OSPF Area Designer
Describe the routers, their areas and the link costs. It computes the paths, shows where traffic splits across equal-cost links, and checks the area design against the rules that actually stop OSPF working.
Local toolYour network plan stays in your browser. Nothing you enter is uploaded, logged, or sent to any server.
Routers
Links
Design checks
Path between two routers
The rules that actually matter
Area 0 is the backbone and everything hangs off it. Inter-area routing goes through area 0 by design — there is no path from area 1 to area 2 that avoids it. So every non-backbone area needs at least one router that is also in area 0. That router is an area border router.
The backbone must not be in pieces. Two groups of area 0 routers that cannot reach each other within area 0 partition it, and inter-area routing breaks even though every area looks correctly attached. This is checked separately because it is easy to create and hard to see.
Both ends of a link must agree on the area. If they do not, no adjacency forms at all — the interfaces come up, OSPF says nothing useful, and the routes never appear.
Router IDs must be unique. Duplicates cause adjacencies to fail or flap, and the symptom rarely points at the cause.
A virtual link can attach a detached area to the backbone. It is a repair for a design that went wrong, not a design, and treating it as one produces networks nobody can reason about.
Why areas exist at all
A single OSPF area is a single link-state database: every router holds a full description of every link, and any change is flooded to all of them, each recomputing its shortest-path tree. That is fine until it is not. The cost grows with the size of the area, and more importantly so does the blast radius — one unstable interface in a corner of the network makes every router in it recalculate, repeatedly.
An area boundary stops the flooding. Detail inside an area stays inside it, and what leaves is a summary. A link flapping in area 2 is invisible to area 1 as long as the destinations it reaches are still reachable, so the instability is contained where it happens. That containment, rather than any saving in memory, is the reason to divide a network up.
The stub types, and what each one hides
Areas at the edge of a network usually have only one way out, which makes carrying full routing information into them pointless. The stub variants trade detail for quiet, and they differ in how much they suppress.
A stub area refuses external routes — anything redistributed into OSPF from elsewhere — and replaces them with a default route towards the area border router. A totally stubby area goes further and drops inter-area summaries too, so routers inside it know their own area and a default for everything else. A not-so-stubby area is the awkward case made workable: an area that needs to be stubby but has an external route of its own to inject, which it carries in a distinct LSA type that the border router translates on the way out.
The common thread is that every one of them assumes there is nothing to decide locally — that the way out is the way out. Where an area genuinely has two exits with different reachability, hiding the detail removes the information the routers needed to choose between them.
Related tools
OSPF Cost Calculator explains where costs come from. BGP Best-Path Simulator covers route selection between autonomous systems.