Category: fabric

One of the attendees of the Building Next-Generation Data Center online course solved the build small data center fabric challenge with Virtual Chassis Fabric (VCF). I pointed out that I would prefer not to use VCF as it uses centralized control plane and is thus a single failure domain.

Here are his arguments for using VCF:

In the market overview section of the introductory part of data center fabric architectures webinar I made a recommendation to use larger number of fixed-configuration spine switches instead of two chassis-based spines when building a medium-sized leaf-and-spine fabric, and explained the reasoning behind it (increased availability, reduced impact of spine failure).

One of the attendees wondered about the “right” number of spine switches – does it has to be four, or could you have three or five spines. In his words:

Evil CCIE concluded his long list of leaf-and-spine fabric myths (more in part 1 and part 2) with a layer-2 fabric myth:

Layer 2 Fabrics can't be extended beyond 2 Spine switches. I had a long argument with a $vendor guys on this. They don't even count SPB as Layer 2 fabric and so forth.

The root cause of this myth is the lack of understanding of what layer-2, layer-3, bridging and routing means. You might want to revisit a few of my very old blog posts before moving on: part 1, part 2, what is switching, layer-3 switches and routers.

The next set of Leaf-and-Spine Fabric Myths listed by Evil CCIE focused on BGP:

BGP is the best choice for leaf-and-spine fabrics.

I wrote about this particular one here. If you’re not a BGP guru don’t overcomplicate your network. OSPF, IS-IS, and EIGRP are good enough for most environments. Also, don’t ever turn BGP into RIP with AS-path length serving as hop count.

Apart from the “they have no clue what they’re talking about” observation, Evil CCIE left a long list of leaf-and-spine fabric myths he encountered in the wild in a comment on one of my blog posts. He started with:

Clos fabric (aka Leaf And Spine fabric) is a non-blocking fabric

That was obviously true in the days when Mr. Clos designed the voice switching solution that still bears his name. In the original Clos network every voice call would get a dedicated path across the fabric, and the number of voice calls supported by the fabric equaled the number of alternate end-to-end paths.

As I explained in a previous blog post, most leaf-and-spine best-practices (as in: what to do if you have no clue) use BGP as the IGP routing protocol (regardless of whether it’s needed) with the same AS number shared across all spine switches to implement valley-free routing.

This design has an interesting consequence: when a link between a leaf and a spine switch fails, they can no longer communicate.

One of the attendees of my Building Next-Generation Data Center online course tried to figure out whether you can build larger broadcast domains with VXLAN than you could with VLANs. Here’s what he sent me:

I’m trying to understand differences or similarities between VLAN and VXLAN technologies in a view of (*cast) domain limitation.

There’s no difference between the two on the client-facing side. VXLAN is just an encapsulation technology and doesn’t change how bridging works at all (read also part 2 of that story).

Here’s a question I got from an attendee of my Building Next-Generation Data Center online course:

As far as I understood […] it is obsolete nowadays to build a new DC fabric with routing on the host using BGP, the proper way to go is to use IGP + SDN overlay. Is my understanding correct?

Ignoring for the moment the fact that nothing is ever obsolete in IT, the right answer is it depends… this time on answer(s) to two seemingly simple questions “what services are we offering?” and “what connectivity problem are we trying to solve?”.

Evil CCIE left a lengthy comment on one of my blog posts including this interesting observation:

It's always interesting to hear all kind of reasons from people to deploy CLOS fabrics in DC in Enterprise segment typically that I deal with while they mostly don't have clue about why they should be doing it in first place. […] Usually a good justification is DC to support high amount of East-West Traffic....but really? […] Ask them if they even have any benchmarks or tools to measure that in first place :)

What he wrote proves that most networking practitioners never move beyond regurgitating vendor marketing (because that’s so much easier than making the first step toward becoming an engineer by figuring out how technology really works).

One of my subscribers wondered whether it would make sense to build a traditional leaf-and-spine fabric or go for Cisco ACI. He started his email with:

One option is a "standalone" Spine/Leaf VXLAN-with EVPN deployment based on Nexus equipment. This approach could probably be accompanied by some kind of automation like Ansible to ease operation/maintenance of the network.

This is what I would do these days if the customer feels comfortable investing at least the minimum amount of work into an automation solution. Having simpler technology + well-understood automation solution is (in my biased opinion) better than having a complex black box.

When I created the Data Center Infrastructure for Networking Engineers webinar, I wanted to reach these goals:

• Understand the data center acronym soup;
• Build a conceptual framework of the data center technologies and solutions.

Every now and then I get feedback from a happy attendee telling me how the webinar helped them. Here’s what I got earlier this month:

One of my readers asked me a question that came up in his business strategy class:

Why did routers and switches end up being vertically integrated (the same person makes the hardware and the software)? Why didn't they go down the same horizontal path as compute (with Intel making chips, OEMs making systems and Microsoft providing the OS)? Why did this resemble the pre-Intel model of IBM, DEC, Sun…?

Simple answer: because nobody was interested in disaggregating them.

I got this design improvement suggestion after publishing When Is BGP No Better than OSPF blog post:

Putting all the leafs in the same ASN and filtering routes sent down to the leafs (sending just a default) are potential enhancements that make BGP a nice option.

Tony Przygienda quickly wrote a one-line rebuttal: “unless links break ;-)”

Using EBGP instead of an IGP (OSPF or IS-IS) in leaf-and-spine data center fabrics is becoming a best practice (read: thing to do when you have no clue what you’re doing).

The usual argument defending this design choice is “BGP scales better than OSPF or IS-IS”. That’s usually true (see also: Internet), and so far, EBGP is the only reasonable choice in very large leaf-and-spine fabrics… but does it really scale better than a link-state IGP in smaller fabrics?

As part of his “how does Avaya implement data center fabrics” presentation, Roger Lapuh talked about use cases for SPB in data center fabrics.

I have no idea what Extreme decided to do with the numerous data center fabric solutions they bought in the last few years, so the video might have just a historic value at this point… but it’s still nice to see what you can do with smart engineering.