RFC 8430
RIB Information Model
RFC 8430: RIB Information Model 🎀
Okay so bestie, this one is the RIB Information Model 💅 It's Informational (not a Standards Track spec), it came out in September 2018, and it's by N. Bahadur (Ed.) at Uber, S. Kini (Ed.) at Uber's neighbour in the author list, Hmm, let me stop. Correction, real talk: the authors are N. Bahadur, Ed. (Uber), S. Kini, Ed. (Uber), and J. Medved (Cisco). It's published under ISSN 2070-1721. It doesn't obsolete or update anything 🦋 and it's the "record the design decisions" kind of doc that other RIB builders can learn from ✨
The parts I'm rewriting in this first chunk are the front matter, the Abstract, Status, Copyright, the Table of Contents, Section 1 (Introduction), and Section 1.1 (Conventions). The rest comes in the next parts 💖
Abstract ✨
Okay so routing in enterprise and carrier networks is usually handled by network devices like routers and switches using a Routing Information Base (RIB) 🌸 Protocols and configs shove data into the RIB, and then the RIB manager installs state into the hardware so packets actually get forwarded 🚀
This doc specifies an information model for the RIB, which is basically the blueprint you need before you can make a standardized data model 🧚 The IETF's I2RS working group used it to design the I2RS RIB data model. It's being published so that other RIB developers can benefit from the higher-level design concepts 💜
Status of This Memo 🌈
This is NOT an Internet Standards Track spec. It's published for informational purposes only 🫶
It's still an IETF product though, so it reflects the consensus of the IETF community, got public review, and was approved by the Internet Engineering Steering Group (IESG). Not every IESG-approved doc is a candidate for any level of Internet Standard, though. See Section 2 of RFC 7841 for that 💅
You can find the current status, errata, and how to send feedback at https://www.rfc-editor.org/info/rfc8430 🔥
Copyright Notice 🌸
Copyright (c) 2018 IETF Trust and the persons identified as the document authors. All rights reserved. 💖
This doc is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication. Please read these carefully, because they spell out your rights and restrictions. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions, and they're provided without warranty as described in the Simplified BSD License 🎀
Table of Contents 🦋
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1. Conventions Used in This Document . . . . . . . . . . . . 6
2. RIB Data . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1. RIB Definition . . . . . . . . . . . . . . . . . . . . . 7
2.2. Routing Instance . . . . . . . . . . . . . . . . . . . . 7
2.3. Route . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4. Nexthop . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4.1. Base Nexthops . . . . . . . . . . . . . . . . . . . . 12
2.4.2. Derived Nexthops . . . . . . . . . . . . . . . . . . 14
2.4.3. Nexthop Indirection . . . . . . . . . . . . . . . . . 15
3. Reading from the RIB . . . . . . . . . . . . . . . . . . . . 16
4. Writing to the RIB . . . . . . . . . . . . . . . . . . . . . 16
5. Notifications . . . . . . . . . . . . . . . . . . . . . . . . 17
6. RIB Grammar . . . . . . . . . . . . . . . . . . . . . . . . . 17
6.1. Nexthop Grammar Explained . . . . . . . . . . . . . . . . 20
7. Using the RIB Grammar . . . . . . . . . . . . . . . . . . . . 20
7.1. Using Route Preference . . . . . . . . . . . . . . . . . 20
7.2. Using Different Nexthop Types . . . . . . . . . . . . . . 20
7.2.1. Tunnel Nexthops . . . . . . . . . . . . . . . . . . . 21
7.2.2. Replication Lists . . . . . . . . . . . . . . . . . . 21
7.2.3. Weighted Lists . . . . . . . . . . . . . . . . . . . 21
7.2.4. Protection . . . . . . . . . . . . . . . . . . . . . 22
7.2.5. Nexthop Chains . . . . . . . . . . . . . . . . . . . 22
7.2.6. Lists of Lists . . . . . . . . . . . . . . . . . . . 23
7.3. Performing Multicast . . . . . . . . . . . . . . . . . . 24
8. RIB Operations at Scale . . . . . . . . . . . . . . . . . . . 25
8.1. RIB Reads . . . . . . . . . . . . . . . . . . . . . . . . 25
8.2. RIB Writes . . . . . . . . . . . . . . . . . . . . . . . 25
8.3. RIB Events and Notifications . . . . . . . . . . . . . . 25
9. Security Considerations . . . . . . . . . . . . . . . . . . . 25
10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 26
11. References . . . . . . . . . . . . . . . . . . . . . . . . . 26
11.1. Normative References . . . . . . . . . . . . . . . . . . 26
11.2. Informative References . . . . . . . . . . . . . . . . . 27
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . 28
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 281. Introduction 🌸
Okay so routing in enterprise and carrier networks is traditionally handled in network devices. Routers usually run routing protocols, and those protocols (plus any static config) fill up the router's Routing Information Base (RIB) 💜 The RIB manager runs the RIB, and it gives its clients (the routing protocols) a northbound interface so they can insert routes into the RIB ✨ The RIB manager checks the RIB and decides how to program the Forwarding Information Base (FIB) of the hardware, by talking to the FIB manager 🧚 Here's how all those pieces fit together in Figure 1 🦋
+-------------+ +-------------+
|RIB Client 1 | ...... |RIB Client N |
+-------------+ +-------------+
^ ^
| |
+----------------------+
|
V
+---------------------+
| RIB Manager |
| |
| +--------+ |
| | RIB(s) | |
| +--------+ |
+---------------------+
^
|
+---------------------------------+
| |
V V
+----------------+ +----------------+
| FIB Manager 1 | | FIB Manager M |
| +--------+ | .......... | +--------+ |
| | FIB(s) | | | | FIB(s) | |
| +--------+ | | +--------+ |
+----------------+ +----------------+
Figure 1: RIB Manager, RIB Clients, and FIB ManagersRouting protocols are inherently distributed, and every router makes its own independent decision based on the routing data it gets from its peers 🌈 Newer deployment paradigms and specialised apps mean there's a growing need to guide the router's routing function [RFC7920] 💅 The traditional protocol-based way of filling the RIB is totally fine for most cases where distributed control is used 👑 But operators are also using static routes, policies, and RIB import/export rules on routers to cover some use cases 🎀 There's also a growing list of situations where an operator might want to program the RIB with data that isn't just about routing, within their own domain 😭 That data could be about routing in the adjacent domain, or the load on storage and compute 🫶 Or it could just be a programmatic way to set up on-demand dynamic overlays, like GRE tunnels, between compute hosts, without making those hosts run traditional routing protocols 🍓 A standardized, publicly documented programmatic interface to a RIB would unlock even more networking apps for all sorts of use cases [RFC7920] ✨
A programmatic interface to the RIB involves two kinds of operations: reading from the RIB, and writing to it (adding, modifying, or deleting) 💖
To see what's in a router's RIB, people currently use things like per-protocol SNMP MIBs and screen scraping 😭 Those don't scale, because they're client pull mechanisms rather than proactive push from the router 💀 Screen scraping is error prone (the output format can change) and it's vendor dependent 🙈 Building a RIB from per-protocol MIBs is also error prone, since MIB data represents protocol data and not the exact information that went into the RIB 🔥 So honestly, just getting read-only RIB info from a router is a hard task 💅
Adding content to the RIB from a RIB client can be done today with static config mechanisms that router vendors provide 🌸 But what you can change in the RIB varies from vendor to vendor, and so does how you configure it 👑 That makes it really hard for a RIB client to program a multi-vendor network in a consistent, vendor-independent way 🥺
So the whole point of this doc is to specify an information model for the RIB 🦋 With the information model, you can build a detailed data model, and a RIB client could then use that data model to program a network device ✨ One data model based on this doc is the I2RS RIB data model [RFC8431] 💜
Here's the rest of the doc's layout 🧚 Section 2 gets into what makes up a RIB and what can be programmed in it 💖 Sections 3 and 4 give guidelines for reading and writing the RIB, respectively 🎀 Section 5 gives a high-level view of the events and notifications going from a network device to a RIB client, to update the client on asynchronous events 🌈 The RIB grammar is in Section 6 🫶 Examples of using the grammar are in Section 7 🔥 And Section 8 covers what to think about when doing RIB operations at scale 💅
1.1. Conventions Used in This Document 🌟
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here 🔐 Translation: the capital-letter key words are the real rules, and lowercase ones are just regular words 💜
Part 1 of 6 is done! 🌈 More parts are coming, so stay tuned, bestie 💖
2. RIB Data ✨
okay so this section is the RIB deep dive 🧠💖 it's got forward references to the RIB grammar (see Section 6) and a big-picture drawing in Figure 2. quick heads up: the drawing only shows ONE routing instance, ONE RIB, and ONE route, because ASCII art can only handle so much, bestie 🙈 Sections 3 and 4 are where the high-level read and write operations live.
network-device
|
| 0..N
|
routing instance(s)
| |
| |
0..N | | 0..N
| |
interface(s) RIB(s)
|
|
| 0..N
|
route(s)
Figure 2: RIB Information Model2.1. RIB Definition
okay the RIB is an entity that holds routes 🗂️🌸 it gets identified by its name and lives inside a routing instance (see Section 2.2). a network device MAY have routing instances, and each routing instance MAY have RIBs. the name MUST be unique inside its routing instance. every route in a given RIB MUST be the same address family (like IPv4), and each RIB MUST belong to a routing instance. no exceptions, main character rules 👑
a routing instance can hold two or more RIBs of the same address family (like IPv6) 💅 the classic use case is multi-topology routing [RFC4915] [RFC5120] ✨
each RIB MAY be tagged with an ENABLE_IP_RPF_CHECK attribute, which turns on Reverse Path Forwarding (RPF) checks for all the IP routes in that RIB 🛡️ the whole point is to stop spoofing and cut down on malicious traffic. here's how it works: for an IP packet, the IP source address gets looked up, and the RPF interface(s) tied to the route for that source address are found. if the packet came in on one of those RPF interfaces, it gets forwarded based on its IP destination address. if not, it gets dropped 🚫 bye bye, sus packet 💀
2.2. Routing Instance
a routing instance is a collection of RIBs, interfaces, and routing parameters 🧚 it basically creates a logical slice of the router, and it lets different logical slices across a bunch of routers talk to each other. Layer 3 VPNs, Layer 2 VPNs (L2VPNs), and Virtual Private LAN Service (VPLS) can all be modeled as routing instances ✨ BUT, lowkey important: modeling an L2VPN this way only covers the Layer 3 (RIB) side. it doesn't model any Layer 2 info (like ARP) that might go along with the L2VPN, so don't get confused 🙃
the interface set says which interfaces belong to this routing instance. the RIBs say how incoming traffic gets forwarded, and the routing parameters control what info goes into those RIBs 🎀 now here's the tea: the interface sets of two routing instances MUST have an intersection of the null set. in other words, an interface MUST NOT show up in two routing instances. no sharing interfaces, bestie 🫶 so a routing instance is basically the routing info and parameters across its own set of interfaces.
a routing instance MUST contain these mandatory fields:
INSTANCE_NAME: a routing instance is identified by its name,INSTANCE_NAME. this MUST be unique across all routing instances on a given network device 🏷️rib-list: the list of RIBs attached to this routing instance. each routing instance can have multiple RIBs for different kinds of routes 🌈 like, you might put IPv4 routes in one RIB and MPLS routes in another. the list of RIBs can be an empty list, which is valid 🍓
a routing instance MAY contain these optional fields:
interface-list: the list of interfaces attached to this routing instance. this helps constrain the boundaries of packet forwarding 🚧 packets arriving on these interfaces are directly associated with the given routing instance. the list holds identifiers, and each identifier uniquely identifies one interface.ROUTER_ID: this field identifies the network device in control plane conversations with other network devices 🛰️ it's meant for when you want to virtualize one physical router into several virtual routers. each virtual router MUST have a uniqueROUTER_ID, and aROUTER_IDMUST be unique across all network devices in a given domain. unique everywhere, periodt 💜
a routing instance may be made purely for packet processing, with no interfaces attached 🦋 for example, a packet arriving in routing instance A might have a nexthop of routing instance B, and after processing in B, the nexthop might be routing instance C. so routing instance B isn't tied to any interface. and since it doesn't do any control-plane talking with other network devices, it doesn't need a ROUTER_ID either 🧠✨
2.3. Route
a route is basically a match condition plus an action that follows the match 💖 the match condition says what kind of route it is (IPv4, MPLS, etc.) and which fields to match on. Figure 3 shows the overall contents of a route. quick note: for the sake of ASCII art, only a single instance of the route-attribute, match flags, and nexthop is drawn 🎨
route
| | |
+---------+ | +----------+
| | |
0..N | | |
route-attribute match nexthop
|
|
+-------+-------+-------+--------+
| | | | |
| | | | |
IPv4 IPv6 MPLS MAC Interface
Figure 3: Route Modelthis document specifies these match types:
- IPv4: match on destination and/or source IP address in the IPv4 header 🌐
- IPv6: match on destination and/or source IP address in the IPv6 header 🌍
- MPLS: match on an MPLS label at the top of the MPLS label stack 🏷️
- MAC: match on Media Access Control (MAC) destination addresses in the Ethernet header 🔌
- Interface: match on the incoming interface of the packet 🚪
a route MAY be matched on one or more of these match types by policy, either as an "AND" (to restrict the number of routes) or an "OR" (to combine two filters) ✨ it's giving flexible 💅
each route MUST have these mandatory route-attributes:
- ROUTE_PREFERENCE: a numerical value that lets you compare routes from different protocols 🥇 static configuration counts as a protocol for this field too. it's also known as "administrative distance". the lower the value, the higher the preference. for example, there can be an OSPF route for 192.0.2.1/32 (or IPv6 2001:DB8::1/128) with a preference of 5. if a controller programs a route for 192.0.2.1/32 (or IPv6 2001:DB8::1/128) with a preference of 2, then the controller's route is preferred by the RIB manager 👑 preference should be used to dictate behavior. for more examples of preference, see Section 7.1 🎀
each route can also have one or more optional route-attributes:
- route-vendor-attributes: vendors can specify vendor-specific attributes using this 🛍️ the details of this attribute are outside the scope of this document.
each route has a nexthop associated with it. nexthops are described in Section 2.4 🦋
last thing, and this one's a bit of a tease 😭: multicast matching was considered (like using the TTL of a packet to limit the range of a multicast group), but it wasn't added to this information model. future RIB information models should look into those multicast features 🔮 to be continued, bestie 💜
2.4. Nexthop ✨
okay so a nexthop is basically the answer to "where does this packet go next?" 💅 It's the object you get back from a route lookup. Like, if the lookup says "send it out of this interface," then the nexthop is that interface 🦋
Nexthops come in two flavours: fully resolved or unresolved 💜 A resolved nexthop has enough info to send the packet on its way by forwarding it out an interface to a neighbor that's directly connected. Like, a nexthop pointing at a point-to-point interface, or at an IP address on an Ethernet interface, is already resolved 🌸
An unresolved nexthop is still waiting on the RIB manager to figure out the final resolved nexthop 🫠 For example, a nexthop could be just an IP address. The RIB manager has to work out how to reach it. Is it reachable by regular IP forwarding? By an MPLS tunnel? By both?? 👀 If the RIB manager can't resolve it, the nexthop stays unresolved and is NOT a candidate for installation in the FIB. Future RIB events can totally flip it to resolved, like when an IGP neighbor advertises that IP address. But it can also go the other way. A resolved nexthop can go back to unresolved, like when a tunnel goes down 💀 At that point it's no longer a candidate to be installed in the FIB either.
Now here's the vibe check ✨ When at least one of a route's nexthops is resolved, the route can actually be used to forward packets. That route is eligible to be installed in the FIB, and we call it a FIB-eligible route 🎀 If all of a route's nexthops are unresolved, the route can't forward anything, so it's ineligible for the FIB, and we call that a FIB-ineligible route 🥀 The RIB information model lets a RIB client program routes whose nexthops might start out unresolved. And whenever an unresolved nexthop gets resolved, the RIB manager sends a notification about it (see Section 5) 🔔
Here's the overall structure and usage of a nexthop. Quick heads up: to keep the ASCII art readable, only one instance of each component is drawn in Figure 4 🧚
route
|
| 0..N
|
nexthop <-------------------------------+
| |
+-------+----------------------------+-------------+ |
| | | | | |
| | | | | |
base load-balance protection replicate chain |
| | | | | |
| |2..N |2..N |2..N |1..N |
| | | | | |
| | V | | |
| +------------->+<------------+-------------+ |
| | |
| +-------------------------------------+
|
+-------------------+
|
|
|
|
+---------------+--------+--------+--------------+----------+
| | | | |
| | | | |
nexthop-id egress-interface ip-address logical-tunnel |
|
|
+--------------------------------------+
|
+----------------------+------------------+-------------+
| | | |
| | | |
tunnel-encapsulation tunnel-decapsulation rib-name special-nexthop
Figure 4: Nexthop ModelThis spec is giving very generic, extensible, and recursive grammar for nexthops 🌈 A nexthop can be a base nexthop or a derived nexthop. Section 2.4.1 covers base nexthops, and Section 2.4.2 explains the different kinds of derived nexthops. There are some special nexthops too, and those are in Section 2.4.1.1. Last but not least, Section 2.4.3 dives into nexthop indirection and how it's used 🧠 And if you want examples of when and how to use tunnel nexthops and derived nexthops, check out Section 7.2 🍓
2.4.1. Base Nexthops
okay so at the lowest level, a nexthop can be one of these 🍬
- Identifier: This is an identifier the network device hands back to represent a nexthop. You can reuse it when you're programming derived nexthops, which is so efficient 🧚
- Interface nexthops: These point to an interface 🔌 The attributes are:
- Egress-interface: This represents a physical, logical, or virtual interface on the network device. Address resolution must not be required on this interface. This interface can belong to any routing instance 🌸
- IP address: A route lookup on this IP address gets done to figure out the egress-interface. Address resolution might be needed, depending on the interface 🤷♀️
- An optional rib-name can also be specified to say which RIB the IP address should be looked up in. You can use the rib-name field to send a packet from one domain into another domain 🚪 By default, the RIB is the same one the route belongs to.
These attributes can be combined like this:
- Egress-interface and IP address: This works when, for example, the IP address is a link-local address 🔗
- Egress-interface and MAC address: The egress-interface has to be an Ethernet interface. Address resolution is not required for this nexthop ✨
- Tunnel nexthops: These point to a tunnel 🚇 The types of tunnel nexthops are:
- tunnel-encapsulation: This can be an encapsulation for an IP tunnel, an MPLS tunnel, or others defined in this document 📦 You can chain an optional egress-interface to the tunnel-encapsulation to say which interface the packet goes out on. The egress-interface comes in clutch when the network device has Ethernet interfaces and you need address resolution for the IP packet 💅
- tunnel-decapsulation: This is for decapsulating a tunnel header 🎁 After decapsulation, you can do another lookup on the packet by chaining it with another nexthop. Or the packet can go straight out an egress-interface 🏃♀️
- logical-tunnel: This can be an MPLS Label Switched Path (LSP) or a GRE tunnel (or others defined in this document). It's represented by a unique identifier, like a name 🏷️
- rib-name: A nexthop pointing to a RIB 📍 This means the route lookup needs to keep going in the specified RIB. It's a way to do chained lookups 🔗
Tunnel nexthops let a RIB client program static tunnel headers 🎀 Sometimes the remote tunnel endpoint doesn't support dynamic signaling (like, no LDP support on a host). In those cases, the RIB client might want to program the tunnel header on both ends of the tunnel 🫶 The tunnel nexthop is kept generic, with specs for some commonly used tunnels. The data model is expected to model these tunnel types with complete accuracy 💯
2.4.1.1. Special Nexthops
Special nexthops do specific, well-defined jobs (like DISCARD) 🗑️ Here's what each one does:
- DISCARD: The network device should drop the packet and increment a drop counter 💀
- DISCARD_WITH_ERROR: The network device should drop the packet, increment a drop counter, and send back an appropriate error message (like an ICMP error) 🙅♀️
- RECEIVE: The traffic is destined for the network device itself, like protocol packets or Operations, Administration, and Maintenance (OAM) packets 📬 All locally destined traffic SHOULD be throttled, so nobody can launch a denial-of-service attack on the router's control plane 🛡️ You can specify an optional rate limiter to say how to throttle the traffic headed to the control plane. The description of the rate limiter is outside the scope of this document 🤐
2.4.2. Derived Nexthops
okay so derived nexthops can be any of these 🌈
- weighted lists, which are used for load-balancing ⚖️;
- preference lists, which are used for protection using primary and backup 🛟;
- replication lists, which are lists of nexthops to replicate a packet to 👯;
- nexthop chains, which are for chaining multiple operations or attaching multiple headers; or 🔗
- lists of lists, which are a recursive application of the above 🪆
Nexthop chains (see Section 7.2.5 for usage) let you do multiple operations on a packet by logically combining them ✨ For example, you could chain together "decapsulate MPLS header" and "send it out a specific egress-interface" 🫧 Chains can also specify multiple headers over a packet before it gets forwarded. One simple example is MPLS over GRE, where the packet has an inner MPLS header, then a GRE header, then an IP header 🎀 The outermost IP header is decided by the network device, but the MPLS header or GRE header is specified by the controller 🕹️ Not every network device can support all kinds of nexthop chains, or an arbitrary number of headers chained together 😭 So the RIB data model SHOULD provide a way to expose the nexthop chaining capability a given network device supports 💜
It's expected that all network devices will have a limit on how many levels of lookup they can do, and not all hardware will support all kinds of nexthops 🧱 That's why RIB capability negotiation is SO important 🙌 A RIB data model MUST specify a way for a RIB client to learn about the network device's capabilities 👑
2.4.2.1. Nexthop List Attributes 💅
okay so when a nexthop is a list (or a list of lists), every member can have its own attributes that say what job it's doing 🧚 There are two of them, and they're both kinda iconic:
NEXTHOP_PREFERENCE: this one is for protection schemes 🛡️ It's an integer between 1 and 99, and lower = higher preference (so the lowest number is the main character 👑). To download a primary/standby pair to the FIB, the nexthops that are resolved and have the two highest preferences get selected. Each<NEXTHOP_PREFERENCE>should have a unique value within a<nexthop-protection>(see Section 6).NEXTHOP_LB_WEIGHT: this one is for load-balancing ⚖️ Each list member MUST be assigned a weight between 1 and 99. The weight decides how much traffic goes over a nexthop that's used for forwarding. It's a ratio: this nexthop's weight divided by the weights of all the nexthops of this route that are used for forwarding. For equal load-balancing, one MAY specify a weight of "0" for all the member nexthops 🍓 The value "0" is reserved for equal load-balancing and, if applied, MUST be applied to all member nexthops. Also, a weight of 0 is special because of historical reasons, so it's giving legacy energy 🕰️
2.4.3. Nexthop Indirection 🦋
okay this one is so clever. Nexthops can be identified by an identifier, which creates a layer of indirection ✨ The identifier is set by the RIB manager and handed back to the RIB client when it's asked for.
A classic example is a nexthop that points to another network device (like a BGP peer). The returned nexthop identifier can then be used to program routes to point to this nexthop. Now, once the RIB manager has set up that indirection with the nexthop identifier, if the transport path to the network device (BGP peer) changes, the RIB client doesn't even notice 🫶 All the routes that point to that network device just automatically start going over the new transport path. Literally seamless. Indirection with identifiers isn't just for unicast nexthops either, it also works for nexthops that contain chains and nested nexthops. See Section 2.4.2 for examples.
3. Reading from the RIB 📖
A RIB data model MUST allow a RIB client to read entries for RIBs created by that entity. The network device administrator MAY allow reading of other RIBs by a RIB client through access lists on the network device. The details of access lists are outside the scope of this document, so we're not getting into that here 🙈
The data model MUST support a full read of the RIB and then subsequent incremental reads of changes to the RIB. When sending data to a RIB client, the RIB manager SHOULD try to send all dependencies of an object before sending that object 💜
4. Writing to the RIB ✍️
A RIB data model MUST allow a RIB client to write entries for RIBs created by that entity. The network device administrator MAY allow writes to other RIBs by a RIB client through access lists on the network device. Again, the details of access lists are outside the scope of this document 🌸
When writing an object to a RIB, the RIB client SHOULD try to write all dependencies of the object before sending that object. The data model SHOULD support requesting identifiers for nexthops and collecting the identifiers back in the response 🎀
Route programming in the RIB MUST result in a return code that contains the following attributes:
- Installed: Yes/No (indicates whether the route got installed in the FIB)
- Active: Yes/No (indicates whether a route is fully resolved and is a candidate for selection)
- Reason: E.g., "Not authorized"
The data model MUST specify which objects can be modified. An object that can be modified is one whose contents can be changed without having to change objects that depend on it and without affecting any data forwarding. To change a non-modifiable object, you'll need to create a new object and delete the old one 😭 For example, routes that use a nexthop identified by a nexthop identifier should be unaffected when the contents of that nexthop change. Main character energy for the routes, honestly 👑
5. Notifications 🔔
Asynchronous notifications are sent by the network device's RIB manager to a RIB client when some event happens on the network device. A RIB data model MUST support sending asynchronous notifications. Here's a brief list of suggested notifications:
- Route change notification (with a return code as specified in Section 4)
- Nexthop resolution status (resolved/unresolved) notification
6. RIB Grammar 📜
This section specifies the RIB information model in Routing Backus-Naur Form (rBNF) [RFC5511]. The grammar is there to help you understand Section 2 better so you can derive a data model. It's the blueprint, bestie 🧚
<routing-instance> ::= <INSTANCE_NAME>
[<interface-list>] <rib-list>
[<ROUTER_ID>]
<interface-list> ::= (<INTERFACE_IDENTIFIER> ...)
<rib-list> ::= (<rib> ...)
<rib> ::= <rib-name> <address-family>
[<route> ... ]
[ENABLE_IP_RPF_CHECK]
<address-family> ::= <IPV4_ADDRESS_FAMILY> | <IPV6_ADDRESS_FAMILY> |
<MPLS_ADDRESS_FAMILY> | <IEEE_MAC_ADDRESS_FAMILY>
<route> ::= <match> <nexthop>
[<route-attributes>]
[<route-vendor-attributes>]
<match> ::= <IPV4> <ipv4-route> | <IPV6> <ipv6-route> |
<MPLS> <MPLS_LABEL> | <IEEE_MAC> <MAC_ADDRESS> |
<INTERFACE> <INTERFACE_IDENTIFIER>
<route-type> ::= <IPV4> | <IPV6> | <MPLS> | <IEEE_MAC> | <INTERFACE>
<ipv4-route> ::= <ip-route-type>
(<destination-ipv4-address> | <source-ipv4-address> |
(<destination-ipv4-address> <source-ipv4-address>))
<destination-ipv4-address> ::= <ipv4-prefix>
<source-ipv4-address> ::= <ipv4-prefix>
<ipv4-prefix> ::= <IPV4_ADDRESS> <IPV4_PREFIX_LENGTH>
<ipv6-route> ::= <ip-route-type>
(<destination-ipv6-address> | <source-ipv6-address> |
(<destination-ipv6-address> <source-ipv6-address>))
<destination-ipv6-address> ::= <ipv6-prefix>
<source-ipv6-address> ::= <ipv6-prefix>
<ipv6-prefix> ::= <IPV6_ADDRESS> <IPV6_PREFIX_LENGTH>
<ip-route-type> ::= <SRC> | <DEST> | <DEST_SRC>
<route-attributes> ::= <ROUTE_PREFERENCE> [<LOCAL_ONLY>]
[<address-family-route-attributes>]
<address-family-route-attributes> ::= <ip-route-attributes> |
<mpls-route-attributes> |
<ethernet-route-attributes>
<ip-route-attributes> ::= <>
<mpls-route-attributes> ::= <>
<ethernet-route-attributes> ::= <>
<route-vendor-attributes> ::= <>
<nexthop> ::= <nexthop-base> |
(<NEXTHOP_LOAD_BALANCE> <nexthop-lb>) |
(<NEXTHOP_PROTECTION> <nexthop-protection>) |
(<NEXTHOP_REPLICATE> <nexthop-replicate>) |
<nexthop-chain>
<nexthop-base> ::= <NEXTHOP_ID> |
<nexthop-special> |
<egress-interface> |
<ipv4-address> | <ipv6-address> |
(<egress-interface>
(<ipv4-address> | <ipv6-address>)) |
(<egress-interface> <IEEE_MAC_ADDRESS>) |
<tunnel-encapsulation> | <tunnel-decapsulation> |
<logical-tunnel> |
<rib-name>
<egress-interface> ::= <INTERFACE_IDENTIFIER>
<nexthop-special> ::= <DISCARD> | <DISCARD_WITH_ERROR> |
(<RECEIVE> [<COS_VALUE>])
<nexthop-lb> ::= <NEXTHOP_LB_WEIGHT> <nexthop>
(<NEXTHOP_LB_WEIGHT> <nexthop) ...
<nexthop-protection> = <NEXTHOP_PREFERENCE> <nexthop>
(<NEXTHOP_PREFERENCE> <nexthop>)...
<nexthop-replicate> ::= <nexthop> <nexthop> ...
<nexthop-chain> ::= <nexthop> ...
<logical-tunnel> ::= <tunnel-type> <TUNNEL_NAME>
<tunnel-type> ::= <IPV4> | <IPV6> | <MPLS> | <GRE> | <VxLAN> | <NVGRE>
<tunnel-encapsulation> ::= (<IPV4> <ipv4-header>) |
(<IPV6> <ipv6-header>) |
(<MPLS> <mpls-header>) |
(<GRE> <gre-header>) |
(<VXLAN> <vxlan-header>) |
(<NVGRE> <nvgre-header>)
<ipv4-header> ::= <SOURCE_IPv4_ADDRESS> <DESTINATION_IPv4_ADDRESS>
<PROTOCOL> [<TTL>] [<DSCP>]
<ipv6-header> ::= <SOURCE_IPV6_ADDRESS> <DESTINATION_IPV6_ADDRESS>
<NEXT_HEADER> [<TRAFFIC_CLASS>]
[<FLOW_LABEL>] [<HOP_LIMIT>]
<mpls-header> ::= (<mpls-label-operation> ...)
<mpls-label-operation> ::= (<MPLS_PUSH> <MPLS_LABEL> [<S_BIT>]
[<TOS_VALUE>] [<TTL_VALUE>]) |
(<MPLS_SWAP> <IN_LABEL> <OUT_LABEL>
[<TTL_ACTION>])
<gre-header> ::= <GRE_IP_DESTINATION> <GRE_PROTOCOL_TYPE> [<GRE_KEY>]
<vxlan-header> ::= (<ipv4-header> | <ipv6-header>)
[<VXLAN_IDENTIFIER>]
<nvgre-header> ::= (<ipv4-header> | <ipv6-header>)
<VIRTUAL_SUBNET_ID>
[<FLOW_ID>]
<tunnel-decapsulation> ::= ((<IPV4> <IPV4_DECAP> [<TTL_ACTION>]) |
(<IPV6> <IPV6_DECAP> [<HOP_LIMIT_ACTION>]) |
(<MPLS> <MPLS_POP> [<TTL_ACTION>]))Figure 5: RIB rBNF Grammar
6.1. Nexthop Grammar Explained 🧠
A nexthop is used to specify the next network element to forward the traffic to. It's also used to say how the traffic should be load-balanced, protected using preference, or multicast using replication. All of this is explicitly spelled out in the grammar. The nexthop has recursion built in to handle complex use cases, like the one defined in Section 7.2.6 🔥
7. Using the RIB Grammar 🌈
The RIB grammar is very generic and covers a variety of features. This section gives examples of using objects in the RIB grammar and examples of programming certain use cases 💖
7.1. Using Route Preference 🦋
Using route preference, a client can preinstall alternate paths in the network. For example, if OSPF has a route preference of 10, then another client can install a route with a route preference of 20 to the same destination. The OSPF route gets precedence and is installed in the FIB. When the OSPF route is withdrawn, the alternate path gets installed in the FIB instead ✨
Route preference can also be used to prevent denial-of-service attacks by installing routes with the best preference, which either drops the offending traffic or routes it to some monitoring/analysis station. Since those routes are installed with the best preference, they supersede any route installed by any other protocol. Very dramatic, very effective 💀
7.2. Using Different Nexthop Types 💅
The RIB grammar lets you create a variety of nexthops. This section describes uses for certain types of nexthops 🌸
7.2.1. Tunnel Nexthops 🚇
A tunnel nexthop points to a tunnel of some kind. Traffic that goes over the tunnel gets encapsulated with the tunnel-encapsulation. Tunnel nexthops are useful for abstracting out the details of the network by letting traffic route seamlessly between network edges 🫶 At the end of a tunnel, the tunnel gets decapsulated. So the grammar supports two kinds of operations: one for encapsulation and another for decapsulation, and honestly both are giving 🔥
7.2.2. Replication Lists
okay so replication lists are giving main character energy 🦋 One can make a replication list for copying traffic out to multiple destinations at once! 💖 The destinations can themselves be derived nexthops, as long as the network device supports that level 🫶 Point to multipoint and broadcast are the classic examples that involve replication 🌸
A replication list at the simplest level looks like this 🎀
<nexthop> ::= <NEXTHOP_REPLICATE> <nexthop> [ <nexthop> ... ]And you can derive it from the grammar like so ✨
<nexthop> ::= <nexthop-replicate>
<nexthop> ::= <NEXTHOP_REPLICATE> <nexthop> <nexthop> ...7.2.3. Weighted Lists
A weighted list is how you load-balance traffic across a set of nexthops 🍓 Honestly it's super similar to a replication list, with one big difference: every member nexthop MUST have a NEXTHOP_LB_WEIGHT attached to it 💅
A weighted list at the simplest level looks like this 🧚
<nexthop> ::= <NEXTHOP_LOAD_BALANCE> (<nexthop> <NEXTHOP_LB_WEIGHT>)
[(<nexthop> <NEXTHOP_LB_WEIGHT>)... ]Derived from the grammar, it comes out like this ✨
<nexthop> ::= <nexthop-lb>
<nexthop> ::= <NEXTHOP_LOAD_BALANCE>
<NEXTHOP_LB_WEIGHT> <nexthop>
(<NEXTHOP_LB_WEIGHT> <nexthop>) ...
<nexthop> ::= <NEXTHOP_LOAD_BALANCE> (<NEXTHOP_LB_WEIGHT> <nexthop>)
(<NEXTHOP_LB_WEIGHT> <nexthop>) ...7.2.4. Protection
Primary/backup protection is a whole vibe 🛡️✨ It can be represented as ✨
<nexthop> ::= <NEXTHOP_PROTECTION> <1> <interface-primary>
<2> <interface-backup>)The above can be derived from the grammar as follows 💜
<nexthop> ::= <nexthop-protection>
<nexthop> ::= <NEXTHOP_PROTECTION> (<NEXTHOP_PREFERENCE> <nexthop>
(<NEXTHOP_PREFERENCE> <nexthop>)...)
<nexthop> ::= <NEXTHOP_PROTECTION> (<NEXTHOP_PREFERENCE> <nexthop>
(<NEXTHOP_PREFERENCE> <nexthop>))
<nexthop> ::= <NEXTHOP_PROTECTION> ((<NEXTHOP_PREFERENCE> <nexthop-base>
(<NEXTHOP_PREFERENCE> <nexthop-base>))
<nexthop> ::= <NEXTHOP_PROTECTION> (<1> <interface-primary>
(<2> <interface-backup>))Traffic can be load-balanced among multiple primary nexthops with a single backup 🌈 In that case, the nexthop looks like this 👑
<nexthop> ::= <NEXTHOP_PROTECTION> (<1>
(<NEXTHOP_LOAD_BALANCE>
(<NEXTHOP_LB_WEIGHT> <nexthop-base>
(<NEXTHOP_LB_WEIGHT> <nexthop-base>) ...))
<2> <nexthop-base>)A backup can also have its own backup 🙈 Then the list looks like this ✨
<nexthop> ::= <NEXTHOP_PROTECTION> (<1> <nexthop>
<2> <NEXTHOP_PROTECTION>(<1> <nexthop> <2> <nexthop>))7.2.5. Nexthop Chains
A nexthop chain is a way to do multiple operations on one packet by logically stringing them together 🔗 Say a VPN packet arrives on the WAN interface and needs to get forwarded to the correct VPN interface. You have to pop the VPN label before sending it out 💀 With a nexthop chain, you chain together "pop MPLS header" and "send it out a specific egress-interface" and boom, done 💅
The above example can be derived from the grammar like this ✨
<nexthop-chain> ::= <nexthop> <nexthop>
<nexthop-chain> ::= <nexthop-base> <nexthop-base>
<nexthop-chain> ::= <tunnel-decapsulation> <egress-interface>
<nexthop-chain> ::= (<MPLS> <MPLS_POP>) <interface-outgoing>Elements in a nexthop chain are evaluated left to right ➡️🦋
A nexthop chain can also put one or more headers on an outgoing packet 🎁 One example is a pseudowire, which is MPLS over some transport (MPLS or GRE, for instance) 🧚 Another example is Virtual eXtensible Local Area Network (VXLAN) over IP 🌸 So a nexthop chain lets a RIB client break up programming the nexthop into independent pieces, one per encapsulation 🫶
A simple example of MPLS over GRE looks like this 💖
<nexthop-chain> ::= (<MPLS> <mpls-header>) (<GRE> <gre-header>)
<interface-outgoing>The above can be derived from the grammar as follows ✨
<nexthop-chain> ::= <nexthop> <nexthop> <nexthop>
<nexthop-chain> ::= <nexthop-base> <nexthop-base> <nexthop-base>
<nexthop-chain> ::= <tunnel-encapsulation> <tunnel-encapsulation>
<egress-interface>
<nexthop-chain> ::= (<MPLS> <mpls-header>) (<GRE> <gre-header>)
<interface-outgoing>7.2.6. Lists of Lists
Lists of lists is a derived construct 🪆 One example of using it is replicating traffic to multiple destinations with load-balancing. For each branch of the replication tree, there are multiple interfaces where traffic needs to be load-balanced. So the outer list is a replication list for multicast, and the inner lists are weighted lists for load-balancing 🎀 Let's take a network element that has to replicate traffic to two other network elements. Traffic to the first one should be load-balanced equally over two interfaces: outgoing-1-1 and outgoing-1-2. Traffic to the second one should be load-balanced over three interfaces: outgoing-2-1, outgoing-2-2, and outgoing-2-3 (in the ratio 20:20:60) 📊✨
This can be derived from the grammar as follows 💅
<nexthop> ::= <nexthop-replicate>
<nexthop> ::= <NEXTHOP_REPLICATE> (<nexthop> <nexthop>...)
<nexthop> ::= <NEXTHOP_REPLICATE> (<nexthop> <nexthop>)
<nexthop> ::= <NEXTHOP_REPLICATE> ((<NEXTHOP_LOAD_BALANCE> <nexthop-lb>)
(<NEXTHOP_LOAD_BALANCE> <nexthop-lb>))
<nexthop> ::= <NEXTHOP_REPLICATE> ((<NEXTHOP_LOAD_BALANCE>
(<NEXTHOP_LB_WEIGHT> <nexthop>
(<NEXTHOP_LB_WEIGHT> <nexthop>) ...))
((<NEXTHOP_LOAD_BALANCE>
(<NEXTHOP_LB_WEIGHT> <nexthop>
(<NEXTHOP_LB_WEIGHT> <nexthop>) ...))
<nexthop> ::= <NEXTHOP_REPLICATE> ((<NEXTHOP_LOAD_BALANCE>
(<NEXTHOP_LB_WEIGHT> <nexthop>
(<NEXTHOP_LB_WEIGHT> <nexthop>)))
((<NEXTHOP_LOAD_BALANCE>
(<NEXTHOP_LB_WEIGHT> <nexthop>
(<NEXTHOP_LB_WEIGHT> <nexthop>)
(<NEXTHOP_LB_WEIGHT> <nexthop>)))
<nexthop> ::= <NEXTHOP_REPLICATE> ((<NEXTHOP_LOAD_BALANCE>
(<NEXTHOP_LB_WEIGHT> <nexthop>)
(<NEXTHOP_LB_WEIGHT> <nexthop>)))
((<NEXTHOP_LOAD_BALANCE>
(<NEXTHOP_LB_WEIGHT> <nexthop>)
(<NEXTHOP_LB_WEIGHT> <nexthop>)
(<NEXTHOP_LB_WEIGHT> <nexthop>)))
<nexthop> ::= <NEXTHOP_REPLICATE>
((<NEXTHOP_LOAD_BALANCE>
(50 <outgoing-1-1>)
(50 <outgoing-1-2>)))
((<NEXTHOP_LOAD_BALANCE>
(20 <outgoing-2-1>)
(20 <outgoing-2-2>)
(60 <outgoing-2-3>)))7.3. Performing Multicast
IP multicast means matching a packet on (S,G) or (*,G), where both S (Source) and G (Group) are IP prefixes 🌊 After the match, the packet gets replicated to one or more recipients 📨 How recipients subscribe to the multicast group is outside the scope of this document 🙅♀️
In PIM-based multicast, packets are IP forwarded on an IP multicast tree. The downstream nodes at each point in the multicast tree are one or more IP addresses. These can be represented as a replication list (see Section 7.2.2) 🌳
In MPLS-based multicast, packets are forwarded on a Point-to-Multipoint (P2MP) LSP. The nexthop for a P2MP LSP can be represented in the nexthop grammar as a <logical-tunnel> (P2MP LSP identifier), or as a replication list (see Section 7.2.2) of <tunnel-encapsulation>, with each tunnel-encapsulation representing a single MPLS downstream nexthop 🦋
- RIB Operations at Scale
This section is about the scale requirements for a RIB data model 📈 The RIB data model should be able to handle a large scale of operations so RIB applications can be deployed in large networks 🏗️
8.1. RIB Reads
Bulking (grouping multiple objects into a single message) MUST be supported when a network device sends RIB data to a RIB client 📦 Similarly, the data model MUST enable a RIB client to request data in bulk from a network device 🙌
8.2. RIB Writes
Bulking (grouping multiple write operations into a single message) MUST be supported when a RIB client wants to write to the RIB ✍️ The response from the network device MUST include a return-code for each write operation in the bulk message 🔢
8.3. RIB Events and Notifications
Sometimes one network event causes multiple events and/or notifications from the network device to a RIB client 🎇 On the flip side, because of timing when lots of things happen at once, a network device might have to send multiple events and/or notifications to a RIB client ⏱️ The network-device-originated event/notification message MUST support bulking multiple events and notifications in a single message 💌
- Security Considerations
Okay bestie, this part is SERIOUS 😭🚨 The information model in this document defines a schema for data models that are meant to be accessed through network management protocols such as NETCONF [RFC6241] or RESTCONF [RFC8040] 🔐 The lowest NETCONF layer is the secure transport layer, and the mandatory-to-implement secure transport is Secure Shell (SSH) [RFC6242] 🐚 The lowest RESTCONF layer is HTTPS, and the mandatory-to-implement secure transport is TLS [RFC8446] 🔒
The NETCONF access control model [RFC8341] lets you restrict access for particular NETCONF or RESTCONF users to a preconfigured subset of all the available NETCONF or RESTCONF protocol operations and content 🚪
The RIB information model specifies read and write operations to network devices 🖥️ These network devices might be considered sensitive or vulnerable in some network environments 🥺 Write operations to these devices without proper protection can have a negative effect on network operations 💥 Because of this, it is recommended that data models also consider the following in their design 👇
- Require use of the authentication and authorization features of the NETCONF or RESTCONF suite of protocols 🔑
- Augment the limits on how much data can be written or updated by a remote entity, built to include enough protection for a RIB data model 🧱
- Expose the specific RIB data model implemented via NETCONF/RESTCONF data models 📣
- IANA Considerations
This document has no IANA actions 🫡
- References
Okay bestie, let's finish strong! 💖 Picking up right where we left off with the References. Here's the normative and informative crew 🦋
11.1. Normative References
- [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997, <https://www.rfc-editor.org/info/rfc2119>.
- [RFC6241] Enns, R., Ed., Bjorklund, M., Ed., Schoenwaelder, J., Ed., and A. Bierman, Ed., "Network Configuration Protocol (NETCONF)", RFC 6241, DOI 10.17487/RFC6241, June 2011, <https://www.rfc-editor.org/info/rfc6241>.
- [RFC6242] Wasserman, M., "Using the NETCONF Protocol over Secure Shell (SSH)", RFC 6242, DOI 10.17487/RFC6242, June 2011, <https://www.rfc-editor.org/info/rfc6242>.
- [RFC8040] Bierman, A., Bjorklund, M., and K. Watsen, "RESTCONF Protocol", RFC 8040, DOI 10.17487/RFC8040, January 2017, <https://www.rfc-editor.org/info/rfc8040>.
- [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, May 2017, <https://www.rfc-editor.org/info/rfc8174>.
- [RFC8341] Bierman, A. and M. Bjorklund, "Network Configuration Access Control Model", STD 91, RFC 8341, DOI 10.17487/RFC8341, March 2018, <https://www.rfc-editor.org/info/rfc8341>.
- [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, <https://www.rfc-editor.org/info/rfc8446>.
11.2. Informative References
- [RFC4915] Psenak, P., Mirtorabi, S., Roy, A., Nguyen, L., and P. Pillay-Esnault, "Multi-Topology (MT) Routing in OSPF", RFC 4915, DOI 10.17487/RFC4915, June 2007, <https://www.rfc-editor.org/info/rfc4915>.
- [RFC5120] Przygienda, T., Shen, N., and N. Sheth, "M-ISIS: Multi Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)", RFC 5120, DOI 10.17487/RFC5120, February 2008, <https://www.rfc-editor.org/info/rfc5120>.
- [RFC5511] Farrel, A., "Routing Backus-Naur Form (RBNF): A Syntax Used to Form Encoding Rules in Various Routing Protocol Specifications", RFC 5511, DOI 10.17487/RFC5511, April 2009, <https://www.rfc-editor.org/info/rfc5511>.
- [RFC7920] Atlas, A., Ed., Nadeau, T., Ed., and D. Ward, "Problem Statement for the Interface to the Routing System", RFC 7920, DOI 10.17487/RFC7920, June 2016, <https://www.rfc-editor.org/info/rfc7920>.
- [RFC8431] Wang, L., Chen, M., Dass, A., Ananthakrishnan, H., Kini, S., and N. Bahadur, "A YANG Data Model for the Routing Information Base (RIB)", RFC 8431, DOI 10.17487/RFC8431, September 2018, <http://www.rfc-editor.org/info/rfc8431>.
Acknowledgements 🙌
The authors would like to thank Ron Folkes, Jeffrey Zhang, the WG co-Chairs, and reviewers for their comments and suggestions on this document. ✨ The following people contributed to the design of the RIB information model as part of the I2RS Interim meeting in April 2013: Wes George, Chris Liljenstolpe, Jeff Tantsura, Susan Hares, and Fabian Schneider. 🫶
Authors' Addresses 💌
Nitin Bahadur (editor) Uber 900 Arastradero Rd Palo Alto, CA 94304 United States of America
Email: nitin_bahadur@yahoo.com
Sriganesh Kini (editor)
Email: sriganeshkini@gmail.com
Jan Medved Cisco
Email: jmedved@cisco.com
That's the whole thing, bestie! 🦋👑 Authors, thank you for the RIB magic 💜
📦 where this page comes from
- ietf: rfc8430 2018-09
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms