When and why should you go to cloud?

A while ago, @killerloop asked my opinion on an article in TechCrunch that drew my attention. The article is titled “The Enterprise: I’m not sexy and I know it.” Frankly the author Rodney Rogers clearly articulates what many CIOs tell me—large enterprises are complex and not the least bit “sexy.”

Rodney, I’d like to respond if you allow me. I agree with you, but I also believe that positioning “cloud computing” properly in the enterprise allows the company to adapt quickly to the evolving economic environment. But let’s start from the beginning.

The need for agility and responsiveness

The marketplace is becoming more and more global. If you don’t believe it, read “The World is Flat” from Thomas Friedman and you’ll be convinced. Competition can show up from anywhere. It doesn’t take much to start-up a company these days. You just need a good idea, and there are many people with great ones.

To make things more complex, the market is extremely volatile at the moment. Because of this, companies have to respond quickly to opportunities and threats. Agility and responsiveness is of the essence.

Let’s look at an example: for years, telecommunication companies have been talking about Voice Over IP (VoIP). But not a lot has happened. Then from nowhere, a couple of people, specialized in pair-to-pair file exchanges (you remember KaZaa),  created Skype—the largest telecommunications company by number of phone calls at the moment. It was a disruptive innovation.

Most telcos did not see this coming, and some are still in denial, pretending Skype is no competition. The successful ones were capable of giving their customers the capability to call at a price point in line with Skype. This implied re-engineering systems and quickly rolling out new technologies. Now, Microsoft is looking at adding automatic translation to Skype, the next disruption. You know if you are an interpreter, your job might be gone soon. Sure, translation will not be at the same level day one, but it might be good enough.

So, the lesson from these examples are how to react. As Rodgers points out, enterprises have a vast array of technologies as well as a complex and heterogeneous landscape. This represents hundreds of millions of dollars of technology assets deployed on laddered refresh cycles through a variety of delivery models, supporting highly differing divisional requirements across a number of development platforms. With these assets, how do you react to change? Do they work with you or against you?

Become responsive?

The first question we need to ask ourselves is what responsiveness means for IT and the IT environment. Are we asking for more of an existing application or are we asking for something new? If the answer is related to an existing application, and this application is a traditional one running on the assets described above, increasing capacity on short notice may be something difficult to achieve. For example, if the number of orders that need to be taken increases drastically, we may have to create a front-end application that buffers the orders so they can be processed by the traditional environment when capacity is available. Creativity may have to be deployed to address the needs in a timely manner.

If a new application or service needs to be created, a cloud environment may prove an excellent approach. Let me come back, once more to a point made by Rodgers. He speaks about a number of development platforms. In a traditional environment, each platform uses its own hardware, or the developers spend time re-configuring environments. A private cloud environment, in which the development platforms can be provisioned, is a nice way to allow the developers to quickly respond to the needs of the business.

Does that imply the enterprise needs to buy brand-new hardware and acquire one of those “gold-plated appliances?” Not really. The advantage of the integrated environments such as HP’s CloudSystem is that they can be taken into production very quickly. But the software powering them is perfectly capable of running on top of the enterprise legacy environment, converting some of the technology assets in a private cloud, limiting the upfront investment. Time to production may be somewhat slower because customization is required, but it’s a perfectly viable alternative. Another way of doing things is to start with an appliance and then enlarge the development cloud by integrating existing assets. As you can see, there are a lot of alternative possibilities here.

One size does not fit all

First, let me tell you I fully agree with Rodgers, ERP is absolutely not dead. It’s out there for quite a while. Because of this, any cloud vision needs to incorporate the traditional environment. If your enterprise spent millions in implementing SAP or any other ERP environment, you’re not going to dump that any time soon. You may not even be willing to invest in porting this application to the cloud. It runs well, supports your day to day business effectively, but is no longer a competitive differentiation because most of your competitors run the same environment.

What you would probably like to do is to complement your ERP with new functionalities that allow you to differentiate yourself from your competition. You may want an effective CRM solution. You may also want a top notch collaboration environment to streamline your supply chain, interact better with your suppliers and understand where improvements are possible. Why not use cloud computing for such functionalities? As the amount of new requirements increase, the functionality you use from your existing ERP environment will reduce. Over time you may want to think about sun-setting the application, but that will only happen in several years.

In the meantime, you may be looking at Salesforce.com or Microsoft Dynamics CRM. The first question to ask yourself is whether CRM will differentiate you from your competitors. Is this something you need to develop yourself to have a unique differentiator, or are you just looking for the functionality out of the box? In the first case, you may want to develop the functionality and have it available in your private cloud. In the second case; you will probably lean to the sourcing of SaaS services. But do you want your users to have to know which decision you made? Do they need to remember where they can find the service? Probably not. This is where you can really make a difference.

What if you had an integrated portal/service catalog, in which your users found all the services at their disposal? What if you allow them to choose, within a list of accessible services, the one they need at any moment in time? Your private cloud platform would then either deliver the service (if this is a core service you are providing yourself) or intermediate/aggregate the service from an external provider. You can even go one step further and encapsulate the traditional environment using web services technologies so it can be accessed by your users in very much the same way as the new cloud functions you have implemented or sourced. This gives your business users one entry point to IT, shielding the complexity of the delivery models.


Migrating to the cloud does not require throwing all the existing hardware away. In many situations the same hardware can be used for traditional and cloud based (private cloud) environments. The traditional world and the private cloud world will co-exist for the forcible future. Providing the users with one user experience to access IT shields them from the actual implementations and allows you to migrate services from traditional to private, from private to public cloud over time. OpenSource will facilitate such migration and facilitate the migration of functionality from one environment to another. In that context OpenStack is quickly becoming a de-facto standard for infrastructure as a service.

Should your first task in migrating to the cloud, be to migrate your ERP or any other of your traditional environments? I do not believe so. It is costly, requires very specific capabilities as Rodgers points out, and does not really add a lot of new capabilities to address the business requirements of agility and responsiveness.

What is critical though is that IT understands the needs of the business and reviews what functionalities need to be provided when. Then the decision can be taken to source services from a service provider or develop them depending on the competitive nature of the functionality and data provided. This can only be achieved by building strong governance between business and IT.

So, Rodney, you’re asking six key functions to migrate your SAP to the cloud. They make a heck of a lot of sense and can mostly be addressed by existing technologies today. However, I’m not sure that is where I would start. I’d use cloud for business innovation and collaboration first, helping the enterprise grow in the current business environment. Yes your enterprise is large and makes a good deal of money. You may feel it’s not sexy. But let’s work together to make it more responsive by employing the right technology in the right place, regardless its name. What do you think?

Christian Verstraete is the Chief Technologist Cloud at HP and has over 30 years in the industry, working with customers all over the world, linking business and technology.

Follow Christian's blog