I had the pleasure of presenting at the VTUG Summer Slam this year. This event is always great, if for no other reason than the “pull out all the stops” Lobster Bake at Gritty’s following conference sessions. This year, I even had the great fortune of sitting down and eating lobster with Hans Bernhardt from VMware, a perennial presenter and long term supporter of this event and user group.
Working with our cloud engineering team, my presentation focused on our experience here at Markley in helping customers to extend their in-house VMware infrastructure into the cloud. Our primary focus was to highlight strategies and best practices to make these projects not only go smoothly, but also meaningfully add value. Here is a quick summary of some of the ideas presented:
Do it for the right reasons
Implementing a new infrastructure is non-trivial, so it is important that goals are well considered early in the project, and continually evaluated. Cloud has become the mantra, yet many project leaders end up frustrated by end results that don’t measure up to a set of ill-defined, ambiguous objectives. Express goals in terms of their ability to tangibly increase efficiency and value. If you can’t identify how you are going to save time, lower cost, reduce risk, or enhance value for your organization, it’s important to pause and reset.
Plan the work, work the plan
At Markley, we actively work with our customers to plan projects in five phases (Discovery, Assessment, Design, Implementation, and Management), and to consciously set milestones and checkpoints to direct their work. Each phase is wholly dependent on the success and completeness of prior phases. Validate your assumptions each step of the way, for example, to confirm your cost savings and expenditures in the assessment phase, then verify the validity of your design with sample workloads and real-world data (For example, we provide customers free access to cloud resources to implement a proof of concept testbed to validate their design long before moving it into production).
Discovery – Do it first
Discovery is the most important step of the process. This is where you derive a deep dive understanding of your current environment, specifically uncovering all the logical assets under management and hidden, unspoken, and forgotten dependencies these assets have on each other and the underlying infrastructure. You will do the Discovery phase, either painfully – throughout the project, at risk to the success and timeliness of its completion, or more easily, - upfront and to the benefit of all downstream activities.
Now is the time to pay your technical debts
Virtual environments tend to grow like weeds, so now is the right time to weed the garden. Unplanned and amorphous growth begets bad habits. What constraints were imposed in the past that no longer exist? How would you architect your environment differently knowing its current performance and limitations? Is your networking architecture as it should be? What new features would enable the most value? When assessing your go-forward plan, don’t lowball your resource requirements or compromise the integrity of your assessment.
The Cloud is not a data protection strategy
We often witness the misnomer that ‘my infrastructure is naturally more secure or better protected because it is in the cloud’. There are certainly features and capabilities that can help you architect security and data protection effectively in the cloud, but you cannot divorce yourself or your organization from these responsibilities. Your design and implementation must account for proper firewalled access, data backup and recovery, and considerations for high availability and disaster recovery.
If interested, I am happy to share my slides by contacting me at email@example.com . For more information on how Markley helps customers employ our mission-critical facilities, robust connectivity, and in-house cloud infrastructure to engineer VMware hybrid clouds, download our data sheet here.