Seven Steps on the Path to the Cloud

Contributed Article.  Author: Jon Reeve, Senior Director of Product Management, SolarWinds
 

Typically, private cloud initiatives get boiled down to how IT services are provided that meet target goals for:
 

Specifically, agility (time to react to business requests and needs) is a key driver and yardstick to measure the success of a private cloud implementation. When we speak with customers who have a private cloud initiative underway, they are typically highly virtualized (80% or more). This naturally leads to the question, “What is the difference between an environment that is 90% virtualized and a private cloud?” The answer almost always comes down to some combination of agility (and self-service) coupled with chargeback or showback (which is changing how IT services are consumed).

It’s not surprising that those undertaking private cloud initiatives are also highly virtualized, since many of the characteristics that enabled the widespread adoption of virtualization are the same ones that will be needed for successful private cloud adoption (note that most of these have nothing to do with technology!):
 

1. Listening to the Business & Understanding its Requirements
 

As in most relationships, communication between the business and IT is a crucial element in fostering a successful partnership. Some businesses don’t feel that their requirements are met by the IT department, so they reach out to vendors who they think can solve their problems. This leads to the emergence of “shadow IT” within an organization – a collection of software that IT did not vet or purchase, but will likely become responsible for managing. By the same token, IT managers sometimes end up building applications that nobody uses. Maintaining open communication with your organization’s business units is key to ensuring that everyone’s needs are being met in the most efficient way.
 

2. Building Trust
 

From open communication stems building trust with transparency. The reality is that you don’t get to 90% virtualized in the first place (and from there to a private cloud) if you haven’t done a great job at building trust with the business and key application owners, whose critical applications are going to be running on a virtualized infrastructure. When persuading an application owner to virtualize that application, you must build the case for it and demonstrate that it will still perform as needed, when needed. The skills you develop as you become more virtualized are the same skills you’ll need as you move to the private cloud.

3. Picking the Right IT Services to Begin With

In theory (and ideally in practice as well), internal IT departments should have an unfair advantage over their external service provider competition since they are in regular communication with the application users, and should have an understanding of IT and business issues as they arise and how best to address them. Not everything is suitable for the private cloud model — mainstream, high volume (standardized) services are best, whereas legacy one-off applications with highly customized requirements may not be well suited or provide the desired return on investment.

It should be apparent by now that the business has options, and that IT will be compared (in terms of cost, quality and agility) to the competition (external service providers). Competitive analysis in an organization typically comes from the top down. It may stem from the CFO asking something like “how much does it cost us to run application Y?” As C-level executives are being targeted by external service providers, it puts IT in a reactive position. Having a competitive analysis in your back pocket will help demonstrate that you have the needs of the business in mind and are working to address them on a daily basis. It also aids in transparency (key to the aforementioned trust building) by presenting an unvarnished look at what IT is able to provide internally versus what might be better handled by an external provider.

The old adage of “you can’t manage what you don’t measure” is apt here. In the course of your competitive analysis you’ve taken account of what IT can provide internally and at what cost, but it’s also important to factor in the increasing demands within your organization and look at the manpower cost of providing these services. While most IT budgets are done on an annual basis, services like the Amazon cloud (AWS) allow you to pay for only what you use, and to pay by the hour. As the demand for these IT services increases more quickly than the annual IT budget, the question becomes “how does an IT department reconcile the difference between being a yearly budget expenditure to more of a usage-based ongoing service provider?” Showback (showing what resources are being used) and chargeback come into play when IT staff is regularly being asked to provide additional support and services. In an effort to ensure transparency and change behavior, IT should consider tracking showback, even if at first they’re monitoring it for their own knowledge. As many consider chargeback to be a necessary requirement for a true private cloud model, showback will act as a stepping stone to measuring internal use. Being able to pinpoint which areas of the business are using the most resources or services will guide IT in their conversations with C-level managers and other business unit leads.

External service providers are focused on the user experience – how to provide things that users need and make them run properly. They recognize that the end-user doesn’t care about technology for technology’s sake or how the service is delivered, as long as it’s doing what they need it to do. This shift from focusing on the nuts and bolts of internal IT, to an IT service provider mentality and a focus on the customer is perhaps the hardest of all to instill. IT centers of expertise will always exist to some extent, but by keeping user needs in mind, IT can more closely align itself with the business units.

Virtualization is a great tool to enable the abstraction and automation necessary to reach cost, quality and agility goals, and standardizing (reducing the number of moving parts) is key to removing complexity in our environments. With a standardized and finite set of infrastructure parts, there will be a more constrained number of failure patterns and potential things to fix. As some industrial cloud providers have very few standardized infrastructure parts, this speaks to why it is best to move repeatable, high volume applications and services that can be standardized to the cloud.