We talk to many IT organizations these days who want to know how to move their current data center operations to what many are calling a private cloud. They want to know how to transform their current virtual infrastructures into an Infrastructure as a Service (IaaS) model, so that they can add efficiency, consistency, and a deeper understanding of the actual costs of running a data center and “serving” applications to the business.
IaaS allows IT resources (typically comprised of compute, network and storage systems) to be packaged and automatically offered, on-demand, to end users as a selection of services. Public IaaS providers like VMware vCloud Air and Amazon Web Services (AWS) also offer these types of services in a consumption-based, chargeback or “pay-as-you-go” format.
Going on a road trip? Better be prepared...
If you’re like many other companies, your road from virtualization to IaaS might be paved with good intentions. But, like any extended road trip, you need to make sure you first have the essentials in place. Map? Check. Accommodation details? Check. Gas in the car? Check. These initial checks ensure you don’t ruin your trip or get stuck in the middle of nowhere.
Anticipated moves to IaaS are similar. We look for a host of foundational elements in your organization before saying you’re ready to hop in the car and start your journey. As we look, we often find the same foundational issues that need correcting before you go too far down your own road to IaaS. It’s kind of like building a house without the right foundation. You can do it, but don’t blame us when the floors start to buckle or the roof leaks.
Time for a tune-up before you head out
If you’re serious about successfully arriving at your IaaS destination, here are some common mistakes to address first in your current virtual infrastructure:
- 1. Deployments that do not adhere to best practices. There’s a reason vendors work closely together, investing significant time and money in validating deployment of their respective virtual software, server hardware, network components and storage systems. They’ve worked out the right options for end-user success with their products. In customer installations, however, we find that most environments operate with inconsistent, misconfigured, or missing processes for everything from deploying an operating system to tagging a VLAN. For IaaS to be successful, consistency and validated steps are key.
- 2. Lack of documented processes. Many organizations are accustomed to calling on their in-house IT experts to get certain things done. How do you deploy a new virtual machine or set up a new VLAN? They say, “We just call John and he takes care of it.” A documented process is critical for when John’s not around or for creating an automated process. As a prerequisite for automated processes, documentation adds consistency, efficiency, and insight into the future cost of operation of your IaaS resources—whether private or public.
- 3. Resource oversubscription to unhealthy levels. Who in IT doesn’t love the idea of over-subscribing (or thin-provisioning) your resources to cost-effectively maximize their use? The concept, in both storage systems and virtual servers, has developed with some sophisticated features and safeguards to ensure you don’t overdo it and that you still maintain high availability (HA) of your resources. What we sometimes see, however, is intentional overrides of over-subscription safeguards in order to add a few more virtual machines. While this practice may avoid short-term budget costs, it can unduly endanger system availability. It’s also often in direct violation of an organization’s availability-oriented service level agreements (SLAs).
Successful public IaaS providers don’t make these mistakes. Neither should you. If you’re making them, make it a priority to first get your virtual infrastructure in order, and take the time now to schedule your own tune-up. Contact us today to learn more.
