IT Briefcase Interview: Preparing Effectively for a Cloud MigrationAugust 4, 2017 No Comments
As organizations seek ways to optimize their IT operations, almost half reported they will focus on expanding cloud initiatives in the next 12-24 months, according to a 2016 Datalink-commissioned IDG survey. While migrating to the cloud is top of mind when it comes to IT optimization, many organizations tend to rush through the necessary preparations in order to get up and running as quickly as possible.
In this interview, Peter Kraatz, Senior Manager of Cloud Service Management and IT Resiliency at Datalink, an Insight company, shares his expertise on the due diligence necessary for organizations to successfully move to the cloud and the challenges IT professionals should prepare for prior to a cloud migration.
- Q. Organizations have been leveraging the public cloud for a number of years. Have business cloud strategies remained the same or have strategies changed? If so, what prompted the strategy changes?
A. My observation is that the strategy hasn’t changed so much as the attitude. As for the strategy, there continues to be little in the way of a defined strategy from IT leaders, which is troubling. By that I mean there are few defined goals, little engagement with the business to map their drivers to the solutions being considered, or a healthy understanding of the complexity involved with what I’ve called the “change in attitude.” I would characterize the change as moving from “take it slow” or “throw up every objection possible” to “I want to BE Amazon, right now.”
I can’t quite put my finger on what has prompted the change but it does feel like capitulation. Some IT leaders who are approaching us have had their jobs threatened over progress toward public cloud. This is not a rare occurrence, either. I’ve had a half dozen conversations like this in the last month alone. It feels like the business and key IT stakeholders have had enough with science projects and stalling, since their service levels and fulfillment times have not been improving. Clients have not embraced self-service, automation and orchestration. This is the one thing I can identify as a key use case or promise that the business would like to see happen. Public cloud at least seems like it will deliver those benefits (to the stakeholders, at any rate) and thus the pressure on IT to move forward grows.
- Q. What are the challenges that organizations face when preparing to leverage a hybrid cloud model?
A. The first challenge is consumption. Just like the move from physical to virtual, consumption gets a LOT easier. Unlike the move from physical to virtual, the public cloud side of the hybrid model is not forgiving about waste and inattentiveness to consumption. Clients are at least aware of this and are attempting to mitigate these impacts before they occur.
Organizations also struggle with the architecture and integration of their hybrid cloud model. Particularly at the network layer, key decisions have to be made about how to assemble the kit. Firewalls tend to feature prominently in our conversations (virtualize what I have, use the provider’s service, etc.). The default seems to be “go with what you know,” even if that solution isn’t the most efficient, simply because public cloud skills and platform familiarity are a commodity in high demand and short supply.
A third major challenge lies around an organization’s application portfolio and workload placement. The “6 R” model (rehost, replatform, refactor, replace, retain, retire) that analysts and others have cited needs to be taken seriously and integrated into the client’s application portfolio roadmap, but isn’t. We’re still having a lot of conversations around what workloads belong on what platform (this is good) but not enough conversations about remediation and action plans before we move a workload or application. So even if the workload is placed within an optimal container, hosting model or service platform, we have to anticipate rework later to get the greatest benefit out of that final resting place.
- Q. What level of effort should IT teams expect? And, who is typically involved in the process?
A. There’s a lot of “it depends” in any answer I can give here but I’ll jokingly say “nobody expects the Spanish Inquisition!” What they expect when we come to them is that there is a simple way to draw up a new design, just like implementing a new storage or compute platform, and then to stand it up and begin work. They don’t expect us to reel them back in and insist on a solid architecture, build an application portfolio plan (retire, replace, replatform, etc.), and map out a migration roadmap. They also don’t anticipate the work needed to engage the business and assure that the business work flow (not the data flow) will not be impacted. We ask a lot of questions and it does feel like an inquisition at times. They should expect the actual move to public or hybrid cloud to be straightforward. The planning and design around that, however, is the bulk of the actual work.
- Q. Just curious to know about cloud readiness. How can an organization assess if they are operationally prepared to migrate to the public cloud? Can legacy applications be deployed to the public cloud?
Last question first: most of the time, yes. The question, though, isn’t about the technology beyond the application’s dependencies and other workloads that must be moved with it due to factors like latency.
- Q. How can an organization assess its own operational readiness?
A. First: do you have the skills in house with the platform (provider) or are you using contract resources to implement the solution? In the latter case, it’s unlikely the organization will have controls, management tools and processes in place for managing capacity, change and issue resolution. Teeing off those elements, organizations that aren’t regularly monitoring and reporting on consumption, assigning chargeback ownership, or in possession of toolsets to do these things for them are likely to find their journey more difficult, with costs mounting as a result.
Are service levels defined and managed rationally? Organizations with mature roles, processes, the right tools, and good service level expectations tend to be more successful. These KPIs serve to move the majority of the learning curve to the platform or provider.
- Q. How about walking us through Datalink’s framework for assessing business and workload requirements?
A. We treat the assessment and alignment of workloads as though we were planning to conduct a major data center migration and remediation exercise, where the target state needs to be clean and current when we are done. This forces us to review everything about the estate.
Technology is pretty straightforward. After we conduct our baseline inventory and application blueprinting, what does the roadmap for each application look like? Is it destined for retirement? Should we replace it? Replatform it? Or is the organization ready to refactor the application to optimize it on someone else’s platform
Resources for us mean people. We want to make certain those who have to manage the environment after we’re gone are capable of the task or at least have a path to acquiring skills and tools they lack. Otherwise, we have to reconsider our assessment of the workloads and make different future state recommendations.
Processes are the governance, controls and day to day “how things get done” of IT. While not having to worry about the hardware is nice (and eliminates a lot of issues), not being effective change managers, having little experience with chargeback or capacity management, and other potential challenges can introduce risks that we consider for our recommendations. It’s one reason why we end up helping with a lot of phased implementations, putting less important but painful to service workloads in someone’s cloud so we can see better what processes break before we move to the crown jewels.
Additional more specific considerations include:
* Overall cloud strategy (need to start here, need to answer the “why?”)
* Total cost of ownership
* Dependencies of workloads to current environment/IT infrastructure
* Impact cloud will have on people and process
* Will the SLA/SLO of application workloads be the same or better?
* How will the clients leverage both Mode 1 (keep the lights on) and Mode 2 (innovation) technologies?
- Q. Are there any data migration considerations or best practices that need to be considered pre-deployment?
A. There are a variety of considerations, including :
* What are the dependencies of the data to the application and the IT infrastructure?
* Has the data been profiled, in terms of its SLAs back to the business, its rate of change, who needs the data, etc.?
* Has a data migration strategy (how you are moving it) been identified?
* Are the maintenance windows adequate enough for the amount of data being migrated?
* What does the network look like?
Preparation is key is all aspects of a cloud migration. While organizations often try to “go with what they know” in terms of cloud architecture and workload alignment, it is critical to take time upfront to create an action plan and train IT staff. Doing so will allow your business to reap the greatest benefits from a cloud migration.
Peter Kraatz is national portfolio director, cloud services, at Datalink, an Insight company. He has been successfully delivering IT projects for more than 20 years. IT projects spanned business continuity and disaster recovery to information security and public cloud technology deployments. His mission is to help clients realize high-performance offerings that meet or exceed consumers’ requirements by putting customer service at the core.
.CLOUD DATA, DATA and ANALYTICS , DATA SECURITY, Fresh Ink, SOCIAL BUSINESS