Don’t Let Your Cloud Project Grow Up To Be A Failure…
I am the luckiest guy in the world. I get to talk to customers every single day about how to move seamlessly to the cloud. I get to talk to our Microsoft Product Groups about features that are needed to help our customers have an ever better experience during and after migration. And I get to talk to our Worldwide Field, including our Consultants, about how to help our customers navigate around possible blockers. Now, I have an opportunity to write this blog post in hopes of passing on some experience to all of our Modern Service Management (MSM) blog readers.
Recently, I was asked for input on a response to a customer who asked “how can Microsoft help me ensure that my cloud project does not fail?”. That led me to quantify the following musings:
1) Failure-level technical failure is not all that common, but I will list it first due to the importance of security and all things that are cyber
Cloud migrations themselves seldom fail insofar as the technology is concerned. Moving workloads from point A to point B is an achievable task. It is as simple as moving compute and capabilities from one place to another. Smart technical people work through challenges all the time—it is what they/we do. The challenges with cloud migrations fall in the before and the after. The “before” challenges fall in the technical and philosophical planning—will this be more secure or less secure? will we need to change our network ingress and egress design to accommodate the added traffic to and from the cloud since the compute resources are not in our local datacenters? Those “before” questions are usually technical. Note, however, that I do not intend to diminish the importance of cyber security in any way whatsoever. My point, simply, is that our industry is, by definition, focused on the technology aspects, so those are achievable.
The “after” questions are where the most important, yet uncommon questions emerge. Those questions need to be pulled ahead to earlier in the lifecycle—how will I monitor my workloads in the cloud? How will I integrate the alerts and incident workflows for my cloud workloads with my single-pane-of-glass in my Operations Center? How will I audit and report on use? How must I evolve my support and accountability structure given the role-based-access-control capabilities of the cloud? How will I enable continuous integration/continuous delivery of my cloud-born-applications (e.g. automation, machine-learning, and artificial-intelligence next-generation systems) in the cloud while satisfying the compliance requirements of my change and configuration management processes? The list goes on. For the purpose of this blog post, we will focus on the “after” questions; as a reminder, the “after” questions should be predicted, assumed, and pulled ahead in the lifecycle—make the “after” questions “before” questions.
2) End-Users are the second opportunity
There are at least four key classes of cloud migrations:
a) Moving a workload to the cloud as-is (e.g. lift-and-shift to IaaS)
b) Re-writing a workload to really leverage cloud capabilities (e.g. rewriting an app to run on PaaS instead of IaaS)
c) Cloud-born application (e.g. a brand new application that would not likely exist without the power of the cloud)
d) Moving a workload to SaaS
In all four classes of cloud migration, the question that must be asked is “will the end-user experience change”? If the answer is “no”, then you can likely move on to bullet #3. But if the answer is “yes”, then the end-users must be considered. In many projects, the end users should be considered, but they are not considered. When they are considered, they are often considered only with the lens of “how can I avoid help desk calls from these users when I go to production?” But, what we really need to be considering is the business objective of the project in the first place. Why are we doing this cloud project? Are we moving to Office 365 so that our end-users can be measurably more productive? Are we moving our sales application to the cloud so that we can handle holiday spikes via cloud elasticity instead of capital expenditure for on-premise, underused hardware? Are we creating this fancy new IoT, machine-learning based service to transform our industry? Inevitably, there are business goals for the project. Those business goals will only be realized if the end-users effectively use the service(s). So, end-users are the key to successful projects. If they do not use the service effectively, then the business results are not realized, and we fail. It is interesting to realize that the end-user aggregate outcomes are actually the SUM of the results of every individual end-user’s behavior. We need to think about those end-users, and we need to be planful and proactive with how we land any changes on those individuals.
And since the prompt for this blog post is “how can Microsoft help me…”, I need to include that response here. Microsoft recommends that customers be proactive and planful with landing the human-element changes on the end users in order to realize the anticipated business value of their projects. To this end, Microsoft has a worldwide consulting practice dedicated to helping customers land that human-element change on the users. The “Adoption and [Human/User] Change Management” practice enables customers with training and consulting based on Microsoft’s experiential learning (along with industry-standards in [human] change management practices) with millions of cloud customers in every industry, vertical and sector.
3) IT Pros are the third opportunity
Once you have a plan for the end-users (assuming their experience and/or behavior needs to change), we need to turn our attention to the IT Pros. IT Pros are all concerned about devops and cloudops. Many worry that they will no longer have a job. Many worry that the cloud provider, or the development team in the case of devops, will not have the experience to deliver the capabilities that they are signing up for. Many worry about outages because they inevitably have a wealth of experience supporting the workloads in the environment already, and they know there are nuances that exist because of the technical, cultural, and compliance background of the workloads. IT Management needs to be proactive and planful about painting the future for the IT organization, the IT culture, and for the IT Pros themselves early in the lifecycle of the change. Many cloud migrations, and the desired business outcomes thereof, are based on the promises of agility and cost savings. If the IT Pros push against the cloud (cold) or if they drag their feet (warm), then the value realization will not happen as fast (or at all). Management needs to land the message successfully so that people push towards the cloud (hot) so that the business promises are met on-time or ahead of schedule. Wouldn’t it be a shame to be ready to go into production with a fantastic, new, cloud-born application that will change your business’ market position only to be blocked by your global operations team due to lack of operational readiness? I see it too often. We all need to get in front of the IT org, IT culture, and IT Pro issues.
“How can Microsoft help me…” for this bullet is included with bullet #4 below.
4) IT Integration is the fourth opportunity
Most organizations have hundreds or thousands of named applications in their portfolios. They do not move those to the cloud all at once. They move a subset of applications over time. Realistically, if an organization has 1200 applications and moves 1 a month to the cloud, it would take them 100 years to move them all. Thus, it is clear that there will be co-existence for some period of time. It is one thing to operate an application or workload in the cloud. It is yet another thing to operate a cloud workload in concert with the thousands of other workloads in the portfolio without duplicating operational teams, processes, or tools. The integration must be planned where “just enough” changes are made to existing processes, tools, roles, permissions, etc. to enable seamless operations of the cloud workloads. Monitoring, Major Incident Management, Deployment/Change/Release Management, Evergreen Management, Problem Management, Service Desk Readiness, Escalation Protocol, Dependency Mapping, and Workload Cost Accounting are some of the more critical aspects that should be planned.
“How can Microsoft help me…”: Microsoft takes the IT Pro and IT operational integration very seriously. Microsoft has formed a Worldwide IT Service Management organization whose charter is to help customers shorten their learning curve with all of the areas covered above with the move to cloud and devops. The ITSM practice is staffed by trained, experienced consultants who are armed with training and intellectual property directly from our product groups. The intellectual property is not only aimed at “how our customers should do things”, but the IP is based on how our product groups actually run the commercial cloud services internally. There is value in the lessons that Microsoft has learned in the > 13 years that we have been doing the cloud since customer 1 for Office 365, and there is value in the lessons that Microsoft has learned in supporting literally millions of tenants on top of our cloud services. The Worldwide, Customer-Facing ITSM practice's entire mission is about taking those lessons directly to our customers using well-trained and experienced cloud operations consultants. The solutions run the gamut from consulting to training to online training offered on the Premier Portal.
Of course, there are more ways to fail, and I acknowledge that I have painted with broad brushes here. My intent is not to give an exhaustive list, but rather to inspire everyone to think beyond the normal technical implications. It is very important that enterprises consider the end-user behavior change in order to achieve the desired business outcome. And, it is very important to consider the IT elements—IT organization, IT culture, IT Pros, and IT process and tool integration. Often, in my experience, every technical challenge can be overcome. And thankfully, the end-user and IT elements can be overcome as well, but it is always best to deal with those issues proactively earlier in the lifecycle rather than to wait to address them reactively later.
| Carroll Moon ( @carrollm_itsm ) has been at Microsoft for over 17 years. He spent 4 years in Microsoft Services focused on service management, and then he moved to what is now the Office 365 product group. In his time in the PG, he led the service management strategy, engineering/development, and global operations for Office 365. Carroll spent 8 years leading Service Management for Office 365 Dedicated, and he spent 2 years bringing the service management “stack” that his team built in Office 365 Dedicated to Office 365 multi-tenant in order to better enable our enterprise customers. Carroll had the unique opportunity to lead teams in moving critical Office 365 capabilities from “on-premise” to Azure. Carroll returned to consulting in late June 2014 with the mission of helping customers shorten the service management learning curve for consuming and providing “cloud” services. Carroll holds 10 patents or patents-pending for Service Management, and he is helping to define “service management for cloud” for the industry: he recently authored a series of articles on Cloud Monitoring along with a Cloud Ops Webinar and Podcast . In April 2017, Carroll participated on an industry-focused panel for DevOps and Digital Transformation at the Software Risk Summit , and he is currently authoring a series for Office 365 Service Management . Carroll lives in Lynchburg, VA with his wife and two children. |