mesh communication between IBGP speakers to function. Full mesh pairing is not scalable. The number of sessions required grows rapidly as the number of NUF nodes increases. And for N routers, n equals NM lied by n. One divided by two is the total number of BP required sessions, which means sessions per router. Adding new nodes is a configuration-intensive, time-consuming task prone to errors. Some options can loosen the full mesh requirement. Because IBGP routers do not advertise what they learn via IBGP to other IBGP peers, IBGP split horizon rules modify IBGP as needed. This behavior is part of BGP pro behavior that is used to prevent information from circulating between IBGPspeaking routers in a re-information loop or cycle. Because full mesh peering can become unmanageable really fast due to the large administrative overhead of maintaining many peering sessions, it does not scale well in general for N peers. In a GPful mesh, each router would have N minus one multiplied by n minus one peers in total, meaning that each pair would need the CPU memory and bandwidth to mandate and peer status for all the other routers. This action is not a hierarchical design, and it would not be effective to scale for large networks. For example, a router within us that contains 50 routers would require multiplied by 50 minus one divided by two equals one and two, five IBGP sessions, and the associated route per configuration that would be required for a single router that contains 10 routers. There are two IBGP options for scaling IBD reflectors. The root reflectors modify the classic IBGP split horizon rule and allow a particular router to forward incoming IBGP updates to an outgoing IBGP session. Under certain conditions, this router becomes an contraction router or a route reflector. Confederations Small Autonomous Systems Confederations introduce the concept of smaller autonomous systems within the original as the "Small Autonomous Systems Exchange BGP updates between them by using intra confederation EBGP sessions. Now, BGP uses a rolled-split horizon. This is used in distance vector routing protocols to avoid routing loops. The rules state any route that the router learns. It cannot send that route back through the same interface that it learned it from. The split horizon rule is enforced, but we are going to use some different mechanisms to avoid problems with split horizon IBGP scalability solutions. One of those mechanisms is a root reflector, or root reflectors. Now this can reduce the full mesh requirement for any GPD and it modifies the split horizon rule. So a full mesh is needed because the IBGP routers don't advertise routes that they learn via IBGP to their IBGP peers. This is a known behavior of the BGBTQ and it is again used to prevent routing loops. So in a root reflector design, a rootreflector is actually the router that's going to propagate to the root reflector clients the routing updates. A non-client will still have to have a fully meshed relationship with the root reflector. But a root reflector client can be configured with the IP of the root reflector and receive its root information via the root reflector without having to be fully meshed. Clients may have several external gateway proto color EGP peers, but typically have IBDGP sessions and journal sessions only with their root reflector or root reflectors. If a single root reflector fails there, I've got a single point of failure. So it is recommended that we use redundant rootreflectors to avoid a single point of failure. As a result, root reflector clusters prevent IBGP routing loops in redundant route reflectorines, and the network designer must properly identify which rootreflectors and clients will actually foster. Then the network designer assigns it to the cluster. A cluster ID that is unique within the autonomous system can scale your BGP network by introducing BGP root reflectors into the design. A route reflects an IBGP speaker that reflects learned routes from Ibgpiers to other IB Gppers. Root reflectors are enabled by defining their Ibgpir as root reflector clients. The Route reflector club needs no additional configuration. The number of required GP sessions drops significantly, and it may be easier to manage and troubleshoot. A BGP route reflector is an IBGP speaker thatreflects a reariser's routes that are learnt from IBDpeers to some of its other IBGP peers. Remember that the other IBGP-speaking Raptor is not allowed to teach any IBGP routes that it learns from it as.This rule is relaxed by introducing a root reflector, a special type of IBGP node, into your BGP network. You will need to select the best candidate for the role and configure reflector functionality. The configuration of the root reflector is done on the route reflector itself. The configuration identifies which BGP peers are root reflector clients implementing root reflectors as failpull and can be done incrementally. Apart from defining root reflectors as peers, no special configuration is needed on the clients themselves. Each client router needs to be configured as an aclient on the root reflector or audible root reflector. Unnecessary peers can then be removed from the configuration on the client router. Route reflectors only peer with root reflectors in a design route. When root reflectors are introduced, the number of peers in a network with many IBGP routers drops significantly. In general, routerition is the process of redwatching received routes from one peer to another, including the originating node, on a route deflector router. BGP scalability solutions Another way to scale your BGP network is to introduce BGP. Confederations BGP confederates partition VAS into multiple sub-autonomous systems. Confederations provide another way of entering information into BGP routes to prevent loops within them. Rather than adding new attributes, confederations add more information to the A path. In RFC 5065, in a confederate, the A's uses multiple subautonomous systems to scale the outer. A.S. is called the confederate. S is a sub-autonomous system. Each sub-autonomous system has its own private S number. Route reflectors can be used within a subautonomous system to separate multiple subautonomous systems. Confederations insert information using the A S to BGP route to prevent loops within an A S. The basic idea behind confederations is to divide anon BGPaS into multiple sub-autonomous systems. The outer or containing A S is called the confederation, as it is all that is visible to the outside world. Each of the inner autonomous systems is a smaller autonomous system that uses a different ASNumber, which is typically chosen from the private ASNumber range of 64 512 through 65 534. The behaviour is defined in RFC 5065. The BGP confederations contain the full mesh within a subautonomous system, thus lowering the total needed BGP connections in the network, further scaling your network. The use of root reflectors is not allowed within a sub-autonomous system. You can use BGP confederations when you need to partition your network based on structural or geographic requirements. Each synonymous system can run a different IGDP if that is what is needed. Confederations can also be used during a company merger to merge two BGP networks, typically using private A numbers, into a consolidated network, with the newly created confederation using a common public AAS when communicating with the rest of the Internet.

3. BGP Route Reflector Terminology

A root reflector client is a root reflector node's appear. It can send and receive reflected routes from the Root River routers. From the reflector perspective, a nonclient is defined as any root reflector IBGP peer that is not a root reflector client. Each root reflector is also a nontra root reflector in this network. Root reflectors must be fully IBGP meshed with nonclient clients as any IBGP pierto which the root reflector reflects routes. To configure a root reflector clan with a rootreflector, the pair must specifically be designated as a client by a special BGP configuration. A nonclient is a regular peering definition on a route reflector with a peer that is not specifically configured as a root reflector client. Other BP network root reflectors are typically non-clients to other BP network root reflectors. Root reflectors must still be fully IBGP meshed. Non-resilience root reflectors reduce meshing within clusters,but all mesh links outside the cluster must be maintained on the root reflector. The Root Rift clients will get information from IBGPspeakers outside the cluster via the root reflector. Clients may have any number of ebgpos but typically have IBGP set only for their root reflector or reflectors. If the single root reflector fails, its clients can no longer be GP updates to receive them from the rest of the AAS. The root reflector is therefore a single point of failure. To avoid a single point of failure,redundant route reflectors are typically used. An IBP network, which is based on onroute reflectors, relies on them for convergence. If the only root reflector in the network fails,clients will lose the ability to advertise their prefixes. Clients should establish an IBP session with at least two root reflectors by using different physical connections. The concept of clusters is used to prevent IBGP routing loops between root reflectors. The root reflector's functionality is asredundant as the physical network. Clients will establish BGP sessions with multiple rootreflectors by using redundant physical paths. The concept of clusters was introduced to prevent IBGP routines. When you use root reflectors in your network,both root reflectors will receive the same IBGP update from Ions, and they will both reflect the update to the rest of the clients. Also, both root reflectors will get up from the full mesh and reflect those updates to their clients. Under certain circumstances, each client will get two copies of all routes. This is particularly true when you use weights on IBGP sessions to influence BGP selection. Improper route reflection can result in an IBGP routing loop that is impossible to detect. TrabGP attributes are thus necessary to prevent these routing loops. Route reflector clusters prevent IBGProuting loops in redundant routes. The role of the network designer is to properly identify which root reflectors and their clients will form a cluster. The desire to give the cluster a cluster ID number that is unique within the A as redundant rootreflectors and they form a cluster within the cluster,each client peers with all root reflectors. Root reflectors from differentters need to be fully meshed with each other. Routers that do not support root reflectivefunctionality have to be fully meshed with root reflectors from all clusters. The clients will communicate with all of the reflectors in the cluster. For reducing root reflectors from different clusters, they need to be fully meshed with each other. The exception to this rule is in a hierarchical design where a root reflector is a client of another root reflector. Routers that do not support root reflector client functionality should be fully meshed with root reflectors from all clusters and with other non clients.Clusters of Root Reflectors The cluster ID number can be explicitly configured in the root reflectors. If it is not explicitly configured, the router's ID is used as the cluster ID. The clients should not be configured with this information. IBGP behaviour depends on the cluster ID condition. If devices are configured with the BGP cluster itrouter configuration command, the cluster ID is shared. If the BGP cluster command is not configured,the router ID is used as the ClusterId. Cluster ID is ostensibly the cluster list attribute. On reflection, when all root reflectors in a cluster share the same cluster ID,this is referred to as a redundant cluster. because the cluster ID is isomorphic between the root reflectors. The routes that are exchanged between rootreflectors will be discarded because of the thecluster list, thus saving router resources. However, it can lead to missing routes and suboptimal routing in cases where ten BGP sessions go down,either because of a misconfiguration or software fault. Occasionally, root actors do not share an identical cluster IDas in the right hand portion of the figure. Then there is an oven cluster. The clients still connect to two routereflectors for redundancy, but from this perspective,each route represents a separate cluster. To prevent loops, a root reflector adds the router ID of the reflector to the clusterlist of routes that it reflects. IBGP, the BGP route selection process is modified so that it includes the criterion of the length of the cluster list, with a shorter cluster list being preferred to those roots that have a longer list. Nonredundant In some scenarios in a non-rod area, the client routers have a single physical connection to a root reflector router. These routers form a redundant cluster. The router that is designated as the root rootreflector in the cluster is already a single point of view in this physical design because a failure of this router will prevent the clients in the cluster from reaching the regular network. Therefore, there is no actual benefit to introducing another route reflector. The use of root rootreflectors in a basic network allows for redundant or non-redundant clusters. A redundant cluster uses two or more root reflectors. There can be many redundantclusters in one autonomous system. The root reflectors should be pushed with a redundant cluster that has only a single rootreflector and a single point of failure. These should be of If possible, Root reflectors can either be dedicated to notrouting any traffic or in the path of The preferred way of designing the route reflector network is by using a redundant cluster, which relies on the unreliable network redundancy. When using multiple clusters, the route reflectors should be fully meshed unless they're hierarchical root reflector design, which will be covered later in this lesson. An ABGP network can be designed to dedicate root reflectors that are not actually participating in user data packet processing but managing distribution tasks only. Hierarchical Root Reflector Design: In Hierarchies, you can create a root reflector club. With hierarchies, a router that serves as a root reflector in one cluster can act as a client in another cluster. When configuring an Iibgp session on a rootreflector, you must configure the session to reach TT in order for the root reflector's IBGP Split horizon rules to start working. All other IBG sessions that are configured on the root reflector are part of the full mesh. The problem is that this full mesh portion of the network becomes too large to manage. A router that is configured to be a root reflector will still have ordinary IGP sessions that are part of the full mesh. If these sessions are reduced in number and only a few remain, and the route once reaches a second level of root reflectors, a hierarchy of root reflectors is created when you build the first level of clusters. The remaining full mesh is smaller than when all the routers belong to the full mesh, but if the remaining mesh is still large enough, you can build an extra level of root reflectors. The cluster list attribute plays a major role in a hierarchical root reflector BGP network as a key loop prevention mechanism. A single layer of root reflectors may not be enough in very large works. Therefore, a hierarchy of rootreflectors can be established. A root reflector can be a client of another root reflector. The hierarchy can be as deep as needed. The list plays an important role in loop prevention. The cluster list will be long enough to cover each reflection. If a root comes back to the root reflector that has already reflected the route, it will be dropped. The figure shows an example of hierarchical route reflectors. The first level of hierarchy is reduced to a full mesh of twelve routers. All enterprise network routers to a fullmesh of seven routers or three root reflectors, and the upper two root reflectors and two clients. The second level of the root reflector cluster was built by creating cluster 27. This second step further reduced the full mesh of seven routers to a full mesh consisting of only two routers as upper root reflectors. Only the two root reflectors in a cluster should be connected in a full mesh. When a client at the lowest level receives an EBGP update, It will follow an update on all configuredIBGP sessions to a Root Reflector.The Root Reflector recognises BGP updates that are received from configured clients and will forward these updates to all other clients that use normal EGP sessions. The updates sent on a regular basis are second level client updates. Second Level Root Reflector The second level root reflector will recognise that the update was received from a client. It will forward it to all clients and into the full mesh.

4. Describe BGP Split Horizon

In interior routing. A split horizon route advertisement is a method of preventing routing loops in distance vector to calls byprohibiting a router from advertising a route back onto the interface from which it was learned in NGP. The split horizon is enforced in a slightly different way by using extra mechanisms. rule that prevents routing loops. EBGP and IBGP manage the split horizon differently. BGP relies on the A S path to prevent loops. If a router sees its own A S in the A S path, the route is discarded. IBGP relies on a firm rule that IBGProutes will not be read by other Ibgpas. EBGP and GP use different mechanisms to avoid routing. EBGP relies on the AS path to prevent loops. If a BGP speaking router detects its own A S number in the A S path attribute of the routing update, the route is discarded because it originated in A S 65 100. In fact, if the A S number appears in any place in the A, it will be discarded. It is a clear sign of a looped route update. If a root reflector receives a route from EGP,it reflects the route to all clients and nonclients. If a root reflector receives a route from a client,it reflects the route to all clients and all ebgpers. If a root reflector is from a nonclient, it reflects the route to all clients but not to non clients.Nonclients will be meshed. It also sends the route to all Ebgppers. Horizon Peer Rules are split in BGP. The topic discusses different scenarios and modified split horizon rules from the perspective of a root reflector. Split horizon rules are relaxed in the root reflector design scenario, and now you will learn exactly how modified rules apply. As with any BGP route, the route needs to be the best to be reflected. If a route is from an IBGP peer, it will not re-visit that route to another IBGP. If the Igpiers are in the same A S, and they do not add anything to the ASPATH, there is no way to tell if that is advertised through several IBGP speakers as a loop. Full mesh. IBGP must perform r to r four. IBGP peering the IBGP split rising rulesmandate that received updates on EBGP sessions should be forwarded to all IBGP and EBGP sessions. However, updates received within an IBGP session should be forwarded to EBGP sessions only. There is no way to tear the route that is advertised through several IBGP speakers as a loop. Because Ibgpiers are the same, they do not add anything to the A S path. Therefore, they must not really advertise routes that are learned via EGP to other IBGP peers in the network. In the figure, there are no route reflectors Rtwo tises a route that is received from anEbgpa into the IBGP section. The route gets toR, which does not advertise it to R four.because of the split horizon rules of the IBGP. The route can read four only if there are two peers. A full mesh with three routers requires only three IBGP sessions. If a root reflector receives a route from an EBGP peer, it passes that route to all route reflector clients and nonclients. just as in normal IBGP peering behavior. If the root reflector receives a room from a rootreflector client, it reflects the route to the other clients within the cluster and to non clients.It also reflects the route to EB GP peers. The root reflector takes over the communication for the root reflector client, tossing along all the messages that they would normally transmit directly via the appearing section. If a root reflector receives a route from a nonclient, it reflects it to the rootreflector clients but not to other nonclients. This section occurs because nonclients are supposed to be fully meshed with other nonclients and root reflectors. The root actor receives the roots. If it has a direct peering relationship to the original nonclient, the root reflector would send those routes to EBGP peers, which is standard behavior. IBGP routes are repeated to all GPUs as per normal BGP behavior. The route that is to be reflected must be the best route to a specific destination. If an identical root is received on the root reflector from two or more different clients, under normal circumstances,only one will be reflected for a year.

5. Route Reflector Loop Prevention Mechanisms

Two nontransitive optional BGP attributes, the originatorID and the cluster list, are used to prevent dangerous loops of information in BGP networks that use root reflectors. Additional attributes for loop prevention with root reflectors include the originator ID of the root originator sequenceof cluster ID values in the reflection path. A route with an originator ID that is set to rewrite router ID is discarded. A route with a cluster list that contains the receiver cluster ID is discarded. selection modification and follows RFC 4456. The shorter cluster is preferred and makes clusterID configuration obsolete with the originator ID attribute. When it is reflected on the route reflectortoward other clients, the root reflector sets the originator ID attribute of the route to the router ID of the originating router. Any router that receives a route with its own route ID originatorid attribute silently ignores that route. The cluster list attribute offers a loop of whiteand ism when multiple route reflectors reflect the route. In a scenario where there might be many layers of root reflectors within the network, the cluster list is an asequential list of cluster IDs that prepares the router IDs, if configured, of the root reflectors that havereflected the route along the way. You might consider the Russian path that the route has passed. If a root reflector finds its router ID in the cluster list, it will be considered a possible loop. The BGP path selection rules have been modified to select the best route. In this situation, The router might receive reflected and non-reflected routes or several reflected routes. These include the traditional GP path selection parameters such as weight, local preference, origin, and medare compared first, so if these parameters are equal, the routes that are received from EBGP neighbours are preferred over routes from IBGP neighbors. When a router receives two IBGP routes,the nonreflected routes with no originator ID attribute are preferred over reflected roots. The reflected roots with shorter lists are preferred over routes with longer cluster lists. If the additional route reflector-oriented selection and cryado do not yield a decision, the rest of the traditional BGP path selection rules are followed. The introduction of the modified path selection rules makes it possible to use routereflectors without an explicit cluster ID being set. In such a scenario, the cluster list attribute will rely solely on the router IDs of root reflect. This action adds more resiliency to the network at the expense of slightly higher router resources. This behaviour is defined in RFC 4456 LoopPrevention and the root is reflected. The route reflector adds its cluster ID or router ID to the cluster lists. Cluster Topology In this example, the same cluster ID is configured on both root reflectors and an article. Cluster ID prevents a root exchange between RROne and RRTwo, thus preventing loops. When a root reflector adds its cluster ID to the cluster list attribute, the clusterID value outreflector can be taken from the BGPcluster ID if it is explicitly configured, or the Bgpid if it is not explicitly configured. If you consider the first part of the example using a single cluster with two rootreflectors, which are configured with the identical clusterID of 1010, two-five-four, When client A sends a root update to both reflectors, the reflectors reflect other clients and to each other. When RR One sends the update to RTwo, RR Two inspects the call list of the route and finds its ClusterId in it. Therefore, it must discard the route.The same action happens with RR Two-to-RR One loop prevention in the overlappingcluster. Each route reflector has a different clusterID, so root exchange is possible. From client B are preferred because the cluster list is shorter than the reflected route coming from RRThree. One. The cluster list length prevents rooting loops by influencing the best routes. In the overlapping example, you can see the overlapping clusters with two root reflectors that have theirIDs set as the cluster ideas. It puts them into different clusters, yet they share common clients among overlapping clusters. When client B sends a root update to RR Three and RR Four, the root reflects the roots to their clients and nonclients. The difference is that for a route from RR Threeto, RR Four will not detect its own cluster ID in the cluster list, so it will accept the root. RR Four now has the route which is sent from two sources: client B and RR Three. Because the cluster less attribute of the route from tomb is shorter than the same attribute of the route that is coming from RR, three from client is considered the best path for that destination. The clients will end up with multiple copies of routes, but again, each client will select the best path that is based on the same selection process. Problems with BGP Root Reflectors Network design should adhere to strict net design guidelines. Deviations can lead to potential problems. Some of the problems that could occur if you deviate from root reflector network design rules are that some clusters will not have all the roots if root reflectors are not connected with IBGP sessions in a full mesh. If a client has IBGP sessions with some root reflectors in a club but not with all of them, the client might miss some BGP routes. If a client has IBGP sessions to root reflectors that belong to different clusters, the BGP update from the client will be forwarded into the full mesh with different cluster IDs in the cluster list attribute. When the BGP update enters the mesh, it will read other root reflectors, which will unnecessarily accept the root as valid and forward it into their cluster. This causes unnecessary duplication of updates to the clients. If a client has IBGP set to other clients in the same cluster, those clients will receive unnecessary duplications of updates and modifications of root attributes on Root reflect before reflecting the roots, which can lead to suboptimal routing or even potential networks.

ExamSnap's Cisco 300-420 Practice Test Questions and Exam Dumps, study guide, and video training course are complicated in premium bundle. The Exam Updated are monitored by Industry Leading IT Trainers with over 15 years of experience, Cisco 300-420 Exam Dumps and Practice Test Questions cover all the Exam Objectives to make sure you pass your exam easily.