Cloud migration projects continue to garner interest from organizations seeking a variety of benefits. Executives understand that these complex projects require multi-disciplinary teams to increase the odds of success and realize the promised ROI (return on investment).

Unlike piecemeal approaches, a data center move to the cloud presumes you will decommission your existing data center within a reasonable time-frame. It also presumes that you will follow through with necessary steps to realize cost savings such as:

• Actually retiring physical infrastructure
• Actually reducing IT (Information Technology) staff where necessary
• Actually realigning IT roles and processes to provision, monitor, and service cloud-based infrastructure
• Abandoning legacy applications in favor of modern, agile applications that are cloud-ready
• Allocating outside resources to get evaluations done in a timely manner versus internal-only evaluations that commit shot clock violations with regularity

Let’s assume you’ve properly evaluated cloud vendors for security, financial stability, and desired service level agreements. What makes a wholesale move to the cloud different than migrating to a physical data center?

Quite a few things will be different, but let’s discuss just three elements that should command your due diligence.

Performance will be different

It might be better or worse and you certainly don’t want to learn this after you move. You’ll need to model, measure, and test all of your applications before moving. Different applications will require different approaches to getting the performance you want.

It’s likely that your monthly bandwidth budget will need to increase and you might have to spend money on WAN (Wide Area Network) optimization and redundant networks. Most organizations have no idea what performance demands exist with their current applications because they’ve never measured them. Generalizing versus measuring is certain to yield unpleasant surprises.

Incident Response will be different

When you rely on a matrix of service providers, your response to incidents should be well thought out and even rehearsed per application. Your cloud providers may not be able to give you satisfactory estimates of time to fix or even the reason for outage.

Some companies don’t even have a formal Incident Response process documented. Such ad-hoc approaches will be exposed as amateur when you move to cloud. If you don’t revise (or build) your Incident Response plan prior to moving to the cloud, you’re just asking to do it one outage at a time.

ROI will be different

Like a weight loss program, the ROI advertised is not typical and your results will vary. Plan on duplicate carrying costs while you work out the kinks. Plan on application development projects to move in-house or legacy applications to cloud-friendly platforms. Plan on organizational training and education to change the way your company expects to use IT infrastructure. Plan on cultural changes to your IT staff that could involve re-staffing the right kinds of skills. Plan on resistance and clinging to physical infrastructure that will decrease your ROI.

What about a roll-back’s effect on your ROI? Yup, plan on a roll back if the cloud affects your revenue or reputation. Of course, testing minimizes roll-backs and ROI will elongate with more testing time.

Moving a data center to the cloud is a multi-disciplinary project that requires project management, a focus on business objectives, and a deep understanding of the applications being moved. Devote the budget and resources to understand how to plan and execute this move for a successful outcome.

Maximize your move to the cloud with governance and oversight and a healthy does of realistic expectations that the change will introduce.

I’ve reviewed many data center relocation plans. Some are simplistic and some are complex, but most suffer from a single and fatal flaw — Everything has to work correctly in the prescribed order. Everything has to go right.

Is that a likely event in the world you live in?

Think about it in the context of an airline flight. Suppose there’s only one, linear sequence of events that will result in a successful landing of the plane. As long as nothing along that event timeline misfires, the landing will be just fine.

Are you ready to take off knowing that everything has to go right to successfully land?

Yet, data center moving project managers first build a plan where everything has to go right and then expend extraordinary effort tracking in minute detail everything that goes wrong. It’s not unusual to consume a disproportionate share of resources and costs because of this flaw.

Like landing an airplane, the wrong time to ask about alternatives is during final approach. As a leader, you have to recognize this planning flaw and counteract it.

Is there a better way?

Maybe, but it might depend on your own built-in biases for using or refusing outside help. It might depend on your willingness to recognize the confirmation bias that sabotages most internally-led moves. It certainly depends on the schedule pressure you’re already experiencing.

Confirmation bias is the seeking or recalling only the evidence that supports your beliefs. It goes almost unchecked with Information Technology (IT) staffs because they naturally gravitate to their strengths and discount and overlook elements that are not in their wheelhouse. Data center moves are not normal events for most IT staff.

The methodology is flawed.

However, the fundamental issue is the method almost all use to create a complex plan. This method is its own form of confirmation bias and the resultant plan holds a tyranny over the project. Consider how project managers are taught to create plans:

• break the work into smaller tasks with start and stop dates (the work breakdown structure or WBS)
• sum up the resulting WBS pieces into the project schedule plan
• track progress against the resulting plan

What kind of a plan have you created? In reality, you’ve not created a plan at all. When you sum up a schedule based on the pieces of the work breakdown, you’ve created a guess about the future. A hypothesis of how the plan might be completed. Your guess isn’t based on deep experience with data center moves and the trouble starts almost immediately.

What happens when reality doesn’t match your hypothesis?

In most cases, you do two things:

1. You tell offenders to try harder to do exactly what the plan says.
2. You re-plan.

And those two things consume scarce resources while doing nothing to improve the chances of success.

The plan holds a tyranny over actually completing important work. Instead of recognizing that your hypothesis about the future has been proven wrong in the face of reality, extraordinary effort is expended to tame the turbulence with a cycle of blame and re-planning.

How do you overcome this plan tyranny?

Consider the story of how the Empire State Building was constructed in the era before computers. There was no WBS (work breakdown structure) constructed and then summed up into a schedule. The key success factors for the 18 month construction of the Empire State building were:

• Deeply Experienced Builders were used – The building tenants didn’t design the building, experts did.
• Constraints Drove Schedule – The concept of time boxing (what can be done in this time-frame) versus scope boxing (how long will this scope take) was used.
• Dependencies were Broken not Tracked – Figuring out how to schedule dependencies takes far more effort than figuring out how to break the dependency.

When you hand the data center moving planning over to internal staff with built-in confirmation bias and no deep experience moving data centers, should you be surprised to end up with a plan that holds a tyranny over your project?

Your Plan is Still Just A Guess!

The reality is that most WBS-built plans, while impressively detailed, are still just a guess about the future. A guess that is quickly rendered incorrect with a corresponding increase in measurement activities to right the plan.

A merchant operated a market at the top of a hill. She relied on vendors to supply the market who pulled their wagons up the hill several times a day. Wanting to increase sales by decreasing the time it took for the wagon trips, she hired a project manager. However, the wagon trips took even longer! She asked a vendor to explain the problem.

“I could pull it faster if you can get that project manager out of my wagon,” he explained.

When you burden the people responsible for the work to track the hypothesis about the future that has already been proven wrong, you are simply expending resources on the wrong effort. And the tyranny of your plan promises to continue consuming scarce resources at alarming and accelerating rates.

Plans based on experience are more reliable whereas plans based on wishful thinking, the guesses generated by WBS-based methodology, are not.

Get your data center moving plan reviewed by experts with deep experience to break the tyranny of your plan built from a methodology proven to be flawed.

Even with all the information about proper hot and cold aisles, dueling CRAC (Computer Room Air Conditioners) units, and improper return air dynamics, IT (Information Technology) teams are still operating inefficient data centers with glaring problems.

Compounding this problem is the fact that facilities managers and IT mangers rarely collaborate to understand the data center as a system. Jon deRidder, a veteran of data center audits and CEO of Enabled Energy, provides some insights into critical data center improvements in this interview.

Jon brings experience in energy conservation, airflow management, due diligence, data center layout design, and remediation for data center environments.  He has served as a subject matter expert for the State of Colorado in the area of environmental contamination.

You work within many industries to improve their data centers. What do you find is the top misconception held by CIO’s with respect to their data center?

There are two common misconceptions that I see.  These misconceptions stem from how an organization views IT.

The most common misconception in an organization where IT is viewed as a cost center is that the data center is also a cost center only.  In this economy it is wise to consider turning the data center into your organization’s secret weapon.  Reducing the Cost of Goods Sold while improving speed and agility will give an organization a coveted advantage over the competition.

In organizations where IT is viewed as a critical component to the organization’s health and well-being, the misconception is that the data center manager knows everything that is going on in the data center.  There is a great need for improved visibility when managing these complex environments.  Pervasive monitoring systems along with the appropriate training and knowledge of industry trends will save any CIO’s data center significant amounts of time, money, and effort.

Can you elaborate on the first steps a CIO should take to get away from operating a data center as a pure cost center?

The first step to moving in this direction should be to form an understanding of the actual costs involved in running the data center, and then create an internal billing system that allows for IT to bill (internally) for their services.  This will set the stage for meaningful investments to be made not only in the infrastructure, but also the process and human resources managing the center.

Even in these organizations that view IT as a critical component, using outside help encounters resistance because the norm is to search for free information and implement point solutions. What kind of time and effort can you save these kinds of organizations?

Implementing point solutions has, in many cases, made the actual or original problem worse.  “On-boarding” key stake holders in the data center takes a tremendous amount of time along with constant, consistent and clear communication.  Most often times we (Enabled Energy) are brought in as the translator to help everyone understand the mission at hand and how everyone is going to receive benefit.

Is it fair to say that despite all the “green” hype, many data centers are blind to how much power they are using? Where do you find the low-hanging fruit of power waste in the data center?

Absolutely, energy usage metrics like PUE (Power Usage Effectiveness) and CUE (Carbon Usage Effectiveness) are now accepted as industry standard, however PUE and CUE are not fully understood and are therefore not as useful as they could be.

Where is the lowest hanging fruit in an existing data center?

• Most data centers are dramatically over-cooled
  • Raise the set point in the data center.
  • Seal cable cutouts, place perforated panels at computer intakes and implement aisle containment will help to get more aggressive with the temperature increase in the center.
    
  • Convert airflow supply (the fans on your CRAC units) from a static supply (and almost always under-rated at our altitude) to a dynamic supply such as an EC (Electronically Commutated) fan retro-fit to save ~30% and  improve reliability.
  • Implement a refrigerant economizer to a DX (Direct Expansion) system for a reduction of about 3,000 hours of compressor run time.

Colorado Data Center operators are leaving money on the table because of their ignorance of rebates. Can you elaborate on this program? Is there a deadline?

Xcel Energy has recently increased the rebates for participation in their data center program.  Xcel will fund up to 75% of a data center audit (capped at $25,000) and then rebate up to $600 per kW saved.

The program includes rebates for energy savings measures including: virtualization (of your server, network and storage environments), implementing airflow improvements (sealing cutouts and replacing or moving perforated panels), raising set point temperatures (these are almost always too low), improvements in the humidification cycle or methodology (steam vs. ultrasonic etc.), reducing electrical losses (replacing older UPS equipment, etc), and performing a lighting retrofit (or placing motion sensors in the space).  Nearly all data centers have done one or many of the above listed items and very few have taken advantage of the rebates available to them.

It seems that the big gap for data center improvements is the lack of visibility into facilities costs. How do your studies cross that chasm for customers? Are CIO’s surprised at the outcomes?

One of the things that I have learned over the years is to calculate the cost of doing nothing.  This can be very eye opening.  The holistic approach that we have developed helps the CIO to increase their visibility to the hidden costs (and vulnerabilities) that reside in their centers, and then bring out the benefits (and programs) that will bring value to the effort at hand.

Often, data center efficiency concerns begin with IT staff trying to get the attention of executive decision-makers. So, the cost of doing nothing includes all of these individual efforts instead of experts that can reach beyond specific technology concerns. Do you have advice for these professionals?

My advice is to identify the vulnerabilities first, how many single points of failure exist in the center, how much stranded capacity exists that has become unavailable, and then engage experts who can rapidly build a compelling business case for the C-Suite in a language they understand. Often, IT professionals need help to translate their concerns into something upon which the organization will willingly act.

Why go it alone when a catalyst like our firm can expertly contrast the cost of doing nothing with the savings of an effective plan?

What should every CIO know before they decide to move or consolidate their data center?

The facts – all of them…summarized to a fine point!  My concern has evolved over the years as I have seen too many centers wasting too much time, money, and effort solving the right problems, but without maximizing their return.  The problem to solve is a business decision.  After the organization decides where their investments will bear the most return, they need to abandon themselves to the strength of the team they have assembled.  Whether “in-sourced” or “out-sourced,” my job is to maximize that return for the owner of the data center.

Are you facing power, cooling, and efficiency problems in your data center? Contact Jon deRidder at Enabled Energy, a professional services firm focused on critical environments through its process of Discover – Optimize – Sustain.

About E-Oasis: E-Oasis moves data centers and provides the unbiased, independent expertise to make your data center relocation a success. Use the contact information to your right to get started.

Someday soon, someone is going to challenge you to move your entire data center to the cloud. Enough case studies have emerged to entice CIO’s (Chief Information Officers) to explore the promised cloud economics. After all, who can resist the dangled carrot to reduce costs, simplify deployments, and improve business agility?

No one.

And marketing departments know that. The reason is simple: IT (Information Technology), despite all the claims to the contrary, remains a cost center.  Turning this cost center into a value generator often ignores the complexity and legacy that defines Enterprise Information Technology services.

It would be easy to discuss all of the reasons why you shouldn’t move to the cloud. But that’s not why you’re reading this. So let’s complete the thought experiment exercise and discuss what you need to do to move that data center into the cloud.

Prepare a Budget

Nothing happens without money and you’ll need some fuel to complete your data center migration to the cloud. Unfortunately, the cloud has no mature economic models that you can simply plug in your information and arrive at a number to pitch to your CFO (Chief Financial Officer). If you are moving to the cloud, you are by definition, entrusting your working applications to others. Your budget needs to reflect the labor required to engineer an application migration and in some cases a re-architecture prior to that migration.

Application migrations are not new events in the lifecycle of most Enterprises. Therefore, moving these applications to the cloud requires the same discipline, the same use-case testing and certification, and the same skilled project management that was required in the past. The Cloud has not yet suspended the laws of how complex engineering projects are completed successfully. Don’t overlook the cost of educating and re-tooling your internal staff and users for the changes the Cloud will bring. Budget for the security testing required to be confident that your data is secure in its new home.

Dedicate a Team

IT (Information Technology)  projects fail for lots of reasons. Sadly, most of those reasons are well known lessons learned but not lessons remembered. Expecting your IT staff to do their normal jobs while at the same time moving your operation to the cloud is a mistake you can avoid.  A dedicated team will focus on the challenges, overcome them, and provide you with the best chance of success.

Clearly Define the Scope and Milestones

Like every exciting project, a move to the cloud will ignite the demand for other projects to ride along. Some of these will be compelling, some will be silly. Discipline up front with clearly defined scope and expectations for completing the move will be rewarded. Trying to please everyone also has it’s own reward…likely a new job.

Identify the Alternatives

If you haven’t considered alternative solutions to moving to the cloud, you are missing the opportunity to understand what you are doing. The Cloud hype is loud and proud. Believe all of it at your own peril. Being grounded in the alternatives allows you to effectively manage a move to the cloud with a realistic expectation of what you are buying before you buy it. The cloud is not magic and by understanding the alternatives, you will learn what cloud components make sense for your organization.

Performance Matters, Disaster Recovery Matters

Understand the end user performance before you begin the migration. Do you know what performance you are experiencing now per application you want to move? What additional costs are required to maintain or improve that performance? Have you dedicated an experienced architect to review your disaster recovery in a cloud environment? Carefully re-read those cloud case studies and you’ll realize that heavy lifting is involved.

Your IT Staff suffers from Confirmation Bias

Confirmation bias is the practice of enhancing information that supports a preconception and rejecting information that opposes it. CIOs need to understand that their staff suffers from confirmation bias.  And when it comes to the Cloud,  that confirmation bias can keep a CIO out of the Cloud with misinformation collected through the lens of confirmation bias. How many of these do you recognize?

• We already are our own cloud.
• There is no security in the cloud.
• There are no use cases for us to use the cloud.
• It’s too expensive.
• It’s unproven.

Clearly, you’ll need to overcome your staff’s confirmation bias against the cloud if you want to explore its benefits.  Moving your data center to the cloud should begin with you solving this problem first.