Introduction to SRv6

Oh, or with our experts here, our customer success specialist Gopal and Moise will be leading the session today on the topic of.00:00On the series. This is the 1st session of the SRV six series that we have. The upcoming ones will be later during in the following weeks. But the 1st session, the topic of the 1st session is to get you introduced to SRV six.00:20And, also this is an interactive session, so please feel free to raise your hand if you have any questions and I'll unmute you to ask your question live with the experts and enjoy the session. With that.00:40Gopal, please take it away. Sure thing, thank you. Hello everybody. Good day. Welcome to introduction to SRV six webinar. My name is. I am a customer success specialist in.00:58Customer experience organization in Cisco. I have strong experience in service provider technologies. I hold two ccies, one in service provider and one in enterprise track. Today.01:17You want to say hello? Yeah, good morning, everyone. Thanks for joining. My name is and I will be doing the mostly the lab demo for this session. I have been working with service provider networks. I have a CCI in SP and data center and.01:32And passionate to work with sharing the knowledge.01:51Alright, today Moz and I, have the pleasure of walking you through this webinar on SRV six. Like Bhavana said, the location for you to either ask question.01:56There are two ways you can do it. One is by raising hand, which is on the bottom right hand side of the Webex screen or by clicking on the Q and A and typing in your question. We will try to answer the questions as much as possible if if not.02:13But we will cover them in the future webinars are touch base with you one on one so we are going to discuss SRV six today, but before diving in, I just want to set the handraise from Dennis Springer.02:33Okay, ok, so how will I unmute? Anyway to unmute? I don't see the hand raised anymore, maybe.02:54Okay. Just for the sake of testing, I will try to unmute Randy and see if it works. Hello it was sent. Oh yeah, hey Randy. Okay, thank you for testing. Okay, so I will mute you back again. Thanks, Randy.03:12All right. Okay, let's move on alright so what is SRV six? SRV six is a source routing underlay networking architecture.03:32So here the source node will define the path as well as the packet treatment to be done within the rest of the network. SRV six provides an elegant method to transport service traffic such as L three VPN and L two VPN.03:48Right? And this helps the service providers provide differentiated services. So one thing at a at the beginning I want you to get the idea that we are talking about networks that support multi tenancy. This is applicable to.04:04Single tenant network as well, but most of the benefits you will reap when the networks are service providers, large enterprises and data centers. Throughout this presentation, at least I will go with that mindset.04:23And then we will cover basically the benefits and SRV six fundamentals in this session, and followed by a demo at the end. This is a three part series on SRV six. In the upcoming sessions, we will discuss.04:41The features and services and the another session for implementation where we cover them some part of design as well as migration. Alright, so let's talk about some some introduction introduction and benefits.04:57So basically, you know, how did we get here? So in nineties, Cisco, had this tag switching, right? And then which was evolved into MPLS and broadly.05:16Adopted by other vendors after it was standardizing 2000s. So then service providers adopted MPLS to provide services to their customers, especially L two VPN and L three VPN services.05:36This is the earlier there was similar technologies, but then MPLS gives any to any connectivity over a shared infrastructure. And in addition to that.05:54It was also good, it's a good underlay architecture because it provides the the typical traditional service providers requirement of 50 milliseconds hit hit time, right? So the if there is a network outage that it should recover within.06:10In 15 milliseconds so that it can support wise traffic.06:30So, MPLS is great it it requires some complex protocols and complex operations team to, to, to operate the network. So in in early 2010s, Cisco.06:34Introduced something called segment routing. Segment routing, is an effort to simplify MPLS and also make it more SDN friendly and programmable. It supports segment routing supports the same technology.06:54Be as MPLS, actually it uses the same data plane.07:13Only the control plane changed and its control plane was simplified. So this is this is the progression of the service provider's technologies. Though I am saying service provider and MPLS, the same thing is applicable to any network that that is there, who wants to be.07:17You make use of MPLS benefits, they can actually even a small enterprise can use MPLS if they want to use some benefits that are unique. Okay, so.07:36At a very high level, what is segment routing? So segment segment routing is a modern approach to MPLS where, each segment is considered as a, as an instruction.07:54To the network or the router that receives the packet. So you can have a set of instructions in the packet header and the router, the transit routers process the packet as per the instructions within that segment.08:10So whatever were called labels in MPLS, the terminology changed to segments or instructions in segment routing. In traditional or legacy MPLS, the labels are dynamically allocated as well as distributed.08:27In the case of segment routing, the label allocation is, there there is a static label allocation per node, but there is also a dynamic allocation for all the interfaces, right? And the distribution of the labels is still.08:43Same as the MPLS. No, the distribution of label there is difference. In MPLS we have a dedicated label distribution protocols such as LDP, RSVP. In the case of segment routing, it uses the IGP routing protocols with the extensions to distribute the label.09:02Right, so it uses the same data plane and it is simplified, it's SDN friendly and programmable, right? Most importantly segment routing does not require LDP or RSVP.09:24Right? So the core network can be stateless. It need not know the information e.g. of the traffic engineering tunnel ID and things like that. The.09:39The reservation of the bandwidth on the interfaces, so on and so forth, need not be, the state need not be maintained on the transit routers. It uses the same.09:55Use cases such as traffic engineering, fast reroute, network slicing, but more importantly segment routing comes in two flavors. Segment routing using MPLS.10:13And segment routing using IPv six. Segment routing you using MPLS has been there since early 2010s. Now what we are talking about is segment routing using IPV six as the.10:27Data plane as the transport, right? So SRV six is segment routing where the data plane is IPv six. All right. How does the packets flow? How does the.10:44Source routing works in segment routing. So what happens is, in this, in this scenario, R one has two packets to send to R eight. Here what R one is doing is it's sending packet one with.11:03One set of segments and the packet two has a different set of segments. So based on the segments, the routers in the transit will forward the traffic. That is why.11:24Because R one is the one which pushes these segments into the packet header. So R one is the one which is controlling the traffic flow of these packets, right?11:40Let me be a little bit more specific. Assume in this network what happens is each router has one node segment ID.12:01Right, so e.g. R eight, it's static segment ID is 16008. You you can have more than one, but I'm just giving an example. What happens is these segments are distributed across the network using OSPF or ISIS.12:16So when R one, based on the policy of the, and then the packet that it received and it based on if it matches the policy, it will make the decision what segments to inject into the packet header.12:34Okay, so ISIS are OSPF are the ones that distribute the labels. And then based on the policy, the ingress router or the router that is initiating the packets will.12:51Inject the labels into the packet header, right? Because, segment routing uses IGP, and for other reasons as well.13:08The native ECMP or multicast equal cost multi path load balancing works fine in segment routing. This is a big challenge for legacy MPLS because for the the labels change and for a given label, there is only one next hop.13:24So there there is a mechanism that later came.13:44Entropy label and so on, but it is a complex thing. So in segment routing load balancing is inbuilt natively, right? As you can see here.13:50This is the set of instructions I was talking about. The source node has these instructions, so the packet goes to R2R2 will receive the packet and pop the label and then we'll forward the packet based on the next label, which is 16006 in this case.14:06So in summary, if we look at the data plane, the evolution is that instead of MPLS, we are using MPLS labels, we are using IPV six header itself to forward the traffic.14:27The transport protocols in legacy MPLS, there is a IGP such as ISIS or OSPF, then MPLS LDP, RSVP RSVP, protocol and then BGP labeled unicast.14:44All these protocols together work as the control plane for the legacy MPLS, whereas in SRV six we simply have ISIS version six or OSPF six, right? The service protocols.15:02Earlier we have L three VPN and L two VPN there with the help of LDP and MPBGP are supported. The same are supported in SRV six as well but simply by using MPBGP. Okay, so with that introduction, let me talk about some of the.15:20Benefits but before that let me let me let me just go over the SRV six features. Moz I just want to do a check. Am I going too fast? Are there any.15:43Thanks.16:01Okay, I will try to slow down, but if there are any questions, please let me know. SRV six supports very powerful features. Some of these features are also supported by legacy MPLS as well as well as MPLS segment routing SR SRMPLS.16:04But this is, these are the features.16:24Partial list of features that SRV six supports. The traffic engineering using SR policies so that it will be the configuration is much more efficient. Flexible algorithm are it's also called network slicing. Load balancing a big.16:26Upgrade from the legacy MPLS egress dynamic egress peer engineering micro loop avoidance.16:46Very, network observability and automation, topology independent loop free alternative SRV six on the host with NFV so we can have SRV six all the way to the host.16:55Right, so performance measurement path tracing and traffic traffic accounting. These are some of the features, but these are the features which helps the any typical underlay architecture because when the, when the network is.17:13Hosting services and carrying the services across the network. These combination of features are very useful to provide, very high uptime and some some provide monitoring and operational efficiency.17:30Engineering efficiency and so on the benefits. So with SRV six, we, we have the extreme flexibility to create service, services based on the service provider.17:48Or large enterprise definition, they can.18:08Use a SDN controller and their software is the, you know, limiting factor in creating those services. So end to end policies can be played on the ingress and egress routers so that.18:12The packets can take a specific path and then I'll be going over some of the features where a function and then arguments can be passed to the routers so that they can process the packet as per the the the function of the argument.18:28Right, the biggest and the most benefit is unified data plane. The in the current MPLS networks, the network to to take care of the complexity and the high scale of the network, service providers.18:48Create these, you know, imaginary domains in the network where, you know, call it core domain, you know, edge domain distribution, domain, data center, what not. With the SRV six.19:07SRV six, what happens is you can have one domain, right? You can still have, IGP areas or OSPF areas and iss areas so that your routes are controlled in a, in a simplified manner, but.19:25The data plane point of view, you don't need to have these domains and especially at the edge of the domain, the kind of stitching takes place and you know huge configuration and high expertise of operations engineers, all that is reduced and you can have.19:43One unified data plane from the host to host, right? Less is more it simplifies the network by eliminating LDP.19:59And RSVP, right? This is a big elimination because lot of control plane activity takes place. In a, in a tunnel, if you have a typical service provider network with 10000 tunnels, you can imagine the kind of RSVP messages keeps on going across the network. It's huge.20:15Right, and also load balancing is, inbuilt especially when you, when you transport traffic such as mobility traffic. It's a, you know, for all the mobility operators, it's a big pain point. That, that is not there right from the design itself.20:36So the natively load balancing is supported and simplicity and scalability, it it is simple because there are NO interests to deal with the prefix summarization is easy and then you can use IPVC.20:55Six routing, and then summarize it and then.21:15It's a simple to operate. All right, so those are the high level view of the benefits. Let's look at some fundamentals of SRV six. So I just want to check with you if there are any.21:20Yeah, let me see.21:40I don't see any questions very good. Good. All right, note to self slow down, ok. All right, so you're doing good spot.21:42Thank you. So, like I said, segment routing and SRV six leverages the source routing paradigm. The source node encodes the traffic encodes the instructions into the packet so that the.22:04The network will, do the correct treatment of the packet in terms of forwarding and other way like such as QoS as well, right? So.22:21And then, as you can see, packets can be, The source node can influence the packet the path. So at a very high level, so the source node carries the payload and adds its.22:37Set of instructions into the header and that is how the packet, the network treats that packet using the whatever is there in this header. Just to recap, if we look at path options.22:55There are, you can simply use IGP in this network and use IGP for calculating the path, right? So router A will receive the entire.23:11Information in either ISIS or OSPF and from its database it can compute the path based on the policy that the packet matches. Another option is it can use an SDN controller and then receive the path explicitly for.23:26For a given packet, right? The control plane, you can use ISIS or OSPF BGP or again SDN controller using BGP and path compute element protocol.23:46Our netconf fang. These are the ones that talk, if it is a PCE talk with PC PCC in this case this router or PCP piece of protocol, if we have, if the customer wants to maintain multiple domains and.24:02The the information is leaked through PSAP. So these are the options for the control plane. For the data plane segment routing has, can use either MPLS or IPV six today we are going to talk about the IPv six part of the segment routing. So we are calling it SRV six.24:22Yeah, so what what we did is so this is a source node, so typically imagine a situation of a service provider. It it is applicable to anywhere.24:42Like enterprise also, a customer's packet is coming into the SRV six domain or the service provider domain. If you see this, if you are from any enterprise, you can say that this is enterprise core and this is a.25:02Department or a group that is using this enterprise core as the transport. When the packet comes in, what happens is this router will encapsulate the packet. So this payload and this packet header, whether the packet is IPV four or IPv six.25:19That internet. This packet is encapsulated with IPV six header and then will be forwarded to SRV six domain. Okay, so at a high level, that is what happens.25:39Okay, so I want to go back to IPV six a little bit and come because there is a lot of parallels, it helps to connect you with that typical IPv six address or other parallels.26:03So in IPv six we have 01:28 bit address, ok? And that address is broken into two parts, just like IPv four address, it's imaginary breaking point. There is a network part and there is a host part.26:18And the routing takes place using the network part of the address. I told you that SRV six uses IPv six as the data plane. So when a packet is received five by.26:34On either a SRV six router or a IPv six router, they consider it is the same thing. They consider you know part of it as network and part of it is host based on the routing proto routing table that they have, and then.26:50Forward the packet based on the network address. This is the packet header of SRV six. If you notice.27:06It is the same. It is the same address, right? So earlier we were calling it as, network and host. Here, we call it.27:21Locator, and segment, right? So in this case, in this example, this part is sorry these two FCBBB and BB zero zero are.27:38Locators, this is locator, and from here on, these are segment IDs. See, I I want you to I want to take a moment here and then just.27:55Comprehend because I just told that SRV six injects instructions into the packet header and this is how it is injecting instructions. So it is telling that, ok.28:10Route this packet, that's it. So the routers in the transit, they have a routing table to route this packet. So based on the routing match, the highest most significant bit match, the router will forward the packet.28:26Router does not know whether we call it locator or network or host, right? But for our conceptual understanding, so this is the locator. This is a node address, node address, node address, node address. In this example, there are six node addresses.28:45Infused into one header. Not to worry, I will go into details, but I want to warm you up with the change between IPV six address and SRV six address.29:02To a router, to a untrained eye, they, they are the same actually. Actually, they are the same, but how the router process this header is different. Okay, so.29:18Going back again, this is the IPv six header. So this is the destination address, this is the source address. And then below it will be the payload, right? This is a typical 40 byte.29:35IPV six header. So what happens is in the SRV six, a new extension is added to the header.29:52So, there is a field here in the, in the, in the previous slide called next header. So next header if it is 43, it is a routing header. Routing header can be the TCP UDP voice.30:06SPF and so on and so forth. And then if the, if, if the field of a routing type is, four, then the this is a SR header. So the routers know.30:26If a packet is a SRV six packet or a IPv six packet or whatever packet based on the, these fields in the packet, right? So.30:45Hold your thought there. So that's all I want to cover because there are others other other slides explained a little bit more. So before I, again, I want to show you a high level and then go into details and details and high level.31:01So I want to take a minute, step back and talk about, at a high level, what have you have, what have I spoken or covered so far? SRV six domain. So.31:17Typically this is a service provider or a large enterprise core network or service provider network, and in this case, customer at site A is talking to the same customer within another site, site B So in this case, this domain is a trusted domain.31:33And the packet is coming from a partner, right? So basically, the entire network works as a large computer, and then process this packet based on how this ingress.31:53Node, inserts that packet into this domain, right? So for a customer, this entire thing is a large computer, so it's sending it to the computer and it is sending it to the to the destination. In this IPv six domain or SRV six domain.32:12Some routers are maybe SRV six capable. Some need not be SRV six capable. That's an important thing to know. Even even if your network, some of them don't have SRV six capability, you can still have them.32:32In the transit. However, they should be able to support IPv six. Given that now all routers, as far as I know, they support IPv six, it is so prevalent. So, you can have them in the SRV six domain.32:51The ingress node. SR ingress node is the one which will inject the traffic into the SRV six domain. When it injects the traffic into SRV six domain, what it does is it takes the customer's packet or the packet.33:12Get coming from?33:31External and then add some encapsulation and put it into the network. The ingress packet can be a, IPV four packet or IPv six packet or a or a regular ethernet packet without.33:33IPv six. So are any IP headers? So, another thing, thing to remember, the ingress node has to be SRV six capable because it, it has to do understand SRV six and it has to properly.33:51Implement or inject the instructions into the packet ok transit node. Transit node is the router.34:09Which simply forwards the packet based on the destination address and the destination address of the packet is not itself, right? So when the destination address is someone else, all these routers, whether they are SRV six capable or IPV six capable, they operate the same.34:28Same manner. They simply use destination address and their own forwarding information base or routing table to.34:47Forward the packet. So in this case, it is simply forwarding the packet based on the destination address without changing any headers. Obviously, some headers will change the TTL and so on, but.34:54I don't know if, yeah, some, but at large, nothing else, NO SRV six operation takes place in the transit node. Endpoint node. Endpoint node is the.35:10Is the router where destination addresses itself, right? So, any SRV six router in the network can be an endpoint node. So if the, so destination has it has to be itself so that the.35:28Router will do the further processing on the header. If the, if the, if there are NO more segments in the packet, right? So that means this is the end of the SRV six domain.35:44Typically, there will be functions within the packet header.36:04Such as e.g. decapsulation, right? In this example, whatever is sent by the partner received by the ingress node and injects the SRV six headers and then here.36:08This endpoint node, in this case the egress node will decapsulate and send the traffic to, the th the destination. In this case actually these intermediate nodes also can be endpoint but with different functions, not necessarily the decapsulation function.36:25Okay alright so in the SRV six domain, the packets integrity is maintained. No, NO, nothing will change in the packet. All operations are applied on the outer header, and.36:43And we in the SRV six domain, we use SRV six or IPV six headers in.37:03Instead of MPLS labels. That's one big difference. That's about the the high level view of the packet treatment within SRV six domain. With, with that I want to go to the.37:10Specifics of the packet and then let's take a look at it. How the packets, the the headers will change within the, within the packet. Okay, so this.37:27Is a typical SRV six domain and this is the ingress node, right? Here I just elaborated the packet header. So the bottom two rectangles are the.37:43Payload and customers packets header. And this is the service provider IPv six header added to it as part of the encapsulation. If you look inside this header, there is, there is a destination address.38:01And there is an optional SR segment routing header, right? So we can talk about when we use this SRH, but in this case, let us say that there are say ten.38:17Hops, so then six hops can be fit into this, the destination address itself and for the rest of the ones we use SR header. If there are a lot more, we can use multiple SR headers.38:36Okay, now let us look at the format of segment ID. So the segment ID is nothing but the actually a few few slides ago.38:54I said the destination address. So the entire destination address now has become set of instructions. Now, in, in the SRV six parlance, we call it.39:10Segment ID container, right? Because it contains several segment IDs. So the segment ID container is a 01:28 bit field, which is a IPv six address.39:27Yeah so within that segment micros are SID container segment ID container, there are.39:44It is divided I mean imaginarily imaginary division within these 28 bits. Wait a bit for the micro, I will talk about the micro segment part, but consider this as a segment ID.40:03Right in within the segment ID, 01:28 bits, it is divided into locator block, locator node, function, argument, segment location.40:20Later, right? So locator block and locator node together is called locator.40:40Okay, so this is not complex at all. Let us take a IPV six address. When a customer receives IPv six address from, what happens is they get a block.40:48Right a 64 block or 48 block or something like that or even 32 block/thirty two. So the customer can divide that block into 01:28 bits, right? So they keep the same block across the network. Some customers cannot also get multiple blocks or some customers.41:05Get one block and they get merged and they end up having multiple, address blocks.41:25But in any case, segment locator consists consists of locator block and locator node. So if a customer gets a 32 bit address, they can further, they can keep the 32 bit as a locator block and.41:31They can arbitrarily allocate a locator node. E.g., router one, I allocate 0001 as the locator node. And my locator block is FCBB 2100.41:51And then this will be 0001, right? That will become the segment locator. Though I'm trying to explain this, if you are familiar with IPv six, this is a NO brainer.42:12It is, calling some bits something. So in this case, I'm calling the 1st 48 bits as the segment locator function. Okay, while segment locator is used to.42:27Recheck given SRV six node.42:47Right using routing because this is address as well, right? So this will be advertised in the ISIS or OSPF and used by the routers to reach a given destination. Locator is nothing but you can, a equivalent is loopback address.42:50Right, so here the loopback address of router one, I allocate 0001, right? So locator block and look the locator node together can be used to route the traffic to a given router. The function on the other hand defines the behavior.43:09Are the type of activity for to do. This can be decapsulate and forwarded through your interface. Or this can be queuing.43:30You know, put this pack packet into a, a a given queue, right? So any behavior that customer can define and then that will be the function. And.43:46This is a optional, segment ID means if you don't want if you want purely your, your regular router configuration to take care of packet processing and then you don't need to use this. But obviously if the pa on the egress rou.44:01If you want to decapsulate, you need to use the function. Okay, next field is argument. While the function defines the behavior and type of action to take, argument defines the.44:22Value or quantification or quantity, how much or what specifically to do. E.g., if this says that, ok, put this packet into so much so Q and argument tells what should be the queue depth, right? And also, you know, customers are becoming very.44:39Creative and they are doing very advanced thing is the com with the combination of function and argument. Locator and functions are, 16 bit fields. Our argument is a variable.44:59Now that we talked about the locator and locators that can be used as addresses, right? We will take an example where.45:24A ingress router needs to send traffic to router two which whose address is 002. Router three whose node address is 003. Router four whose address is 004 but with a function of E001.45:35And router 5005 and E000.45:55So this is the instruction set that the router needs to inject into the packet and encapsulate it and put it into the SR domain. If you notice this here, the FCBB BBB.46:01Zero zero is the locator block is common across and that is typical because the customer has, in this case at least one large block of IPv six addresses. So SRV six has a efficiency mechanism wherein.46:21These four addresses is combined into one single address in 01:28 bits. FCBB BB zero zero, which is the common locator block and 002 that is router two.46:41You know one of the routers node ID or segment ID 003004 and E00. Here the segment ID as well as the function, right? 00400E001 and 0005E000.46:57The SID segment ID and function of the last router. So all these can be compressed in 1128 bit container. So this type of compression is called, this is.47:18Is it called, compressed SID listing and each one is a compressed SID. It is also called microsid. Though it appears a little bit a lot but.47:38It is what we are doing in a sense is we are compressing several instructions into one large instruction where the router knows how to deal with it. Alright, so IPv six routers.47:56Because they don't change the destination address, they need not make any changes to this header. They simply route this packet based on the most significant bit match, right? In their forwarding information base.48:12Okay, so we talked about compressed SRV six segment list encoding, so basically if if a router does not support the segment compression.48:30The packet looks like this. So this is the, the partner packet that is that that is sent into the are encapsulated into the SR SR domain SRV six domain.48:47This is the V six header. So here the destination address is the next hop address and the rest of addresses are put in SR header.49:07As a result of the compression, the routers that support compression, what they do is, they simply put all of this into a smaller 01:28 bit header in the one destination.49:17Address this compressed one. So the advantage here is we are saving overhead, right? And also, I will show you in the future that the forwarding also becomes efficient for the hardware ASICs.49:36Okay, let's talk about more of the segment locators. Okay, in SRV six domain, the SRV six locator.49:55Address, which is a kin or very similar to loopback address will be advertised into the SRV six domain using IGP. No brainer, NO brainer, right? So basically the locator block is advertised into SRV six.50:12So the SRV six capable routers understand that as locator and non SRV six and routers like regular IPv six routers consider that as a IPv six address block.50:31Simply advertised by router four here, right? If, this is a locator but the router can also for for this locator router here advertised locator block.50:47The locator node, like block plus the node address and different functions, E000E001. These functions are known to the only that specific router for only.51:03Okay, others forward this traffic based on the node right here, and then once the, once it is received by this router it, it treats the packet as per the functions. Locator allocation.51:21So a router can have a one or more locators. The reason to have more locators is if you have one, one typical example is that if you have a network, if you use network slicing in your network, where you have you.51:42Provide the.52:01Topology of the network into multiple imaginary topologies, you can use a specific locator for a specific topology so that the routing becomes easier. That is one common example. You can have other examples based on.52:02Your use case for examples you want to create multiple tunnels to this, but with a different.52:22Head end IP address, so you you can use the locators for that also. That is the locators allocation. You can have multiple locators. You can allocate one locator for, one imaginary topology and other one. E.g., you can.52:31Use one locator for low cost links or low cost topology having low cost links and other one for low delay topology.52:50Right, and like I said, the locators are advertised inside a, in the IGP, right? So in this case locator four locator block column four/forty eight is advertised into the IGP. That is how.53:09This locator is unknown to all other routers in the SR domain SRV six domain. Segment locator reachability. This is NO brainer obvious.53:24Forwarding basically locators are going to routing tables and forwarding information base and each router once in the control plane, this locator is advertised. And in this case 871 these routers.53:42When they when when they need to send traffic here, they use locator to route the traffic. All right. So.54:00How the, how the router, uses the micro seeds to process the packet. So this is the, this is a demonstration of whatever I just.54:18Spoke about but without a diagram wherein the ingress node receives the packet encapsulates with this header, right? And this header will look like this. There is a source address which is whatever is because this is the one which is encapsulating, so it will have its own source address and then the.54:36Destination address is this. How does it know it uses, it should use this destination address based on the constraint SPF computed on this router based on the database of ISIS or OSPF?54:56And then the policy. So this packet has to be treated with this policy, then to match that policy, whatever is the database.55:10It tells about the network. It computes that address and it puts it here. So user is need not know that the CSPF constrained SPF will do it for us. So the destination address is these six micro seats are infused into this destination address. Let's take a look.55:25At how.55:45And then just want to remind you that this destination address is a typical IPv 6128 bit address. So the packet journey wants router A in this case sends the traffic to this destination address.55:47If you see the most significant bit says 0001 here. I don't know if you can see, yeah, 0001. So, these are the transit nodes without SR.56:09They maybe SRV six capable also, but the packet is not destined to them, right? The destination destination of the packet is one because the actual 001 and its address is 001. So the packet will be.56:23Will be processed by this. The process of these routers will be simply to forward, but this is, this does a little bit more than the forwarding. What it does is from this address, it will remove.56:380001, and it shift these rest of the address towards the left, paired the address with zeros and forward it based on the newly constructed destination address. Okay, so again, I will.56:54Again repeat that taking the example of R3R3 received a packet with 003, right? So it, it belongs to me as R three. It will do a shift on forward operation.57:14Means it will take away 003, shift the rest of the address to the left, add zeros to the end and put and make the destination address. As you can see here.57:31The 004 is shifted left, and then more zeros added. That will be the destination address, and then it will forward the packet based on the newly constructed destination address. Alright.57:46I want to conclude this with the reminder of the evolution from legacy MPLS to SRV six. In the legacy MPLS, we use MPLS labels. In SRV six, there are NO labels. There are.58:03IPv six header and compressed segment IDs, which will become part of the destination address. So a simple IPv six router can.58:22Forward the traffic. A simple host with IPv six capable host can inject the IPv six, but it requires NFV based hosts to have some specific capabilities to do it, but.58:37But the bottom line is that the header is IPv six, not label. Transport protocols in MPLS legacy, we, we need IGP LDP RSVP, BGPLU.58:53Here we simply need ISIS or OSPF, right? The service services L two VPN or L three VPN services use LDP and MPBGP in legacy.59:09Whereas we use the SRV six users simply MPBGP, NO need of LDP. With that I conclude my presentation. I would love to answer any questions now or even later, but then with that I want to hand over to.59:25Most. Hey, thank you nice presentation. So with that I will start sharing the lab demo part.59:45Let's see here so thank you everyone for staying with us. This part should be another 20 min so we will go over SR.01:00:00RV6 exactly what described, the.01:00:20The, you know, the very basic fundamentals of SRV six, how it's configured in ISIS, what are all the config that's that are needed and with that of course we understand that all the nodes need to be IPv six capable. The.01:00:24Course and the destination have to be SRV six capable, so this is where our configurations will be and also later on in the next series, we will have traffic engineering which will also show us that anywhere where a traffic engineering decision is made, those nodes.01:00:44Even if they're in the transit node, they need to be SRV six capable also.01:01:04So, so this is the, this is the big topology. So this was, this is what we will follow for our, throughout this series. But in this session, we are mostly focused on this R5R6R1.01:01:10And our three part of the L one ISIS, we will continue on this topology for future presentations where we need to go end to end or where we need to go deploy services or traffic engineering. But for this session just we are looking.01:01:29Into this very simple network.01:01:49Right here so back to the configurations so basically I I will you know break it down into five main for the basic five main configurations, so for the loopback zero on the SRV six capable nodes, we need.01:01:53To address we need to identify an IP address, which will be the global IP address. So every IPv six interface gets a link local address, but for the loopback zero, we will need to get give it assign it an IPv six address which will be the which will identify it with the locator and.01:02:12For the whole network.01:02:33So like you see here under IP interface loopback zero, we have given it the FCBB BB zero zero, and this is for node five. So this is the five part of it, on 01:28. So this is how this loopback one loopback zero will be.01:02:35And it would be routed across the network. Then we need to make sure that the all the ISIS interfaces that are neighbors, they need to be IPv six enabled and that's all we have to do. We do not have to assign an IP V six address. It will get.01:02:54By the link local and it will be enough for this.01:03:15Routing. Next thing for the router ISIS, the in this case the router ISIS core, so we need to make sure that they are IPv six like here you see loopback zero is one of the interface and then gig zero zero is also another interface and then.01:03:20We make sure their IPv six unicast. So everything we are on the whole network is IPv six capable.01:03:40And then again this will make this router segment SRV six segment routing. We we go under segment routing and under segment routing we have SR MPLS how it was explained to you earlier, NSRV six, so we will pick SRV six and we.01:03:49Will make sure that the encapsulation with the source address of.01:04:08What we assigned here for the loopback zero. So double loopback zero will be the source address for this router SRV six. Then the locator needs to be assigned so we will, then this locator name can be anything, right? So in this case we are keeping keeping this as main.01:04:12And then the micro segment behavior we will give it a PSP USD PSPs are where it's you know the the last and the last node takes the label out.01:04:32An ultimate hop, that kind of scenario and USD is the ultimate source and destination. So again the prefix in this case for the locator will be the, the loopback zero. And then finally for we gonna need to make sure that the.01:04:51ISIS is also SRV six, so we will make this bit here we just made it IPv six, but now we will make the ISIS core into a segment routing SRV six with the locator name. So all the ISIS in this domain will, will communicate with.01:05:08Each other with main as the locator.01:05:27Then we have some verifications. This is what we will go over. These are the, I just wanted to show you so it's very clear on the router it may not be very visible, but these are all the verifications that we will be doing. We will just check the interface loopback zero, the IPV six and then.01:05:34Here is.01:05:54Mostly the segment routing part of it, how is a manager, what locator it shows all the said and again the the ISIS database. This is also important so it we will config, we will make sure that our router is learning all the SRV six destinations and the next hop.01:05:55Routes. So with that I will stop this session, this one and bring on the, the lab part of it.01:06:15Okay, so we are in XR five, the router five, and all the configurations are already there so if you look back zero.01:06:48Make sure that it is the locator block is FCBB and this is the node five and then so if you want to quickly if I want to show you like go to router six. Okay.01:07:05Make it visible Okay.01:07:21So it's the same locator FCBB BBB zero zero. The load node name is six. Just wanted to show you how each node will have the loopback address so show IPV six interface loopback zero here, and you can see that it's.01:07:43It's ah IPv six is enabled, local link addresses here which it gets automatically and then the global unicast address is this one right here. So this is what we will route across. Okay, let me give me a minute here.01:08:05Okay. So then, I mean we don't have to see the interface with zero zero. So again, the interface will show.01:08:30So the it's IPv six enabled and the local link address is is automatically there and it's it's up with the worth is default, meaning it's not part of any any worth is with the default worth.01:08:50So we make sure that the ISIS is neighbors are established, so ISIS neighbor and yes, so we are connected with gig zero zero to XR O six and it's it's in the upstate, so.01:09:09The point to point, everything looks good, this basic ISIS.01:09:28Stuff and then so now we'll check the segment routing section of it. So show segment routing and then I just want to show you like if I do a question mark after that, you will see these are the options we have for segment routing.01:09:33One thing we can check SRV six adjacency said mapping server etc. So we'll check SRV six.01:09:52And manager 1st. So here you see it's SRV six enabled it checks the SRV six status mode basically checks the US ID versus classical, so here.01:09:59Here we have micro segment said block is this one. So it's micro segment and then the encapsulation is configured and this is the default, so loopback zero would have become the default, but we configured it with loopback zero. So this is all the information which is important to us.01:10:19Us to confirm is.01:10:40Make sure it's SRV six enabled, then the operational mode is micro segment here. So, so what does the, just to go a little bit more details so basically when it's SRV six enabled, that means that the SRV six control plane.01:10:42And data plane forwarding infrastructure are globally enabled. Segment routing SRV six is config configured basically SRV six across data plane.01:11:02And control plane. Then what is the significance? It confirms the routers and BU pipeline is programmed to process SRV six headers and execute local SID functions. So this is what is needed for the.01:11:13The ingress and the egress nodes for SRV six and also the traffic engineering. So that's how we confirmed these are SRV six enabled. Then the operational mode is micro segment. This route router is operating in SRV six micro SID mode rather than the classical 01:28 bit SRV six.01:11:30So we learn about this micro SIDs, same thing we are confirming here. And then the SID block SID base block is the FCBB.01:11:50The FCBB 0024, that's your that's your the basic block. This is the route you block allocation across the domain and then the rest is the P, you know, the node information and the following stuff.01:11:59So it gives us more information too, so if you go more, it will give us the endpoint behaviors, these are all the endpoint behaviors that are, that are, it's capable of TI LFA.01:12:18It's capable of SRV six micro loop avoidance yes, so this is all the features that Gopal was saying that SRV six has, it's capable of doing all that and then gives us some times on how long the default parameters are.01:12:34For the whole times et cetera. The next we will check is a locator. So here it's like we assigned right so we said that this is part of the main locator named main and the router will assign it.01:12:53ID by itself. We don't have to assign in an ID if we had many different.01:13:14Locators, there will be different IDs in this case only one. Currently we are not doing any flex algo, so it's basically the zero is the algo. The flex algo ranges from 01:28 to 02:55. So when we do assign for traffic engineering of any flex algo.01:13:18Then it will be, shown here under the SRV six locator. Then again the prefix status and flag. So the flag part.01:13:38Is the microsid, ok. And then the flag is up, ok. So or microsite sorry yeah the flag is microsid. So this is what all information is gathered from here.01:13:52If you have any questions, feel free, then next we will check SID all. So in this case again, since it's a very basic configuration, we have the SID so behavior UN and UA, UN is the node itself.01:14:07So this is the final destination when.01:14:27Any traffic, any packet comes on the network with the this IPV six address. This node will take it. This will be the final destination and the next this one if it's going to another function, then it's shift and.01:14:31The process this UA means it will shift the packet and move on. So these are the two behaviors of what when the traffic is received, what it's going to do.01:14:50So, and again PSP and USD are the two behavior there. The LPSP is a penaltimate segment pop, so it will be it's capable like in label MPLS the last node will be able will take the.01:15:01Label out before forwarding to the or sorry, second last node will take the label out of the label out. So that's PSP and then the ultimate segment decapsulation is where the last node.01:15:21Will decapsulate the packet so it's capable of doing that. So this is mostly all of the I will also show.01:15:34So here we can see how we've already went through it. Segment routing was SRV six, source address, locators was assigned here and then the prefix assigned here. So this is all the segment routing configurations that's needed.01:15:50And then for ISIS core level one, ok, then segment routing SRV six, and then the two loopback zero and then the next, next, next, next hop or the neighbor. So these are all the configurations that are absolutely needed.01:16:09Next we can check show route IPV six. So again we can see that it all the routes are learned are IPV six and it has.01:16:29This one the next loopback address where it's connected, it will have a way to go get there by 000, and it will also show SRV six endpoint.01:16:43UA shift or endpoint UA. So these are the under the IPv six routes we will see all all all this information helpful to understand the routing tables and in case there is any troubleshooting needed, if there's anything missing.01:16:59But in this case everything looks good. Next, we will check the ISIS database, and this will be what's.01:17:19So, this database basically shows that how would the gives us the isis information, host name, I will be six unicast, but also.01:18:01The SR algorithm, what it's using and SR locator again here with algorithm zero, and then SID structure gives us that the blocked length is 32, node length is 16, there is NO function, there is NO arguments.01:18:17Here the length is 80 and then.01:18:37Okay, so that's it. Yeah at this point also the N.X SID for, the, this address basically it it assigns across like advertises the n.x, so.01:18:42Behavior for X Connect to an interface. So if this traffic was coming across an xconnect, it will have have assigned this address. So this is all things we need to verify, for the on the on the larger scale.01:19:00And I think that's about it for me, did not did not have much configurations or much to show here because this is a very simple network we wanted to cover.01:19:17And then in the future we're gonna have the more complicated like how do we configure L2L3 services? How do we do traffic engineering and of course the verifications and the configurations will be a little bit more complex.01:19:32So with that, if there is any questions, please go ahead and ask them and if not I can also do you have any ping or trace route in your list of So yeah. So basically from here.01:19:49Yeah, I forgot that part. It's here. So if you wanted to ping, so this is the like one oh seven this is the ping address but we can also ping the loopback, so.01:20:06Thing like the from here, he wanted to ping this guy here. So.01:20:23And then if we can trace route.01:20:43IPV six. So trace route IPv six will be the only option where the versus like MPLS or SRMPLS, we could trace route based on the MPLS or or the label. But in this case because IPv six is the control plane, right? So we will just do.01:20:48IPV six.01:21:09And it's the right next node so it it shows you right here that is one half away yeah so yeah I think this is this pretty much concludes there is anything else I I can add here.01:21:10Yeah, I just want to ask if there are any questions before we conclude the session. All right, I just want to remind you that we also have two other sessions in this series. One is.01:21:30Features and services, and the last one is migration and implementation. If you have any feedback send us now in the chat or raise your hand or.01:21:49I don't know if there is a mechanism to send the feedback. There is a question from how does an SRV six router distinguish between A?01:22:05Micro seed node versus a function. Okay, so I did not cover this in the presentation, but it's a good question. Basically because let me yeah basically because I want, we want to keep this at a very high level conceptual level.01:22:20Not going drilling too much into it, but just to answer your question.01:22:40Just like in segment routing, there is a, address allocation of labels. Here also for the nodes and functions, there is an allocation. The node addresses start from 000.01:22:45The one and go all the way to DFFF.01:23:05Okay, that is Delta Frank Frank Frank. Oh, NO NO NO. Yeah dfff. Under the function addresses start from E000 to FF FF.01:23:10R minus one, whatever, but the the distinction is based on the addressing of the microsite. Thank you, Seth.01:23:26All right. Any other last words for me? Nope. I think this is it. It's, thanks everyone for staying with us and I I think you might get if we can improve this session for anyone you might get a survey let us know, please.01:23:43If this was good or if we can improve, that will help us to ourselves. Awesome. Thank you very much. You guys have a wonderful day. Bye.01:24:03