Glossary

Term	Definition
ARL (Address Resolution Logic)	A forwarding table within TMM that maps MAC addresses to network interfaces. TMM uses ARL entries to determine where to send traffic based on the destination MAC address.
ARP (Address Resolution Protocol)	A network protocol used to map an IP address to a physical (MAC) address on a local network. In EVPN environments, ARP entries are distributed across VTEPs so that remote endpoints can be reached without flooding the network.
ASN (Autonomous System Number)	A unique number assigned to a network or group of networks that share a single routing policy. BGP uses ASNs to identify peers and determine routing paths. In F5 BIG-IP Next for Kubernetes, the GlobalRoutingConfig defines a range of ASNs that the controller assigns to individual TMM pods.
BFD (Bidirectional Forwarding Detection)	A lightweight protocol used to detect link or path failures quickly. BFD runs alongside BGP to provide fast failure detection, typically in milliseconds, so that BGP can reroute traffic faster than its own keepalive timers would allow.
BGP (Border Gateway Protocol)	A standardized routing protocol used to exchange routing information between networks (autonomous systems). BGP determines the best path for traffic based on policies, path attributes, and network reachability. Used in this architecture for both underlay routing and EVPN overlay control.
BGP Unnumbered	A BGP peering method that uses automatically generated IPv6 link-local addresses instead of manually assigned IP addresses. Neighbors are identified by interface name rather than IP address, which simplifies configuration in large data center fabrics.
CIDR (Classless Inter-Domain Routing)	A method for allocating IP addresses and defining IP routing paths using a prefix notation (for example, 192.168.10.0/24). The number after the slash indicates how many bits define the network portion of the address.
CLI (Command-Line Interface)	A text-based interface for interacting with software. In this context, the OcNOS CLI (cmlsh) is used to apply rendered routing configuration inside the routing container.
cmlsh	The OcNOS command-line shell. The grpcconf_watcher process uses cmlsh to apply rendered routing configuration inside the routing container.
CNE controller	The Kubernetes-native controller component in F5 BIG-IP Next for Kubernetes that watches for changes to routing CRDs, validates input, renders Go templates, and pushes configuration to the TMM routing container through gRPC. Also referred to as f5ingress.
ConfigBlob	The rendered configuration payload that the CNE controller produces from a Go template and pushes to the routing container through gRPC. Each reconciliation produces a full ConfigBlob—the controller never sends partial updates.
Control plane	The part of the network architecture responsible for making routing decisions. OcNOS owns the control plane and handles BGP session establishment, route advertisement, route learning, and EVPN route exchange.
CRD (Custom Resource Definition)	A Kubernetes extension mechanism that lets you define your own resource types. In this architecture, CRDs such as GlobalRoutingConfig, RoutingConfig, RoutingTemplate, Vrf, and Vxlan are used to declare routing intent.
Data plane	The part of the network architecture responsible for forwarding packets. TMM owns the data plane and handles tasks such as VXLAN encapsulation and decapsulation, VRF interface creation, and Layer 2/Layer 3 forwarding.
East–west traffic	Network traffic that flows between workloads on different nodes within the same cluster. This traffic uses VXLAN tunnels, where TMM on the source node encapsulates the packet and TMM on the destination node decapsulates it.
eBGP (External BGP)	A BGP session between two peers that belong to different autonomous systems. In BGP Unnumbered configurations, remote-as external indicates an eBGP session where the peer’s ASN is discovered dynamically.
EVPN (Ethernet VPN)	A BGP address family that distributes MAC addresses, IP addresses, and IP prefixes across VTEPs. EVPN acts as the control plane for VXLAN, enabling dynamic VTEP discovery and eliminating the need for static tunnel configuration or network flooding.
EVPN Route Type 5 (RT-5)	An EVPN route that advertises subnet-level reachability by carrying IP prefixes associated with an L3 VNI. Used for inter-subnet routing and north–south traffic.
F5 BIG-IP Next for Kubernetes	The F5 product that provides application delivery and security services in Kubernetes environments. It includes TMM for data plane forwarding and integrates with OcNOS for BGP routing.
F5SPKStaticRoute	A Kubernetes CRD used to define static routes. The vrf field associates the route with a specific tenant routing domain.
F5SPKVlan	A Kubernetes CRD used to define VLANs. The vrf field binds the VLAN to a specific VRF, and the presence or absence of self IP fields determines whether the VLAN is treated as a BGP Unnumbered interface.
FDB (Forwarding Database)	A table that maps MAC addresses to network ports or tunnel endpoints. In EVPN, FDB entries are distributed across VTEPs so that each VTEP knows which remote VTEP to send traffic to for a given MAC address.
GlobalRoutingConfig	A Kubernetes CRD (singleton per cluster) that defines the global base BGP configuration, including the ASN range, BGP neighbors, and a reference to a RoutingTemplate. All tenant RoutingConfig resources depend on it.
Go template	A text templating system built into the Go programming language. The CNE controller uses Go templates (defined in RoutingTemplate CRDs) to render routing configuration for each TMM pod, substituting values such as the per-device ASN and VTEP IP.
gRPC (gRPC Remote Procedure Call)	A high-performance communication framework used for service-to-service communication. The CNE controller uses gRPC to push rendered configuration payloads to the TMM routing container. The channel is secured with mutual TLS (mTLS).
grpcconf_watcher	A process inside the f5-tmm-routing container that receives gRPC messages from the CNE controller, performs post-processing, and applies the configuration through the OcNOS CLI.
IRB (Integrated Routing and Bridging)	An interface type that enables both Layer 2 (switching) and Layer 3 (routing) on the same interface. IRB interfaces are defined in the RoutingConfig CRD for tenant-specific configurations.
kubectl	The command-line tool for interacting with Kubernetes clusters. Used to create, view, and delete CRDs such as GlobalRoutingConfig, RoutingConfig, and others.
L3 VNI (Layer 3 Virtual Network Identifier)	A VNI used for Layer 3 routing between subnets within a tenant VRF, or between a tenant and external networks. Each tenant VRF maps 1:1 with an L3 VNI, and these mappings must be consistent across all VTEPs.
LLA (Link-Local Address)	An IPv6 address automatically generated on every IPv6-enabled interface (in the fe80::/10 range). BGP Unnumbered uses these addresses for neighbor discovery and session establishment without manual IP assignment.
MAC (Media Access Control) address	A unique hardware identifier assigned to a network interface. EVPN distributes MAC address information across VTEPs to enable Layer 2 forwarding without flooding.
mTLS (Mutual TLS)	A security mechanism where both the client and server authenticate each other using TLS certificates. The gRPC channel between the CNE controller and the routing container uses mTLS.
MTU (Maximum Transmission Unit)	The largest packet size (in bytes) that a network interface can transmit. VXLAN tunnels have a lower effective MTU than the physical network because of the encapsulation overhead.
ND (Neighbor Discovery)	An IPv6 protocol used to discover other devices on the same link, determine their link-layer addresses, and find routers. In EVPN, ND suppression reduces broadcast traffic by answering ND requests locally.
NDD (Neighbor Discovery Daemon)	A software component that manages IPv6 neighbor discovery data. In the BGP Unnumbered architecture, tmrouted runs an NDD server socket that accepts connections from OcNOS and forwards Router Advertisement data.
Netlink	A Linux kernel interface used for communication between the kernel and user-space processes. tmrouted uses Netlink messages to exchange VRF, interface, and route information between TMM and OcNOS.
North–south traffic	Network traffic that flows between internal workloads (such as pods) and external networks. This traffic passes through TMM, which performs VRF lookup, VXLAN encapsulation or decapsulation, and forwards packets to or from the fabric.
oamd	A daemon within OcNOS that configures and manages BFD sessions for fast link failure detection.
OcNOS	A network operating system that runs as the control plane agent inside the f5-tmm-routing container. OcNOS manages BGP sessions, EVPN route exchange, and BFD. It is developed by IP Infusion.
Protobuf (Protocol Buffers)	A data serialization format developed by Google. TMM uses protobuf messages to communicate with tmrouted, sending VRF configuration, VXLAN data, and Router Advertisement information.
RA (Router Advertisement)	An IPv6 message type used for neighbor discovery. In BGP Unnumbered, TMM sends periodic RAs containing its link-local address, and receives RAs from remote peers to discover their link-local addresses for BGP session establishment.
Route domain	A logical partition of the routing table within TMM. Each VRF maps to a route domain, and route domain 0 is the default. Route domains isolate tenant traffic at the routing level.
Route target	A BGP extended community attribute used by EVPN to control which VRFs import and export routes. A VRF only receives routes that match its configured route targets, which enforces tenant isolation.
RoutingConfig	A Kubernetes CRD that defines tenant-specific routing intent, including BGP neighbors and VXLAN or VRF attachment. One RoutingConfig is created per tenant, and it is processed only after the GlobalRoutingConfig reaches a successful status.
RoutingTemplate	A Kubernetes CRD that contains a Go template used by the CNE controller to render the final routing configuration. Both GlobalRoutingConfig and RoutingConfig reference a template.
Self IP	An IP address assigned to a VLAN interface on TMM. Self IPs are used for routing and traffic management. VLANs without self IPs are treated as BGP Unnumbered interfaces.
TCP-AO / MD5	Authentication mechanisms for BGP sessions. TCP-AO (TCP Authentication Option) and MD5 provide integrity and authentication for BGP traffic between peers.
TMM (Traffic Management Microkernel)	The core data plane engine in F5 BIG-IP Next for Kubernetes. TMM handles packet forwarding, VXLAN encapsulation and decapsulation, VRF interface creation, and Layer 2/Layer 3 forwarding.
tmrouted	A component that acts as the communication bridge between TMM and OcNOS. It translates protobuf messages from TMM into Netlink messages that OcNOS understands, and forwards OcNOS route updates back to TMM.
ToR (Top-of-Rack)	A network switch located at the top of a server rack in a data center. ToR switches connect servers in the rack to the broader network fabric. In the NVIDIA Cloud Partner architecture, ToR switches use BGP Unnumbered for peering.
VNI (Virtual Network Identifier)	A 24-bit identifier used by VXLAN to tag traffic for a specific tenant or network segment. VNIs enable up to 16 million isolated network segments, compared to the 4,094 limit of traditional VLANs.
VRF (Virtual Routing and Forwarding)	A technology that creates separate, isolated routing tables within the same network device. Each tenant can have its own VRF, which prevents routing information and traffic from leaking between tenants.
Vrf (CRD)	A Kubernetes CRD that creates a tenant-specific VRF for routing domain isolation. The system generates a deterministic route domain ID from the VRF name.
VTEP (VXLAN Tunnel Endpoint)	The point where VXLAN encapsulation and decapsulation occurs. Each TMM instance acts as a VTEP, wrapping tenant traffic in VXLAN headers for transmission across the fabric and unwrapping incoming VXLAN traffic.
VXLAN (Virtual Extensible LAN)	A network overlay technology that encapsulates Layer 2 Ethernet frames within UDP packets to create virtual tunnels across a Layer 3 network. VXLAN enables scalable multi-tenant networking by using VNIs for segmentation.
Vxlan (CRD)	A Kubernetes CRD that configures VXLAN tunnel encapsulation for tenant traffic, including the VNI, VTEP IP, MTU, port, and VRF association.