Fabric Standardization — Small EVPN-VXLAN Data Centers
Status: Draft — Phase 0 (#31)
Scope: POC for small data centers (2 spines, 3-8 leafs)
Parent: Epic #30
1. Topology constraints
1.1 Spine layer
- Always 2 spines per fabric (redundancy, no single point of failure)
- Spines are pure L3 routers — no VTEPs, no VLANs, no MLAG
- Each spine connects to every leaf via a dedicated P2P link
1.2 Leaf layer
- Leafs always come in MLAG pairs (2 leafs = 1 VTEP)
- Minimum 3 pairs (6 leafs), maximum 4 pairs (8 leafs) per fabric
- Each leaf connects to both spines via dedicated uplinks
- Each leaf pair shares a VTEP loopback IP (Loopback1)
1.3 Host connectivity
- Hosts connect directly to leaf pairs — no access switches
- Every host is dual-homed via MLAG (LACP active)
- Each host-facing port-channel gets a unique MLAG ID
1.4 Containerlab model
- All devices use the
ceos image (Arista cEOS)
- The lab simulates a fixed 12-port model per device
2. Port assignment — Spine
Spines use a simple sequential mapping: one Ethernet port per leaf.
| Port |
Role |
Connected to |
| Ethernet1 |
Underlay downlink |
leaf1 |
| Ethernet2 |
Underlay downlink |
leaf2 |
| Ethernet3 |
Underlay downlink |
leaf3 |
| ... |
... |
... |
| Ethernet{N} |
Underlay downlink |
leaf{N} |
All spine downlinks are:
- Routed (
no switchport)
- /31 P2P addressing
- MTU 9214 (jumbo frames for VXLAN encapsulation)
3. Port assignment — Leaf
Leaf port allocation follows a fixed layout. Ports are divided into 3 zones:
| Port range |
Role |
Details |
| Ethernet1 — Ethernet9 |
Host-facing |
MLAG port-channels (trunk, LACP active) |
| Ethernet10 |
MLAG peer-link |
Port-Channel999 (trunk, trunk-group mlag-peer) |
| Ethernet11 |
Spine1 uplink |
Routed P2P /31, MTU 9214 |
| Ethernet12 |
Spine2 uplink |
Routed P2P /31, MTU 9214 |
3.1 Host-facing ports (Ethernet1-9)
- Each physical port is a member of a Port-Channel
- Port-Channel number = MLAG ID = host index (e.g., host1 → Po1, MLAG 1)
- Mode:
switchport mode trunk
- VLANs: only the VLANs needed by the host
- LACP fallback enabled (timeout 5, individual)
3.2 MLAG peer-link (Ethernet10)
- Always Port-Channel999
- Trunk mode with trunk-group
mlag-peer
- Spanning-tree link-type point-to-point
- Carries MLAG control traffic + VLANs 4090, 4091
3.3 Spine uplinks (Ethernet11-12)
- Ethernet11 → spine1, Ethernet12 → spine2
- Fixed mapping, never changes regardless of fabric size
4. Naming conventions
4.1 Devices
| Role |
Pattern |
Examples |
| Spine |
spine{N} |
spine1, spine2 |
| Leaf |
leaf{N} |
leaf1, leaf2, ..., leaf8 |
| Host |
host{N} |
host1, host2, host3, host4 |
- Leafs are numbered sequentially: odd = primary, even = secondary within a pair
- Pair 1: leaf1 (primary), leaf2 (secondary)
- Pair 2: leaf3 (primary), leaf4 (secondary)
- Pair N: leaf{2N-1} (primary), leaf{2N} (secondary)
4.2 MLAG domain
- Domain ID:
leafs (same for all pairs — Arista uses this as an identifier within the MLAG)
- Peer-link: Port-Channel999
- Peer-link VLAN: 4090
- iBGP peering VLAN: 4091
4.3 Fabric
- Pattern:
{site}-{function} (e.g., dc1-evpn, paris-evpn)
4.4 IPAM identifiers (for resource pool idempotence)
| Object |
Identifier pattern |
Example |
| Site infra prefix |
site-{site}-infra |
site-dc1-infra |
| Site services prefix |
site-{site}-services |
site-dc1-services |
| Device loopback0 IP |
lo0-{fabric}-{device} |
lo0-evpnlab-leaf1 |
| Device loopback1 IP |
lo1-{fabric}-vtep{pair} |
lo1-evpnlab-vtep1 |
| Underlay P2P /31 |
p2p-{fabric}-{spine}-{leaf} |
p2p-evpnlab-spine1-leaf1 |
| MLAG peer /31 |
mlag-peer-{fabric}-pair{N} |
mlag-peer-evpnlab-pair1 |
| MLAG iBGP /31 |
mlag-ibgp-{fabric}-pair{N} |
mlag-ibgp-evpnlab-pair1 |
| Leaf ASN |
asn-{fabric}-pair{N} |
asn-evpnlab-pair1 |
5. IPAM — IP addressing plan
5.1 Supernets (global)
| Role |
Supernet |
Description |
| Infrastructure |
10.0.0.0/8 |
Loopbacks, underlay, MLAG |
| Services |
172.16.0.0/12 |
L2/L3 VXLAN user subnets |
5.2 Site allocation (from supernets)
Each site receives:
- 1x /16 from
10.0.0.0/8 for infrastructure
- 1x /16 from
172.16.0.0/12 for services
5.3 Fabric pools (from site infra /16)
| Pool |
Prefix size |
Allocation unit |
Pool type |
| Loopback0 (router-id) |
/24 |
/32 per device |
CoreIPAddressPool |
| Loopback1 (VTEP) |
/24 |
/32 per MLAG pair |
CoreIPAddressPool |
| Underlay P2P |
/24 |
/31 per spine-leaf link |
CoreIPPrefixPool |
| MLAG peer-link SVI |
/24 |
/31 per MLAG pair |
CoreIPPrefixPool |
| MLAG iBGP peering |
/24 |
/31 per MLAG pair |
CoreIPPrefixPool |
5.4 Service pools (from site services /16)
| Pool |
Allocation unit |
Pool type |
| L2 VXLAN subnets |
/24 per VLAN (customizable) |
CoreIPPrefixPool |
| L3 VXLAN subnets (VRF SVIs) |
/24 per VRF SVI (customizable) |
CoreIPPrefixPool |
5.5 Special VLANs (reserved, not from pools)
| VLAN |
Name |
Purpose |
Trunk group |
| 4090 |
mlag-peer |
MLAG peer-link SVI |
mlag-peer |
| 4091 |
mlag-ibgp |
MLAG iBGP peering |
mlag-peer |
6. BGP — Autonomous System assignment
6.1 Spine ASN
- Single ASN shared by all spines in a fabric
- Defined as an attribute on
InfraFabric
- Default for POC: 65000
6.2 Leaf ASN
- One ASN per MLAG pair (iBGP within pair, eBGP to spines)
- Allocated from a
CoreNumberPool (range: 65001–65099)
- Deterministic via identifier:
asn-{fabric}-pair{N}
6.3 BGP configuration standards
| Parameter |
Value |
Notes |
no bgp default ipv4-unicast |
Always |
Explicit activation per AFI |
distance bgp |
20 200 200 |
eBGP preferred over iBGP |
maximum-paths |
4 ecmp 64 |
Multi-path for spine redundancy |
maximum-routes |
12000 warning-only |
Per neighbor |
ebgp-multihop |
3 |
EVPN overlay (loopback peering) |
send-community extended |
Always |
Required for EVPN route-targets |
next-hop-unchanged |
Spine EVPN peer-group |
Preserve leaf next-hop in overlay |
next-hop-self |
Leaf iBGP peer-group |
Required for iBGP convergence |
6.4 Peer groups (per device)
Leaf peer groups:
| Peer group |
Type |
Remote AS |
Neighbors |
underlay |
eBGP |
spine ASN |
Spine P2P IPs |
underlay_ibgp |
iBGP |
own ASN |
MLAG peer via VLAN 4091 |
evpn |
eBGP |
spine ASN |
Spine loopback0 IPs |
Spine peer groups:
| Peer group |
Type |
Neighbors |
evpn |
eBGP |
All leaf loopback0 IPs (each with its own remote-as) |
Spine underlay neighbors are configured individually (no peer-group) since each leaf has a different ASN.
6.5 Address families
| AFI |
Activated on |
Networks advertised |
| IPv4 unicast |
underlay, underlay_ibgp |
Loopback0/32, Loopback1/32 |
| EVPN |
evpn |
(routes from VLAN/VRF config) |
7. MLAG standards
| Parameter |
Value |
| Domain ID |
leafs |
| Peer-link interface |
Port-Channel999 |
| Peer-link VLAN |
4090 (IP: /31 from MLAG peer pool) |
| iBGP VLAN |
4091 (IP: /31 from MLAG iBGP pool, MTU 9214) |
| Peer VLAN autostate |
Disabled (no autostate) |
| Dual-primary detection |
Enabled (delay 10, errdisable all-interfaces) |
| Heartbeat |
Via Management0 (VRF mgmt) |
| Virtual MAC |
c001.cafe.babe (fabric-wide anycast gateway) |
7.1 Primary/secondary assignment
- Odd-numbered leaf (leaf1, leaf3, leaf5, leaf7): lower IP on MLAG VLANs (e.g., x.x.x.254/31)
- Even-numbered leaf (leaf2, leaf4, leaf6, leaf8): higher IP (e.g., x.x.x.255/31)
8. VXLAN standards
8.1 VTEP interface
- Interface:
Vxlan1 on every leaf
- Source interface:
Loopback1 (shared IP within MLAG pair)
- UDP port:
4789
- Learning:
vxlan learn-restrict any (EVPN-controlled)
8.2 VNI allocation
| Type |
NumberPool range |
Usage |
| L2 VNI |
100001–199999 |
One VNI per extended VLAN (EVPN Type-2) |
| L3 VNI |
200001–299999 |
One VNI per VRF (EVPN Type-5) |
VNIs are allocated from CoreNumberPool with deterministic identifiers.
8.3 Route distinguisher and route target
| Service type |
RD format |
RT format |
| L2 VXLAN (per VLAN) |
{ASN}:{VNI} |
{VLAN_ID}:{VNI} (import/export) |
| L3 VXLAN (per VRF) |
{Loopback0_IP}:{VRF_index} |
{VRF_index}:{VNI} (import/export evpn) |
9. Global parameters
| Parameter |
Value |
Notes |
| Underlay MTU |
9214 |
All P2P and iBGP links |
| Anycast gateway MAC |
c001.cafe.babe |
ip virtual-router mac-address |
| Routing model |
multi-agent |
service routing protocols model multi-agent |
| Spanning-tree |
Disabled on VLAN 4090, 4091 |
MLAG VLANs only |
| LLDP |
Management0 |
lldp management-address Management0 |
| gNMI |
Enabled |
management api gnmi with provider eos-native |
10. Out of scope (for now)
- Access switches — hosts connect directly to leafs
- Multi-fabric / DCI — single fabric per site
- IPv6 underlay — IPv4 only
- BFD — not configured in initial POC
- Route-maps / prefix-lists — no filtering in the underlay
- More than 2 spines — fixed at 2 for the POC
- Non-Arista platforms — EOS only
