What Makes A Cloud?
The most current NIST definition of Cloud Computing states that, in order for you to have a computational “cloud”, you have to have the following essentials:
- On-demand self-service
- Broad network access
- Resource pooling
- Rapid elasticity
- Measured service
So how does the new suite of tools coming in System Center 2012 map to those items? That’s what I wanted to document for you today in Part 4 of our “30 Days of Cloud on Your Terms” series.
On-Demand Self-Service / Broad Network Access
Many of the System Center 2012 tools will provide an administrative console that will display and allow only the features and functions that you have granted, based on the role of that person. That’s great for your administrative team, or maybe for some “Executive Mode” (read-only) use. But ideally you want to have web console portal access for anyone and everyone. For the sake of requesting services in your cloud, System Center Service Manager 2012 is what you’ll want to use. SCSM 2012 allows you to create and manage a self-service portal that you’ll populate with items coming from your Service Catalog.
“What’s a Service Catalog”?
It is a catalog of services.
I know… I’m simplifying here a bit, but the idea is that you probably have a menu of services for your users. It would be great to give that list to them in a way that they can see and pick from through some self-service portal. You can use Service Manager to define and customize these items with fields for requesting all of the information that the service will require. Once the user submits the request, it moves on down (or up) the chain.. perhaps as a work item for someone on your team, or even as the launch-point for some automated task. Ultimate the task is completed, the request or ticket is closed, and the results are recorded. (More about that later.)
The other tool that is coming for this kind of self-service, specifically focusing on private and public cloud management, is System Center App Controller 2012. App Controller connects to and allows you to manage (or delegate management to) all of your clouds – whether they be defined as services (groups of one or more virtual machines representing “apps”) in System Center Virtual Machine Manager 2012, or your cloud-based services (“apps”) and resources hosted in Windows Azure. As with Service Manager, what your user is allowed to do in the portal is limited to what you have allowed them to do, based on the role they play.
Resource Pooling / Rapid Elasticity
As I am sure most of you know, the notion of automation is an important one. When we note which tasks and business processes we perform on a regular basis, and if we’re able to automate – take out the manual steps that cause delays or are prone to human error – we can dramatically improve timely, accurate service. To support our private cloud there are many ways we can automate. PowerShell is one that immediately comes to mind for many of you. And being able to drive ALL configuration, management, and operations of ALL of our server products, virtualization, and System Center tools using PowerShell is a very powerful solution.
“But what if I’m not a very good scripter?”
The good news for you is that, whenever you perform some operation in one of our server products, there is usually a “View Script” button that you can click, so that you can view and learn from (or repurpose) the PowerShell commands that performed that operation.
But what about more complex operations that involve multiple steps, multiple tools… even solutions that don’t support PowerShell? And wouldn’t it be great if you could just draw a diagram that charts out each step, with all inputs and outputs moving on down the line, launching processes and recording results on the way?
“Yeah! That would be awesome!”
I thought you’d like that. That is the purpose of System Center Orchestrator 2012. In Orchestrator, you create “run books”; which are lists of tasks and the order in which they will be launched.
CAUTION: Take advantage of the capability in Orchestrator to TEST your run books. This tool is one that makes it not only easy to shoot yourself in the foot, but to do it rapidly, over and over again. You don’t want to be the guy who launches the run book that captures all servers from out of Active Directory and then shuts them all down.. automatically. You’ve been warned.
Another tool that supports the dynamic nature of our cloud is System Center Operations Manager 2012. OpsManager has been around for awhile (remember MOM?), and with every version we’ve added support for monitoring more and more devices, device types, servers, services (“apps”), and platforms. With the 2012 release, for example, we add broader support for detecting and monitoring your network devices, and have improved DEEP monitoring of applications (the product that was AVICode is now functionality that is included in SCOM 2012).
For your cloud, getting a notification that tells you something is wrong – wherever it is – is important. And then potentially launching some automation (Orchestrator? or PRO?) to resolve and then report on the issue and it’s resolution is a key piece of a dynamic cloud.
But let’s talk more specifically about elasticity…
Scale up. Scale down. Scale out. Scale in.
I have to assume it’s the opposite of “scale out”. But whatever we call it, the idea is that we need to be able to adjust our service / capacity / scale for whatever reason comes out way; whether it be as simple as a user request for more app storage, or as complex as scaling out and re-shifting massively virtualized services that span multiple virtualization hosts and even out to the public cloud. When the request comes, either resulting from a manual request or via some automation (Orchestrator?) that was triggered based on some performance threshold being overstepped (Operations Manager?), we need to do the work.
Let’s bring it ‘round full-circle, shall we? Once some (or any) operation happens in our cloud, we want the ability to track it. Here’s a question for you: How are you keeping tabs on your SLAs?
“Service Level Agreements? Um.. we’ve defined them, but reporting on how we’re doing against them, or tracking where we’re having problems.. that’s a tough one.”
Exactly. And that’s why I said we’ve come full-circle. System Center Service Manager 2012 – the tool that was the beginning of our work in the form of service requests entering the system, is also the place that tracks the completion of those requests. And along with that, it now allows you to define your SLAs and then to keep track of and report on how you’re doing against those. Worst case: It’s a way to quickly learn where you may have areas to improve. Best case: It’s a great way to prove that you’re keeping your promises to the business, justifying the IT spend, and adding real business value. And that’s really what it’s all about, after all.
Another key improvement in System Center 2012 is in the amazing things you can do based on the integration of the products. I haven’t yet touched on System Center Configuration Manager 2012, but like previous versions, SCCM collects and maintains a very rich inventory of your infrastructure resources. What if you were able to combine that information with the collected monitoring details from OpsManager and the service history recorded in Service Manager? The addition of the Configuration Management Database (CMDB) as the central connector and repository for information becomes
“System Center 2012 is still a ways off, isn’t it? What do you have that I can start using today?
Short of purchasing your own Windows Azure Appliance, Microsoft’s current solution for private cloud involves the current set of System Center tools, and leans heavily upon System Center Virtual Machine Manager 2008 R2 SP1 along with the SCVMM Self-Service Portal 2.0 SP1. For more information on that solution, I refer you to my blog post here.
And yes, the date of the release of System Center 2012 hasn’t been announced, but every part of the suite is now currently available as either a beta or a release candidate. I blogged about it HERE, and you can download one or all of these pre-releases by going HERE.
Did you find this information useful? Do you have any questions? Please post a comment and we can discuss it further.
Keep watching this blog, plus the blogs of my good friends Brian Lewis, John Weston, and Matt Hester, for additional posts in this series. And if you have missed any of the series posts, check out my summary post for links to all of the articles available so far at https://aka.ms/cloudseries.