The New Network
Explore Juniper’s vision for network innovation and how the company and industry are shaping the future with the new network

When is a Fabric not a Fabric?

by Juniper Employee on ‎04-05-2011 05:14 PM


On March 30, Cisco announced an update to their data center “vision.”  We are pleased to see that they have endorsed our view of the problem—that the legacy multi-tiered data center network architecture must change to enable a truly virtualized data center.  This evolution will begin with the architecture, where hierarchical tree structures will give way to flat fabrics.  We at Juniper have been promoting this vision for the last two years and Cisco has now joined the chorus.  Where we fundamentally differ is in our view of what constitutes a fabric and how one implements that fabric.  Whereas Cisco is attempting to build a flatter, non-blocking network using their existing and evolving Nexus switches, we have focused on building a fabric by rethinking the concept of the switch.


These different approaches are best illustrated by how each company expresses the “key principles of a fabric.”  In the column on the left in the table below, I have listed the seven key principles of a fabric as defined in a paper published in February by Pradeep Sindhu, the founder of Juniper Networks and the architect behind QFabric.  In the column on the right are Cisco’s five key principles of a “Unified Fabric” as presented on slide 32 during the March 30 announcement.





1.    Any-to-any connectivity with fairness and full non-blocking behavior for all traffic in a data center


1.   Virtual Switches

2.    Low latency and jitter


2.   Hybrid Switches

3.    No packet drops under congestion


3.  Storage over Ethernet

4.    Linear cost and power scaling


4.  Flattening the network

5.   Support of virtual network and services


5.  A new set of network management capabilities

6.   Modular, distributed implementation that is highly scalable



7.   Single logical device




Notice that the list of principles on the left speaks to how a fabric should behave while the list on the right is more about “building” a fabric from specific existing components or capabilities.  This is a very clear delineation between the approaches each company has taken towards building a fabric.  The question is whether customers should care about these differences.   I would argue the answer is yes.  Here are three reasons why.


1.  The last principle on each list attempts to speak to operational simplicity.  Cisco specifies “a new set of network management capabilities” that, per slide 8 in their announcement, delivers “single pane of glass visibility across LAN and SAN.”  “Single pane of glass” is the term used to describe the ability to monitor and/or manage multiple separate devices from a single user client.  This does not necessarily make it simpler, just more convenient.  In Juniper’s case, QFabric is a single logical device that behaves like and is managed as such.  This is simpler and, we believe, fundamental to a fabric.  For customers, simpler is always better as it lowers OPEX and improves availability.


2.  Cisco introduced their new Broadcom-based Nexus 3000 for “low latency” deployments, primarily because all Nexus 5000s are notoriously slow.  However, the Nexus 3000 is not compatible with FabricPath or UCS, it does not support FCoE to FC gateway functionality, and it does not support their physical or virtual FEXs.  Thus you can be fast or you can be in a Cisco fabric, but you can’t be both.  We believe it is not a fabric unless it has low latency and low jitter.  For customers, faster is always better as it improves application behavior.


3.  Cisco announced their new ability to transport FCoE packets across a Nexus 7000.  However, that ability cannot use the FabricPath protocol.  Thus they offer a “unified fabric” for Layer 2 Ethernet traffic, a different fabric for FCoE traffic, and yet another distinct fabric for Layer 3 traffic.  It is not clear what part of the term "unified" they do not understand.  With QFabric, all traffic is inherently supported by the data path of the fabric and, for customers, one fabric is better as it simplifies operations and reduces cost.

The most important take away from the March 30 announcement is the reaffirmation that the fundamental architecture of the network must and will change in order to enable the promise of the virtualized data center.  Even Cisco now acknowledges that this evolution will require a wholesale replacement of most of the existing switching infrastructure as customers move from trees to fabrics.  The existing installed base of Catalyst switches in the data center will be replaced over the next five to 10 years.  The good news is that customers will have a rich set of alternatives to consider as they retool for the future.


by Abhijeet(anon) on ‎04-06-2012 10:23 AM

Could someone please explain what a non-blocking network fabric is ?

First google search took me here, and nothing found.

PLease good guys at Juniper ! Help me :smileyhappy:

by Distinguished Expert on ‎04-06-2012 11:44 AM



By "non-blocking" they are talking about eliminating the need for spanning tree and therefore blocking ports when building out a large port count datacenter.  The current three tier architecture requires that there be redundant port connections for failover.  These then use spanning tree to block the duplicate port and prevent layer 2 loops.


A non-blocking fabric is able to create that redundancy without "wasting" ports by spanning tree blocks, but allowing all the ports to be used and having the redundant paths built into the fabric connections.


Check out this pdf on page 4 for the comparison of the traditional three tier datacenter reference architecture and the Q-fabric solution.  The rest of the document describes how the traditional switch is abstracted to multiple devices that allow this any port to any port connectivity without the use of spanning tree.


The Q-fabric product general information is here.


Steve Puluka BSEET
Senior Network Administrator
MCP 70-290 - Managing Server 2003
MCTS Windows 7


by Abhijeet(anon) on ‎04-06-2012 10:22 PM

Thanks a lot Steve !

Post a Comment
Be sure to enter a unique name. You can't reuse a name that's already in use.
Be sure to enter a unique email address. You can't reuse an email address that's already in use.
Type the characters you see in the picture above.Type the words you hear.
About The New Network

Exploring the vision for the networking industry and the issues shaping its future.

Subscribe to The New Network    RSS Icon

Our Bloggers

Rami Rahim
Juniper Development & Innovation

Profile | Subscribe

Brad Brooks
Chief Marketing Officer

Profile | Subscribe

Bask Iyer
Senior Vice President and CIO

Profile | Subscribe

Judy Beningson
Vice President, Strategic Planning

Profile | Subscribe

Mike Marcellin
Senior Vice President

Profile | Subscribe

Jonathan Davidson
Senior Vice President

Profile | Subscribe

Ankur Singla
Vice President of Engineering

Profile | Subscribe

Bob Dix
Vice President
Government Affairs &
Critical Infrastructure Protection

Profile | Subscribe

Copyright© 1999-2013 Juniper Networks, Inc. All rights reserved.