JNCIP-ENT JN0-650 · IGP · Advanced OSPF

Advanced OSPF on Junos: OSPFv3, IGP policy, overload and virtual links

Run OSPFv3 next to OSPFv2, control what enters and leaves OSPF with export and import policies (external type, metric, tag), steer transit traffic away from a router with overload, tune preferences, and glue a detached area to the backbone with a virtual link.

40 min read9 chapters3 labs15 quiz7 scenarios15 interview Q&A

This first module is free: read the lesson and take the quiz. Labs need an account: sign up free to run up to 3 hands-on labs.

Log inStart free trial
This reading material cannot be printed. Read it on config.nwkings.com.
01

Advanced OSPF: what you will learn and the big picture

What you will learn in this module. You will take OSPF from "it forms neighbours" to "I can design, steer and repair it". Over nine chapters you will: read the link-state database the way a professional does, build multi-area designs with summaries and stub or NSSA areas, run OSPFv3 for IPv6, control what enters and leaves OSPF with export and import policy, tune preferences, drain a router with overload, repair a broken backbone with a virtual link, and follow a troubleshooting workflow. These match the OSPF objectives of the JNCIP-ENT exam (JN0-650). Check the current objective list on the Juniper certification page, as objectives can change between exam versions.

Prerequisites. JNCIA-Junos and JNCIS-ENT level OSPF: neighbour states, DR and BDR election, the basic LSA types, single-area and basic multi-area configuration, and the Junos commit model. If you have finished the Start here module you are ready.

Analogy: a country with districts and a ring road

Imagine a country split into districts. Each district keeps a detailed street map of itself only. A national ring road connects the districts. A traveller leaving a district consults a short summary: "to reach district C, take the ring road via gateway G2". Gateways are the ABRs. A gateway to another country, with its own rules and maps, is an ASBR. OSPF areas work the same way: details stay inside the area, summaries cross the border, and the backbone, area 0, is the ring road that every district must touch.

Area 0.0.0.0backbone Area 0.0.0.1transit area Area 0.0.0.2no direct link to area 0 R1 R2 R3 R4 virtual link (chapter 7)

R2 and R3 are ABRs. Area 0.0.0.2 touches the backbone only through a virtual link.

The module map

ChapterQuestion it answers
2 Database and areasWhat does each LSA say, and who knows what?
3 Summaries, stub, NSSAHow do I keep routing tables small?
4 OSPFv3How does OSPF carry IPv6?
5 Export policyHow do outside routes enter OSPF?
6 Import and preferenceWhich routes does this router accept and prefer?
7 Overload, virtual linksHow do I drain a router and repair the backbone?
8 TroubleshootingHow do I find a fault quickly?
9 SummaryChecklist, glossary, cheat sheet.

The lab devices you will meet

The module labs use small Junos topologies. In the Delhi lab R1, R2 and R3 run OSPFv3 across two areas. In the Chennai lab R1 to R4 form a square with a head-office PC and a server LAN, and R4 exports partner routes. In the Hubli lab R1 to R4 sit in areas 0, 1 and 2 and a virtual link is broken. Router IDs follow a pattern such as 10.255.4.2, links use /30 networks such as 10.4.12.0/30.

Worked example. In the Chennai lab R1 learns the server LAN 172.16.4.0/24 with metric 3 and two equal-cost next hops, 10.4.12.2 through R2 and 10.4.13.2 through R3. After you put R2 into overload, R1 shows only the R3 next hop, but the ping to R2's loopback 10.255.4.2 still works. The whole module is about explaining and controlling results like this.

Common mistake. Treating OSPF as "turn it on and forget it". Area boundaries, summaries and policy decide what each router knows. Without a plan, routing tables grow and failures spread.

Exam trap. OSPF area IDs are written like IPv4 addresses, 0.0.0.0 for the backbone. Area 0 and area 0.0.0.0 are the same area. Router IDs and area IDs are separate things even though they look alike.

The merger that created an orphan area

Two companies merged. The new Hubli site was connected behind an existing branch router that sat in area 1. Nobody remembered that every area must touch area 0. Hubli had its own neighbours but head office could not reach it. The quick repair was a virtual link between the two ABRs; the long-term plan was to renumber the areas.

Lesson: a working adjacency does not mean a working design. The area rules matter.

"Why does OSPF need a backbone area?"

Inter-area routing uses a distance-vector-like exchange of summaries through area 0, which prevents loops between areas. Every non-backbone area must attach to area 0 directly or through a virtual link. ABRs summarise routes between areas through the backbone.

Key takeaways

  • This module covers the OSPF objectives of JN0-650: areas, policy, OSPFv3, overload, virtual links, troubleshooting.
  • Details stay inside the area; summaries cross the border; area 0 connects all areas.
  • ABRs sit between areas; ASBRs bring in outside routes.
  • Area IDs look like IPv4 addresses; 0.0.0.0 is the backbone.
  • Every chapter pairs concepts with output from the simulator labs.

Upgrade to keep reading

The free plan lets you read chapter 1 of this module. Upgrade to open the remaining chapters.

  1. Reading the OSPF database: LSA types, scope and path types
  2. Summaries, stub areas and NSSA: keeping tables small
  3. OSPFv3: OSPF for IPv6 on Junos
  4. Export policy: how outside routes enter OSPF
  5. Import policy, preferences and metrics: steering what a router uses
  6. Overload and virtual links: draining a router and repairing the backbone
  7. OSPF troubleshooting workflow
  8. Summary and exam checklist

Join the community