So, if Green is the new black, what is becoming clearly evident is that Virtualization is the new Green. I was astounded by the fact that after speaking with an old friend this week, he “knew” of virtualization, and it had been on the radar for the past couple of years, but he had no plans to move forward on it, as of yet. None. And, he owns his own hosting company. What a great use case for virtualization. 

Anyway, back to green. Look around, processors are getting denser, with 2, 4 and soon 8 cores per processor, consuming less and less power per core. If the inclusion of “VT (virtualization technology)” at the hardware level by Intel (and somewhat AMD with AMD-v) wasn’t the tell tale sign of the intended use of these servers, then I don’t know what is… Really, do you believe that running a Windows OS on a dual quad-core cpu server (8 cores total), with IIS is the best use of this server’s capabilities? Sure, with Oracle or SQL Server, you could “dedicate” that machine to your needs, if it warranted. (Remember, virtualization rule of thumb: High I/O or High CPU Workloads = Dedicated Server, anything else, and I don’t see a reason not to virtualize.)

Again, back to green. How could virtualization enable Green in my data center, you ask? Quick example:

• Step 1. Take your most powerful, recently purchased (12 – 18 months) server, and pack it with a little more RAM (at least 8-12GB, if possible). 
• Step 2. Back it up, possibly with something like Acronis or Platespin, in preparation to virtualize.
• Step 3. Install and configure VMware ESX 3.0 on it, and start it up. 
• Step 4. P2V (Physical to Virtual) convert several of the servers you have running on similar machines, and load them as virtual machines your new VMware host. Let’s assume you converted 4 servers, just as a starting point. 
• Step 5. Shutdown the original servers. Now check the usage on your metered PDU or UPS. If this is not a shared data center, but a company DC in your facility, then wait 20 minutes, and check the temperature in the room. Is it down? Even one degree is a significant atmospheric change. 

Like your results? Well then, repeat steps 2 – 5, as many times as you would like.

Feel the green glow in the room getting stronger? You may be surprised, but those aren’t plants growing in your data center. That’s money pilling up in the corner as you reduced your power, cooling costs AND x86 server costs. 

Have you had a similar experience, or do you disagree? Tell me what you think.

- clikdude

Hmmm… I pondered those words for a while. Back then, I was dealing with issues around the procurement of physical server resources, and uniform provisioning of those resources in the environment, while maintaining order and manageability in the infrastructure. Risk management, and more specifically backup and recovery, and disaster recovery controls and measures were paramount, and considered the priority of the day. The features of ESX I saw were compelling.

Virtualization, as presented with all of the features offered in the ESX platform, brought a change in my mindset for many areas, including backup and recovery via snapshots, redundancy and high availability via vMotion and HA, and heterogenous systems manageability (removing vendor lock-in at the hardware level). But, his words stuck with me: No “Stupid IT Tricks” with virtualization.

Fast forward many years. I was there, last week, sitting in a locally promoted presentation on Microsoft’s Hyper-V platform (their new hypervisor) for a Microsoft User Group. The presenter seemed knowledgeable, and at times, extremely enthusiastic about virtualization and its benefits. This is probably why he felt compelled to start at the beginning, and talk about virtualization as a whole, and not just in the context of Microsoft. He did a decent job… in the beginning.

His enthusiasm turned dangerous, when, to a crowd of about 30 individuals (many who seemed nascent to the area of virtualization hands-on) he began shouting about how “snapshots” of Virtual Servers totally replaced backup and recovery as we know it! (Personally, I was stunned.) Literally, he stated “look at this, with snapshots, backup and recovery of files is old school. I can just take a snapshot of a server, running anything…SQL Server, Exchange, and other services, while they are running, and if there is a problem, I can revert back…like this. Whola!” His demo of snapshots and reverting of a plain vanilla install of Windows Server 2003 in a VM worked, of course. It was plain vanilla.

What’s the problem, you might say? He did not mention the need to quiesce those services, like SQL or Exchange, while the snapshot was taken in order to maintain data integrity, and prevent corruption. And for such block type data, having a true data backup via specific agents serves a different function then just a snapshot, which is best served as a disaster recovery solution. How many of those individuals left that presentation thinking “wow, I can get rid of ArcServe and its agents…”, or worse, “no more tape drives or near line storage for backup versions”. 

He continued the presentation, and made another grave error in judgement, in my opinion. He said “clustering as you know it, and as you need it, is gone!”. He described, very loosely, the clustering of the Hyper-V servers (similar to an ESX farm), and how you could “quick migrate” a VM from one server to the other. And, he added, if a VM goes down, its “data files are copied to another server in the cluster, and automatically started”. Now, I will admit, I do not know much about Hyper-V, so hopefully, his rough description of this process results in a similar process as that of VMware HA. Nonetheless, he went on to say that “Microsoft Clustering, SQL clusters and Exchange clusters are a thing of the past”. Wow.

Again, clusters at the application services level serve a different purpose than clusters at the utility computing level, as in ESX and Hyper-V. I know I have always err’d on the side of having critical application services clustered, whether via VM, physically or both, because other than VMotion, the amount of time necessary to restart that service proved an unacceptable RTO (recovery time objective) for the business.

He continued on, and again, his enthusiasm was great. He loves virtualization. But, the issue I have, is “does he understand what he should and should not do with it and the features a virtual platform provides?”

Virtualization is a great, and probably the best, strategy for x86 infrastructures of the 21st century. However, the key is this: “Virtualization is NOT an excuse for ‘Stupid IT Tricks’.”

What do you think? Let us know and comment below.

- clikdude

As you read this, you may be asking yourself, why another “virtualization” blog? Aren’t there at least a few hundred thousand blogs (that’s an exaggeration) out there on this technology subject already? As a matter of fact, yes, there are. Many of the top blogs on virtualization are written by well respected members of the IT and media communities, and they offer the latest technology news out there on this subject. (I will be linking to several of them shortly.)

So, if that’s the case, why create this blog? Well, simply put, I hope many of you, business and tech geeks alike, will take the opportunity to read this blog, and use it as a resource to educate, and if necessary evangelize, your senior IT management and business operations counterparts on the merits of taking this technology seriously.

Many, including some CIOs I know, still look at virtualization as a “fringe” technology. How, or why, after nearly 10 years of the rapid evolution of this technology in the PC server world, would this still be the case? Well, as IT leaders and innovators, we haven’t done an effective job of communicating the real value proposition to business management. Most virtualization deployment or conversion projects in many IT departments fail to bring about the value for which they were originally signed off on.

In fact, as with many mis-directed “good intentions” in IT, virtualization deployment or conversion projects add a level of burden, uncertainty and volatility into the infrastructure, resulting in yet another “perceived” failure, which business operations must bear, now that IT is fully “committed” to it.

Want examples of their reactions to virtualization?

You have probably experienced one, or many, of these:

• Development departments blame poor performance of their applications, rightfully or wrongfully so, on the new “Virtual Servers”, rather then doing sufficient analysis to determine the correct cause.
• Sales and Marketing departments blame their failure to reach the proposed level of targeted responses on the web site which NOW seems to be “slower” in this new virtual infrastructure.
• P&L business leaders blame their missed goals on the IT department’s insistence on “converting a perfectly good server infrastructure to virtual”, which has added no real value, but rather resulted in 80% or higher human resource re-allocation during the project, causing less focus on the areas that could’ve had real impact on the bottom line, according to them.
• Accounting finds that the ERP package, now resting on brand new virtualized infrastructure, slows to a crawl several times throughout the day. The unprepared and untrained IT group, managing the “systems running on the virtual infrastructure”, simply states that “all of the Windows OS servers and Databases servers are running at normal levels, so it is not the systems”. Frustration in Accounting sets in, since they, as the customer, know that there has been an obvious change as to how their application performs.

This is not the case everywhere. But, after the last 5 years of mainstream acceptance of these virtualization platforms, starting with VMware, virtualization adoption is still at its infancy. Whether caused by an influx of products and platforms to choose from, causing confusion and fear of “vendor lock-in”, or a miscommunication of its real value and impact to the infrastructure now and in the future, CTOs and IT Directors are finding it more and more difficult to introduce this technology into the environment where it can make the greatest impact: Production.

So, hopefully, together with your comments and real-world experiences, we can work together to accelerate the acceptance and adoption of this technology, because for those of us who have experienced, virtualization is truly transformational, in all aspects of IT.

In the current economic conditions, most CFOs, CEOs and CIOs are asking the same question: How can this proposed deployment of, or conversion to, a Virtual Infrastructure impact the bottom line 12, 18 and 24 months from now?

I am confident that together we can answer this question, and many others, in a pragmatic fashion.

This is the state of the virtual infrastructure today.

-clikdude