7 Reasons Cisco UCS Rocks | #7: Manageability

Continuing in our series on the Cisco Unified Computing System (UCS) and its rise to #1, we now turn our focus to part two of seven. Here's the outline:

1. Abstraction
2. Agility
3. Simplicity
4. Efficiency
5. Scalability
6. Performance
7. Manageability 

Why Cisco UCS Rocks #7: Flexible Manageability

So what good is all this abstraction, agility, simplicity, efficiency, scalability, and performance . . . if you can't manage the stinking thing? That's right, I said it. Thankfully, there is a very mature management platform built into UCS, called UCSM. 

Quick Overview of Cisco UCSM

UCS Manager, or UCSM is the heart of the management engine of Cisco UCS. It gives hardware status, provide software-defined policy engines to apply service profiles to blades and rack servers, and configures all the other necessary policies and items. UCSM is what enables the abstraction layer to dynamically configure your environment with rapid agility to meet your changing business needs.

Cisco UCS Manager | Main Display

Cisco UCSM provides good visual feedback. You will notice at the top left the quick status area (which is click-navigable) that provides instant visual knowledge about anything going on globally. In addition, the navigation tree just below it will also outline the corresponding information or warning level so that you know where the problem(s) are within the environment (especially useful for larger data centers).

Cisco UCSM | Quick Visual Troubleshooting

UCSM Service Profiles

I have already covered in brief some of the benefits of service profiles in my discussion on abstraction and how it makes Cisco UCS a real game-changer for software-defined data centers. 

Cisco UCSM | Service Profiles

You can literally have as many service profiles as you want, which is awesome (a server can only be associated to one profile at a time however). So if you are doing a lot of reconfiguration of your environment, or testing different configs out such as boot from SAN, or changing from FC to NFS, or whatever, you can easily do it and then switch back as a fallback option. Nice!

Cisco UCSM | Sub Organizations

You have to look closely at the picture above (sorry I don't have a demo of one right now), but you can create sub-organizations within UCSM. This is super-fantastic for department-level configurations (such as the NetApp case study below). It makes meeting the requirements of all the stakeholders actually possible. 

208 Servers; 2 Points of Management

I have also already talked a bit about the benefits of service profiles for computing agility and how it enables computing to be a resource for businesses instead of a constraint. Here is one such case example. 

NetApp needed to build a converged infrastructure to meet all their testing requirements—which is basically everything, because their customers run everything. Enter Cisco UCS.

By choosing Cisco UCS, NetApp was able to:

• Deploy 10,000 VMs in <1 hour
  
• Maximize storage efficiency: 500GB for 500 VMs
  
• Achieve 100 times client capacity at same infrastructure cost
  
• Enable one-box NFS, iSCSI, FC, FCoE connectivity
  
• Test over 10GbE, 1GbE, FC, FCoE
  
• Simplify with one wire + automation
  
• Reduce cabling by 78%
  

By the way, that's not an exaggeration. From their previous environment, which was blade-based already, they were able shave an estimated 6-12 weeks off their install and provisioning time in installing and configuring all their servers and virtual machines.

When NetApp needed to deploy a scalable testing cloud capable of hosting 23,000 virtual machines, the company’s engineering support services chose Cisco Unified Computing System. The first step consolidated 714 existing servers onto 120 blade servers in a single Cisco UCS platform, reducing 168 management points to just one: the pair of Cisco UCS 6100 Series Fabric Interconnects.
— Cisco UCS: IT Agility Delivered

Here's a little bit more about the Global Dynamic Lab and what NetApp was able to accomplish with Cisco UCS:

Initial Phase

• (20) FAS3170 pairs for unified storage
• 208 Cisco UCS Blades
• 23,000 VMs

Expansion Phase

• 272 Cisco UCS Blades additional (480 total)
• 30,000 VMs additional (53,000 total, another 69,000 planned)

Oh, and by the way . . . 

• It's completely diskless, and
• Using NetApp FlexClone, 500 VMs took only 500GB of storage. 
  

You can read more, if you like, about the NetApp Kilo Client Initiative, and you can watch an awesome tour of NetApp's Global Dynamic Lab below. 

Beyond One Fabric: UCS Central

But what happens when you need to go beyond one pair of Fabric Interconnects? Doesn't my management get a lot more cumbersome? Nope. Enter UCS Central. 

Cisco UCS Central | Architecture

Think about Cisco UCS Central as a kind of meta-UCSM interface; it aggregates all the separate domains (Cisco's term for a UCSM instance) into domain groups (logical collections of UCSM instances) and allows you to manage them all together if you like, or apply changes individually.

Cisco UCS Central | User Interface

So let's say you have six global data centers all running UCS. You don't have to define software policies for your servers six times; you just do it one and apply it across the board! Even more than that, you can aggregate your fault summaries and other notifications into a single interface, giving you that great visual feedback (and SNMP, email, and all the other usual transports) for your entire geographical infrastructure with a single glance. Nice!

Get more info about UCS Central from this episode of TechWiseTV:

By the way, I was asked this recently: what's the difference between UCS Central and UCS Director? The answer is they are fairly similar, but UCS Director targets unified management for FlexPod, so hardware/software deployment combinations, whereas UCS Central is entirely dedicated to managing multiple pairs of UCS Fabric Interconnects and UCSM domain instances. 

And since this blog post is already phenomenally long, I will just mention—there's even more integration available, especially with VMware. Enter the UCSM vCenter Plugin, and vCenter Orchestrator (vCO) for UCS. 

vCenter Plugin for UCSM

With the vCenter Plugin for UCSM, you can gain a lot of visibility into your configuration as well as perform a host of actions on your hosts:

• Lists available Service Profiles (SP), SP Templates, Server Pool, Host Firmware Pack
  
• Reload UCS Domain
  
• Associate Server pool to SP Templates
  
• Associate Host Firmware Pack to SP Template
  
• Create Hardware or Template based Service Profile
  
• Associate and Disassociate Service Profiles
  
• Manage associated BIOS Policy and modify BIOS settings
  
• Manage Host Firmware Pack
  

Awesome!

vCO Plugin for UCS

With vCO for UCSM, you can automate a bunch of stuff by feeding data into the XML API, such as:

• Application of a service profile to a UCS blade
  
• Cloning a service profile template to a service profile
  
• Service profile customization of frequently modified components
  
• Import and export of service profiles from and to UCS Manager instances
  

Integration and automation with VMware. Nice!