Route Reflectors reduce the need for a full iBGP mesh by reflecting selected routes between clients while using attributes to prevent reflection loops.
Learning objectives
- Calculate full-mesh session requirements.
- Explain RR clients and non-clients.
- Apply reflection advertisement rules.
- Explain Originator ID and Cluster List.
- Understand redundancy, path hiding and Add-Path.
1. Full mesh scaling
Without route reflection, iBGP requires a full mesh so routes learned from one iBGP peer are not re-advertised to another iBGP peer. Session count is N(N−1)/2.
| Routers | Full-mesh sessions |
|---|---|
| 4 | 6 |
| 10 | 45 |
| 20 | 190 |
| 50 | 1,225 |
2. Reflection rules
| Route source | RR may advertise to |
|---|---|
| eBGP peer | Eligible iBGP clients and non-clients |
| RR client | Other clients and non-clients |
| RR non-client | Clients only |
Non-client iBGP peers still require a full mesh among themselves unless another supported architecture is used. An RR normally reflects its selected best path, which can cause path hiding.
3. Cisco IOS configuration
router bgp 65001
neighbor 10.0.0.2 remote-as 65001
neighbor 10.0.0.2 route-reflector-client
neighbor 10.0.0.3 remote-as 65001
neighbor 10.0.0.3 route-reflector-client
4. Loop prevention
- Originator ID identifies the router that originated the route into the reflection process; a router rejects a route carrying its own ID.
- Cluster List records cluster IDs traversed; an RR rejects a route containing its own cluster ID.
- These attributes prevent reflection loops but do not replace correct design and policy.
5. Redundancy, path hiding and Add-Path
Multiple RRs improve resilience. Clients may peer with more than one RR. Since an RR generally advertises its selected best path, another useful path can be hidden. BGP Add-Path allows multiple paths to be advertised when supported and configured.
Lab extension
Preserve the existing four-router EVE-NG baseline. Add a separate R5 or clone a snapshot for RR experiments; verify interface mappings and software support.
Senior QA scenarios
| ID | Test | Expected result |
|---|---|---|
| L-01 | Calculate mesh for N routers | N(N−1)/2. |
| L-02 | Client route learned | Reflected to eligible clients/non-clients. |
| L-03 | Non-client route learned | Reflected to clients only. |
| L-04 | Own Originator ID returns | Original router rejects route. |
| L-05 | Own Cluster ID present | RR rejects looped route. |
| L-06 | RR failure with redundant RR | Alternate RR maintains/re-establishes reachability. |
| L-07 | Two candidate paths at RR | Best-path reflection verified. |
| L-08 | Enable Add-Path | Multiple paths advertised/received as configured. |
| L-09 | Client/non-client policy | No unintended propagation. |
| L-10 | RR failure and recovery | Route and forwarding recovery measured. |
Interview questions
- Why use Route Reflectors?
- Full-mesh formula?
- Client versus non-client?
- How are client-learned routes handled?
- How are non-client-learned routes handled?
- What is Originator ID?
- What is Cluster List?
- What is path hiding?
- How does Add-Path help?
- Why deploy redundant RRs?
Student Knowledge Assessment – Module L (reconstructed)
Q1. How many full-mesh iBGP sessions are required for 5 routers?
A. 5
B. 10
C. 15
D. 20
Correct answer: B. 10
Explanation: N(N−1)/2 = 5×4/2 = 10.
Q2. What is the main purpose of an RR?
A. Replace TCP
B. Reduce full-mesh iBGP peering requirements
C. Encrypt updates
D. Assign ASNs
Correct answer: B. Reduce full-mesh iBGP peering requirements
Explanation: An RR reflects eligible routes to clients.
Q3. A route learned from an RR client can be reflected to:
A. Only source client
B. Other clients and non-clients
C. Only eBGP peers
D. No peers
Correct answer: B. Other clients and non-clients
Explanation: Client-learned routes can be reflected to other clients and non-clients.
Q4. A route learned from a non-client is reflected to:
A. Other non-clients only
B. Clients only
C. All peers
D. No one
Correct answer: B. Clients only
Explanation: Non-client-learned routes are reflected to clients only.
Q5. What does Originator ID identify?
A. Original router injecting route into reflection
B. TCP server port
C. Local preference
D. Cluster count
Correct answer: A. Original router injecting route into reflection
Explanation: Originator ID records the originating router ID for loop prevention.
Q6. What does Cluster List contain?
A. ASNs
B. Cluster IDs traversed
C. Interface addresses
D. MED values
Correct answer: B. Cluster IDs traversed
Explanation: Cluster List records the route-reflector cluster path.
Q7. What is path hiding?
A. Route hidden by ACL
B. RR advertises selected best path and may not advertise another useful path
C. TCP encrypted
D. Route not originated
Correct answer: B. RR advertises selected best path and may not advertise another useful path
Explanation: Best-path-only reflection can hide alternate paths.
Q8. What does Add-Path enable?
A. Multiple paths advertised
B. Multiple TCP ports
C. More ASNs
D. Zero Hold Time
Correct answer: A. Multiple paths advertised
Explanation: Add-Path allows multiple paths for a prefix to be advertised.
Q9. Do RR clients need direct sessions to each other in a standard RR design?
A. Yes always
B. No, not for the client mesh served by RR
C. Only eBGP
D. Only with MED
Correct answer: B. No, not for the client mesh served by RR
Explanation: Clients exchange routes through the RR.
Q10. Why use redundant RRs?
A. Remove policies
B. Improve resilience and avoid a single RR failure point
C. Disable AS_PATH
D. Replace IGP
Correct answer: B. Improve resilience and avoid a single RR failure point
Explanation: Redundant RRs improve control-plane resilience.