Cisco UCS From the Ground Up – A Beginner’s Guide Part 5

In my previous post I described how you can achieve 80Gb per blade with a UCS system. I went a little deeper than I had intended, so I thought it would be good to give readers another explanation from another blogger’s point of view.

How UCS achieves 80GbE of bandwidth per blade

With the launch of the third generation of Unified Fabric, Cisco’s UCS system has moved well ahead of the simple 2*10GbE interfaces that a single blade can have. It now delivers up to 80GbE to a single blade in a dual 40GbE way. How is that achieved? What are the options?
There are four main parts that you need to understand in order to answer the above question.

  • Backplane
  • IO Modules
  • Blades, the M2 and M3
  • Adapters

You will see in the following pictures and text an A-side and a B-side. The data plane is split into two completely separate fabrics (A and B). Each side consists of a Fabric Interconnect (UCS6100/6200) and an IO-Module (2100/2200). And each blade is connected to both the A and B sides of the fabric. This is a full active/active setup. The concept of active/standby does not exist within the Cisco UCS architecture.

Continue reading

Cisco UCS From the Ground Up – A Beginner’s Guide Part 4

OK, so in Part 1 and Part 2 I gave high level overviews of the UCS components. In Part 3 I did it all over again, but went a little deeper that time around. In this post, I’ll be doing it again, but this time going even deeper.

Why you ask? Because diving too deeply into UCS too quickly is enough to confuse anyone. By going over the same components over and over again, getting a little deeper each time, I’m hoping to avoid causing confusion. I feel doing it this way is a better approach rather than writing one overview post then diving into why using a VIC 1240 and VIC 1280 may not necessarily give you 80Gb of bandwidth unless you’ve an appropriate FEX/IOM and enough FI uplinks :)

80Gb of Bandwidth!?!

Yes, that’s right. With the right combination (I can’t stress that enough) of equipment, your blades will be able to achieve 80Gb of bandwidth each.

The reason why I’m stressing this point is because I often see unfortunate situations where people have purchased parts which are capable of 80Gb, but they end up being confused and disappointed when they see that they’re only achieving a fraction of that. I hope to prevent that from happening to others in the future with the next few sections of this post.

Continue reading

Cisco UCS From the Ground Up – A Beginner’s Guide Part 3

In Part 1 & Part 2 I provided a high level overview of the UCS components. In this post I’m going to delve deeper into each of these components as well as explain how they connect to one another. When connected correctly, you’ll end up with a fully functional and fully redundant UCS system. That’s right, by fully redundant I mean there is no single point of failure anywhere in the entire system.

I’ll start with the “deepest” part of the system (the blades and their VICs) and then work my way “out” to the furthest part of the system, the Fabric Interconnects (FIs). It may help to keep an eye on the below diagram (taken from this Define the Cloud post) to track which part of the topology I’m talking about.


Note: As newcomers sometimes struggle with B-Series servers and the way in which data is passed between them and the UCS Chassis, most of what I write from this point on will be about these servers.

Continue reading

Cisco UCS From the Ground Up – A Beginner’s Guide Part 2

In my previous post I covered Wikipedia’s and my versions of an overview of the UCS platform. For good measure, here are a couple of additional “Overview” information:

UCS Overview #2

UCS Overview Video

How UCS Differs

  • Reference: YouTube: Cisco UCS Servers Big Picture Differences
  • Traditional blade servers require their own SAN and network adapters. UCS chassis connect to a Fabric Interconnect for this connectivity. This results in less adapters having to be purchased as well as less cabling which results in cost savings.
  • Each traditional blade servers needs to be configured to work with an overarching management system. UCS servers are managed by the Fabric Interconnect they connect to.
  • Service Profiles are the fundamental mechanism by which the Cisco Unified Computing System models the necessary abstractions of server, storage, and networking.

Continue reading

Cisco UCS From the Ground Up – A Beginner’s Guide Part 1

My NetApp From the Ground Up series of posts have proven to be quite popular. Though I still have quite a lot of posts left to write in that series, I thought I’d start a UCS series as well.

As with the aforementioned NetApp series, I’m not going to re-invent the wheel and will instead reference the excellent documentation and information which others have already put together.

You may be wondering “if there are already excellent resources out there, why bother blogging about it?”. Well, it’s for two reasons:

1) I wasn’t able to find a structured guide that took readers right from the UCS basics through to more advanced topics, so I thought I’d attempt to fill that gap. 

2) Everyone has their own way of explaining things. While I have thoroughly enjoyed reading other peoples material, there are bits and pieces which I feel could be explained a little differently in order to give newcomers a better understanding.

So without further ado, let’s get started!

Continue reading

Cisco UCS From the Ground Up – A Beginner’s Guide – Index

Below is a list of all the posts in the “Cisco UCS From the Ground Up – A Beginner’s Guide” series:

Continue reading