Question about traffic engineering/steering for internal borders

[gerry.schmucker]

Hi SDA Experts,

we  have a  streched serverfarm between two buildings, now we connected this serverfarm with 2x Internal-Border per Building all belonging to the same fabric via eBGP and manual L3-out using sub-interfaces.

The serverfarm network is now reachable from the Fabric, however all 4 Internal-Borders have the same LISP Pri/Weight, which means Clients connected to Edge-Nodes may use the Internal-Borders in the other Building instead the local one.
Is there a way to optimize this behaviour ?  

Thank you,

Gerry

[jedolphi]

Hi Gerry, we do have some mechanism in LISP to find the "closest" egress router for a given destination, but it's only for some SD-Access Transit (between fabric sites) use cases, not within a single Fabric Site. Your use case cannot be solved without some code changes. Presumably a single Fabric Site has high bandwidth links meaning latency over any given path (FE1->Border1->DC or FE1->Border2->DC) would be more or less symmetrical. A few kilometers extra distance traveled is negligible at the speed of light ☺. Best regards, Jerome

[gerry.schmucker · Accepted solution]

Hi Jerome,

In this case we have to live with the sub-optimal routing I guess.... 

thanks for your quick reply and confirmation!
Regards,

Gerry