Practices for Scaling ARP and Neighbor Discovery (ND) in Large Data Centers

Approval announcement
Draft of message to be sent after approval:

Technical Summary
As described in draft-ietf-armd-problem-statement, the increasing trend of rapid workload shifting and server virtualization in modern data centers is requiring servers to be loaded (or re-loaded) with different VMs or applications at different times. Those different VMs loaded to one physical server may have different IP addresses, or even be in different IP subnets.   
In order to allow a physical server to be re-loaded with VMs in different subnets, or VMs to be moved to different server racks without IP address re-configuration, the corresponding networks have to have multiple broadcast domains (many VLANs) on the interfaces of L2/L3 boundary routers and ToR switches. Unfortunately, this kind of network can lead to address resolution scaling issues, especially on the L2/L3 boundary routers, when the combined number of VMs (or hosts) in all those subnets is large. 
This document describes some simple practices which can minimize the ARP/ND scaling issues in a Data Center environment. 

Document Quality

The document has incorporated many good practice suggestions from NANOG and IETF operator community. It is readable, accurate on the practices to scale ARP/ND in large data centers.


  Linda Dunbar ( is the Document Shepherd. 
  Ron Bonica ( is the Responsible AD.

RFC Editor Note

  (Insert RFC Editor Note here or remove section)


  (Insert IRTF Note here or remove section)


  (Insert IESG Note here or remove section)


  (Insert IANA Note here or remove section)