Cisco Support Community

Chalk Talk: FabricPath, TRILL & VXLAN

Virtualization has dramatically improved utilization within the  datacenter: now, TRILL , FabricPath and VXLAN overlays can do the same  for data center networks.  This Chalk Talk, by Cisco expert Sanjay K.  Hooda, explains how you can use them to distribute data traffic far more  effectively, utilizing virtually all of your network links.

Editor’s Note: TS Newsletter customers get a 35% discount on Sanjay K. Hooda’s book, Using TRILL, FabricPath, and VXLAN: Designing Massively Scalable Data Centers (MSDC) with Overlays. See discount code at the end of this article.


Over  the last several years, Virtualization and cloud technologies have  become increasingly popular. The realization by the Enterprises that the  server resources (including CPU, memory) were heavily underutilized led  to the move towards virtualized datacenters. In virtualized datacenters  every physical server runs one or more virtual server on top of  hypervisor.

Network architects are now faced with ever  increasing need to provide the connectivity to the virtual machines,  which can move from one physical server to another. To address these  requirements, various overlay technologies including TRILL, FabricPath,  VXLAN, NVGRE, SPBV etc. have been established.  In this article, we will  go through high-level basics of FabricPath, TRILL and VXLAN.

Fabric Path

The  trend towards virtualization of physical servers, especially in large  data centers, requires both support for distributed applications at a  large scale and the flexibility to provision them in different zones of  data centers. This necessitated the need to develop a scalable and  resilient Layer 2 fabric enabling any-to-any communication. Cisco  pioneered the development of FabricPath to meet these new demands.  FabricPath provides a highly scalable Layer 2 fabric with a required  level of simplicity, resiliency, and flexibility.

While  preserving the plug-n-play features of the Classical Ethernet, Cisco  FabricPath provides multi-pathing, high availability, forwarding  efficiency by forwarding traffic over the shortest path by discovering  the shortest path using Layer-2 IS-IS and conversational learning to  meet the challenge to work with small MAC address (Layer-2) table sizes  in large Layer-2 domains.

To forward the frames,  FabricPath employs hierarchical MAC addresses that are locally assigned.  FabricPath encapsulates the original Layer 2 frame with a new source  and destination MAC address, a FabricPath tag, the original Layer 2  frame, and a new CRC. To forward the frames in the FabricPath network,  the outer source and destination MAC addresses contain a 12-bit unique  identifier called a SwitchID. The SwitchID is the field used in the  FabricPath core network to forward packets to the right destination  switch.


Transparent Inter-Connection of  Lots of Links (TRILL) is a technology that addresses the same  requirements as the Cisco FabricPath and has almost the same benefits as  FabricPath. Unlike FabricPath TRILL, is an IETF standard that was  proposed to address the various STP challenges including inefficient  utilizations of links, long convergence times, and MAC address table  scaling in the Datacenter networks.

Like FabricPath,  TRILL uses IS-IS as the control protocol with the idea of taking the  advantages of a Layer 3 routing protocol and at the same time maintain  the simplicity of a Layer 2 network.

To forward frames,  TRILL uses a MAC-in-MAC encapsulation format. The ingress RBridge  encapsulates the original Layer 2 frame with a new source and  destination MAC (the MAC addresses of the source RBridge, and the  next-hop RBridge, respectively), a TRILL Header, (which has the Ingress  and Egress nicknames that identify the source and destination RBridge,  respectively), and the original Layer 2 frame with a new CRC. The  incoming 802.1q or q-in-q tag needs to be preserved in the inner header.  Egress RBridge removes the headers added by the ingress RBridge and  will forward them based on the inner frame.


VXLAN  (Virtual Extensible LAN) has become the most popular overlay protocol  today. VXLAN’s support from a number of networking vendors, including  Cisco, VMWare etc., along with the fact that it runs over IP core, makes  VXLAN a popular choice for deployment in massive scale data centers.  VXLAN is probably one of the few overlay protocols that has been  deployed both as a host-based overlay as well as a network overlay

Today,  the primary mode of deployment of VXLAN is with IP multicast. VXLAN  uses VXLAN tunnel endpoints (VTEPs) to map the endpoints to VXLAN  segments. As we can extrapolate, a VTEP has two interfaces -- one  connecting to the IP network and another one to local segment -- which  support bridging.

VXLANs use of IP multicast has led  to manageability issues that have resulted in exploring multicast-less  options. Multicast-less options, which are gaining popularity, should  lead to Head-end replication based VXLAN schemes as the SDN controller  enabled implementations gain traction.

Various virtual  switches natively support VXLAN encapsulation and decapsulation.   Datacenter customers have already been able to rapidly bring up virtual  applications with integrated services by employing VXLAN as a host-based  overlay.

Over the last couple of years there has been  an incredible rise in the popularity of VXLAN. The MAC over IP/UDP  overlay format, along with the potential to support up to 16 million  virtual network segments, makes VXLAN a leading choice for multi-tenant  data center deployments.

From interoperability with the  existing network devices perspective, VXLAN can interoperate with  existing network using VXLAN gateways, and it can be incrementally  deployed, thereby allowing continuous communication with existing nodes.

About the Author


Sanjay K. Hooda is currently Principal Engineer in catalyst switch software engineering  at Cisco. He has 15+ years of network design and implementation  experience in large enterprise environments, and has participated in  IETF standards activities. His interests include wireless, multicast,  TRILL, FabricPath, High Availability, ISSU (In Service Software Upgrade)  and IPv6. Hooda is co-author of IPv6 for Enterprise Networks.


Using TRILL, FabricPath, and VXLAN: Designing Massively Scalable Data Centers (MSDC) with Overlays

By Sanjay K Hooda, Shyam Kapadia, Padmanabhan Krishnan

Published Feb 7, 2014 by Cisco Press


ISBN-10: 1-58714-393-3

ISBN-13: 978-1-58714-393-9

DISCOUNT: 35% off for Cisco Technical Newsletter customers; use code CISCOTECH at checkout.

This article is featured in the January 2014 issue of the Cisco TS Newsletter. Are you subscribed?

Community Member

Good comparison write up, thank you!

Community Member

Great technical article! Now I have to decide if using VXLAN or Pluribus... Both are software protocols instead of TRILL that is hardware protocol.

Somebody has experience with VXLAN and Pluribus?

Best regards,