In Part 1 and Part 2 of Cloud Wars: Friendly Fire, we covered some of the challenges cloud advocates and early adopters are experiencing within their own organization as well as some of the reasons that they may be receiving pushback. In this third and final segment of the series, we'll cover some additional challenges and offer some potential strategies to work around or work with your IT department that might help you move ahead in your adoption of cloud computing.

"Can you wait six months? We're building a private cloud."
More and more frequently we are hearing from our user community that their IT organizations are building or evolving their existing datacenters into what are commonly known as private clouds. These private clouds are generally based on VMware technology so they are pretty efficient and augmented by the addition of tools like vCloud Director which adds self-service and tools like VMware Chargeback to help make the cost more transparent. While we do believe this is a great step in the right direction, it often takes time and there are numerous examples of failed and delayed private cloud implementations. Simply stated, change like this isn't easy and still ends up conflicting with other critical IT priorities. In all likelihood IT will get their private cloud up and running internally so your best approach is to make sure that anything you place into the public cloud is fully compatible with the technology they are using to build the private cloud. A very common technology for both public and private cloud environments is VMware vSphere. It is essentially a cloud operating system. So any applications you run in the public cloud, as long as you work with a VMware vCloud Provider, will work if and when you need to move them back into your private cloud once IT has it up and running. (a link to a directory of VMware vCloud providers is at the bottom of this post)

"Security, security, security."

If all of the previous attempts to slow you down in your cloud journey are not effective, someone in the organization will ultimately throw down the security gauntlet and take a stance that the cloud is less secure than the existing IT datacenter. I've seen two effective approaches to resolving this. The first of course would be to only run workloads in the public cloud that fit within the level of security capability that IT has today. So if your IT department is SAS-70 Type II, PCI compliant or HIPAA compliant then you should choose a vCloud service provider that meets or exceeds those criteria. The other approach, that takes a bit longer but is a better strategic play is to work with IT to break down exactly what the level of security needs to be in detail. This allows you to determine a checklist of security capabilities you can use to evaluate cloud computing providers as well as pick which additional managed services you might need or be interested in--assuming that they provide them. The list would address things like the regulatory compliance standards mentioned above but also more specific items like firewall, intrusion prevention, data encryption and logging levels.

A more interesting but less common approach would be to look at the cloud as a hybrid solution. We have seen some business units control the workload and application itself, but use their IT department as their managed service provider for security services. That way if your organization has a firewall vendor preference for Checkpoint, they can simply extend their existing management to include the Checkpoint firewall you have in your cloud environment.

If you're interested in digging further into cloud security, we recently released a new whitepaper with VMware titled "Security in the Hybrid Cloud: Putting Rumors to Rest." You can download it and other cloud whitepapers here.

So in conclusion, hopefully our real life experiences working with business units and IT organizations to figure out how to fit our enterprise cloud into their existing world will help you with your own cloud journey.   

As promised, here is a link to a directory of vCloud Providers from VMware. 

So, I'm curious... what do you think are the greatest barriers to cloud adoption?

It's important to know that while the challenges I listed in Cloud Wars: Friendly Fire - Part 1 are very real, the core IT department is behaving that way for some very good reasons. They are struggling with not only the transition to cloud, mobile devices, big data, SaaS and social media but also with trying to do so while under a lot of duress due to economy-driven budget and headcount cuts. Being the sole caretakers of so many diverse mission critical technologies, applications and data in the company is not an easy task. In fact, some groups within core IT are running into challenges in their own department in an attempt to begin leveraging public cloud resources.

Let's look at a couple reactions you might hear from the IT department and some potential ways to address them...

"IT has capacity internally already, why can't you use that?"

If this is the reaction you recieve when you seek approval or get caught using cloud, IT will try to address your needs from an internal infrastructure standpoint and want to know how much memory, compute and storage you need. One of the main reasons you wanted to leverage cloud in the first place was so you didn't have to know what you needed, because you could only pay for what you used. This is a wildly foreign concept in most IT organizations that are accustomed to doing a lot of work to determine your resource needs. This process can often take several weeks if not months. They need to gauge not only your initial requirements when the project starts, but also how much it will grow over time because they have to buying enough hardware and software for the three years over which accounting requires them to depreciate the assets. They may also have to determine if they need to hire additional people. What kills it here is that your $1000 a month idea suddenly becomes a $250,000 or a Million dollar major project. If you tell them that your starting budget is only $1000, they will wonder why you are  wasting their time and move on. Note that IT will come back at some point if and when your project is successful and the dollars become more "real" from their perspective. Transparency is a good approach even though it may seem challenging and if you have the time, it is a better long term play. Bypassing IT or being allowed to try cloud under the guise of a pilot project shouldn't be considered lasting lasting approval.  

"Cloud is more expensive!"

More efficient cost is one of the biggest advantages of cloud computing. When comparing the cost of cloud computing infrastructure to traditional IT costs, the IT organization typically won't include the cost of people and operations management because those costs are hard to scope down to a specific department or application. Recurring costs like these are frequently spread across all lines of business in some form of allocation. This means marketing is paying for Oracle support whether they use Oracle or not. This is why it makes sense to get ahead of the total cost of ownership (cloud TCO) discussion. To do that, take the capital cost they are showing to you and add to that the entire cost of IT that was allocated against your business unit in the budgeting process. I'm pretty sure that will make this a big number. Now compare that to the cloud costs. If you have multiple applications in your business unit or if you use the general IT services like e-mail, you may have to factor those out somehow, but it will likely still be a big number. Now you have the basics to dig into a real TCO discussion and comparison. The good news is that this is going to be much more painful for them than it will be for you. You can simply bring your bill to the meeting, or if you haven’t made the leap yet, bring some vendor quotes.   Top cloud computing providers will have already broken down your cost and have line items like this:

RAM    $320

CPU  $150

Storage Tier 1 $400

Storage Tier 2 $202

Internet $50

Managed Firewall Service $125

Etc.

We provide our VMware Cloud customers a free tool called Bluelock Portfolio. Portfolio breaks cost down by application, component, or even by business unit.

IT generally also thinks in terms of fixed not dynamic capacity. As shown below, the red area represents fixed cost that you pay for whether you are using it or not. In the cloud, you only pay for what you need and that red area can either result in savings or be used for additional projects.

 IT as a Service (ITaaS)

This may seem like an oxymoron and it is one of the biggest challenges for traditional IT departments as they evolve in the cloud era. IT traditionally determines their priorities, and therefore where your support ticket or your project request fits into their queue, based on overall organizational priorities. That would seem to make a great deal of sense. The challenge is that IT is generally highly under-resourced and under a great deal of pressure to be more efficient so deferring smaller more tactical needs and department requests unfortunatley become a recurring byproduct of this approach. With cloud computing, your cloud computing provider is highly incentivized not only to deliver a good service becuase you are voting with your dollars every day or month when you pay your bill. They are also very eager to ensure your success because typically if you have a good experience you will not only stay, but if your project does well it will grow or earn your cloud provider the chance to do additional work for you and grow their own revenue stream. In the future, IT may work this way as technologies like VMware Chargeback and Showback become more prevelant and the cost of IT becomes easier to understand. While this may take some time, you can use the cloud today and even share your experiences with IT to help them understand some examples on how they might evolve their service to better fit the business.

In the third and final chapter of Cloud Wars: Friendly Fire we'll cover the myths and challenges around security, private cloud and hybrid cloud.

Battles over formats and standards in the technology industry aren't new. Whether it was e-mail, word processing, graphical images or even some more current like the apps on your smart phone, each new innovation typically starts out somewhat proprietary and incompatible.

Today we live in a world where a lot of those battles have been fought and won while some are only starting to heat up. Formats tend to resolve themselves through standards so that things like e-mail and web pages "just work." That or at a minimum the technology we use evolves and hides it all from us via various forms of automatic conversion. In a world where so many technologies seem to get along - why shouldn't clouds? The answer isn't so simple and, as is often the case, history has a way of repeating itself.

Different Approaches
Just as email initially emerged inside of private datacenters, so has cloud infrastructure. It was initially based on virtualization technology and, depending on what kind of IT shop was involved, you most likely ended up on VMware if you were trying to make your core datacenter more efficient by virtualizing legacy servers; Xen if you had a significant Linux or Java developer presence where the need to rapidly provision test and develop machines was important; and maybe even Citrix if you were using Metaframe for serving up applications to remote users or thin clients. For some companies, cloud adoption started outside the corporate datacenter inside of the VMware vCloud or Amazon Web Services (AWS).

Different Platforms
These approaches initially evolved into distinctly different virtualization technologies. VMware's vSphere platform grew up in the heart of enterprise data centers so it focuses more on performance, manageability, stability and uptime. Amazon's platform is directed at rapid provisioning of instances that are great for developers and dynamic workloads. While they both promote the concept of virtual machines as the building blocks of their clouds, the formats are very different. Amazon breaks its workloads into images called Amazon Machine Images (AMIs), which contain an operating system image from a limited library, memory and CPU resources. VMware uses a format called a Virtual Machine Disk (VMDK) that contains not only the memory, but also a more flexible operating system image that can be based on any x86 and, as you can tell by the name, the disk storage itself. With Amazon's AMI model, any additional storage outside of the operating system image must be kept on a separate disk image on the network using either their S3 or EBS storage solutions. With VMware's VMDK, because the disk is included, you can move everything around. They call this a vApp.

Competitors
It also goes without saying that the war for cloud market share has started and is being waged in earnest. Given the amount of dollars at stake, it makes complete sense. The major players all want their share of the anticipated $241 billion that corporations are predicted to spend in the cloud over the next 9 years (according to Forrester Research). This kind of growth opportunity resembles a land grab of significant size and proportion that fosters only self-serving forms of compatibility.

History Will Repeat Itself
As we now know from the email platform wars, once the bulk of companies picked their initial email standard to implement, the green fields of opportunity started to dry up. As a result, email platform vendors found themselves in a situation where they could only gain market share by taking customers away from each other. This caused them to start offering various forms of migration services. The initial offerings were assistance with your email strategy, which was closely followed by professional services teams that could perform the migration for you. At some point, the professional services teams developed tools to make the migrations easier and those tools ended up being packaged and offered directly to end users.

We are starting to see this trend emerge in the cloud. Most Software-as-a-Service (SaaS) providers have either a toolkit or a professional service offering that helps customers migrate from one SaaS provider to another. You can see this with Salesforce and Netsuite. VMware and Amazon are no exception. Amazon offers a tool that allows you to convert your VMDK virtual machines into AMI images. VMware has an extensive professional services offering and numerous channel partners that can help an Amazon customer migrate to a vCloud or back to their internal VMware vSphere environment if a customer found themselves locked in to the Amazon cloud.

Interoperability and Open Standards
It's clear that more help is on the way, but there may be some solutions that can help in the short term. The Distributed Management Task Force, also known as the DMTF, is a standards body that represents 160 member companies and organizations and more than 4,000 active participants crossing 43 countries. They have endorsed and are actively promoting the Open Virtualization Format (OVF) standard for Virtual Machines(VMs) and Virtual Applications (vApps). OVF support is the foundation of the VMware vCloud and VMware vSphere 4.1. Workloads can be quickly converted to OVF and then moved or copied between clouds as the needs dictate. This gives the IT department and power users the ultimate flexibility as they begin to adopt a cloud-enabled approach to IT.

Which One Is Best?
Choosing the best cloud platform is a lot like deciding to upgrade from a tube television to a flat screen. It can be hard to choose when you see all of the different flat screen television technologies like DLP, LED, Plasma and LCD but once you've made the leap to High Definition, regardless of which technology you chose, it still puts you light years ahead of where you were with your old TV set.

How to Choose
The most important thing is to determine what the right use cases are for your organization to adopt cloud. For some companies that may be to make their existing IT resources more agile and efficient so placing some of the workloads they have already virtualized using VMware into a public vCloud make the most sense. For other companies that may need to perform thousands of calculations on a moment's notice, consuming that capacity from a commodity cloud like Amazon's EC2 might be the best fit. The real key is that this landscape is evolving and changing rapidly. Any workload placed into the cloud should conform to existing IT practices and should be able to be de-provisioned, redirected or moved as the cloud market matures and the understanding of which clouds best fit which workloads becomes clearer.

The Cloud Wars have already begun and as in all conflicts, the most unfortunate incidents are caused by friendly fire. From a cloud computing perspective, it occurs most often when people who want to try or have tried public or hybrid cloud are challenged by peers or counterparts in their own companies. 

I’ve been thinking a lot about the challenges my fellow technologists have experienced in trying to sell their public, private or hybrid cloud initiative inside an organization. I also realized that not only was this a difficult task for them during their initial project, but that they are compelled to continue to justify their cloud initiatives on an ongoing basis. I know that in most situations the team that needed to be convinced of moving forward with a cloud project was corporate IT. In some cases the "cloud champions" were trying to convince the CIO to try something new. In some cases a cloud committee or cloud strategy working group was already working on it and the project was a distraction to their process, or it ended up under a microscope as a learning opportunity instead of the business critical application that it was for the department. In other cases it was actually IT leadership trying to encourage their teams that the old world might be changing, just a little bit, but the going was pretty slow because no one was really sure what that meant.

In each case a lot of the core challenges were similar but there were also some differences.

Most of my fellow technologists were headed to or ended up in the cloud in the first place because they needed or place a high value on agility. I define agility as being able to:

• Act quickly
• Deal with the unknown
• Change your mind

And they generally felt like going outside of core IT made the most sense because:

• Cloud was less expensive because you only pay for what you use
• They perceived the service was better
• The response to issues would be faster
• Promises about service delivery and uptime were written with financial penalties for under-performance, typically in a Service Level Agreement (SLA ) document
• They or their team members could provision resources rapidly, and via self service
• There was transparency and true visibility on the cost of IT
• As a cloud customer, they were a priority and could vote with their dollars
• They felt wanted because multiple vendors were competing for their business
• It was a more cost effective and lower risk way to try newer technology or new idea

And though those things made a lot of sense, when they exposed their plans to the internal team or in some cases, when the internal team found out they were already using the cloud, the reaction was generally:

• We have capacity internally already, why can't you use that?
• Buy some servers and a SAN, we'll manage it for you.
• Cloud is more expensive.
• We do understand service, we can act fast, just submit a ticket.
• Our standard cloud offering is [fill in cloud vendor here], use them.
• Can you wait six months? We are implementing a private cloud.
• Security, security, security!
• They were in a busy with a big company-wide project and were unavailable for the meeting.

Those questions and challenges are legitimate and can be quite difficult hurdles to navigate especially if you are trying public cloud for the first time.

In my next post I’ll share some cloud adoption challenges in more detail as well as ways that I’ve seen people overcome them.