What's the reason for OSPF E2 metric type being the default

[Paheeradan Nagulan]

Any reasons on why "OSPF E2" metric type being the default instead of "E1" in Type 5 LSA when redistributing the default route into OSPF? I understand the if we use E2, the routers within the OSPF area won't be able manipulate the metric in the routing table for that particular route.

[M02@rt37 · Accepted solution]

Hello @Paheeradan Nagulan 

Great question!

The key takeaway is that E2 routes abstract the internal OSPF topology and prioritize external metric values, ensuring predictable path selection regardless of the internal cost to reach the ASBR. This makes E2 ideal for scenarios where the primary concern is egress consistency, such as balancing outbound traffic between multiple ASBRs.

Conversely, E1 routes factor in the internal OSPF cost, leading to more dynamic behavior where routers select the nearest ASBR based on both internal and external costs. This can be advantageous when internal topology changes impact traffic flow, ensuring optimal exit points.

Neither choice is universally superior; rather, the decision depends on whether the network design prioritizes internal path efficiency (E1) or external path consistency (E2).

[Paheeradan Nagulan]

Hi M02@rt37 ,

Understood. Thank you for the detailed explanation!

[paul driver]

Hello
E2 routes does not add the link costs towards the route being advertised just the redistributed cost (default 20), so the cost is the same no mater how many links it routes through to reach the ASBR advertising it

E1 routes - does count these links cost so OSPF is more deterministic is what path it takes and is preferred over E2 routes

[Paheeradan Nagulan]

Hi @paul driver , it makes sense. Thank you for the explanation!

[MHM Cisco World]

it only default value there is no reason like default value of redistribute
dont care about this point 

MHM

[Joseph W. Doherty]

I suspect the default choice was made as to minimize CPU resource demand on the routers.

[Paheeradan Nagulan]

Hi @Joseph W. Doherty , Yes, that's a valid point. I can see why Cisco went that route. Thank you for taking your time to answer my question!

[Joseph W. Doherty]

Not just Cisco, believe it's part of the OSPF RFC.

Often it's easy to forget decades ago issues we would consider non issues currently, were big issues then.

For example, Classful IPv4.

[Paheeradan Nagulan]

Good to know that. I guess that's the beauty of Technological Evolution! 

[Giuseppe Larosa]

Hello  @Paheeradan Nagulan ,

it  is  a good  question indeed.  My   personal preference when multiple ASBR are present is to use O E1 type with route-maps   to  provide granular control on the exit point .

As  suggested by @Joseph W. Doherty there  are historical reasons why O E2   is the  default  external route  type.

With  modern 64bit CPUs used in  the routing engines/route processors but even before the  number of routers in the same  area   has reached hundreds  of routers by using point to point links.

When OSPFv2 was   created  50 routers per area was Cisco recommended   design

Look for Russ White considerations on  OSPF scalabilty   on  his  blog

https://rule11.tech/archive/

Hope to help

Giuseppe

[Joseph W. Doherty]

@Giuseppe Larosa an interesting site, but could not find the OSPF reference you allude to.  Could you provide either the archive page's name or provide reference links to specific articles?

[Giuseppe Larosa]

Hello @Joseph W. Doherty ,

Russ White has written about tests on OSPF scalability with up to 800 routers in the same area here in the netpro now Cisco Community forum in years 2002-2003  I follow him on linkedin like I follow Ivan Pepelnjak and other people.

[Paheeradan Nagulan]

Hello @Giuseppe Larosa , 

I like the idea of using O E1 when we are having multiple ASBRs. I'm going to bookmark the link you posted. So much interesting information to read on that link. Thanks again for taking your time to answer my question.