John Cupit, Director, Cloud & Data Center Services, Global Solution Expert Team, Huawei Technologies USA Inc.
Not all roads lead to Nirvana....
I believe that most Enterprise CXOs today recognize the power of cloud technologies to transform the Enterprise and to create certain strategic imperatives moving forward. From my perspective, cloud services (if deployed properly) can shift IT from a “Support Platform” paradigm to a “Value Generating Platform” environment. The Value Generating Platform leverages IT resources to provide elastic services to unlock constrained opportunities and create new business models that surpass the well-known limitations of the Non- Cloud Model. It is this transformation, focusing on the creation of service catalogs, the rise of APIs in support of new services and the creation of elastic infrastructure that begins to generate value and competitive differentiation for the Enterprise.
“The key to successful cloud adoption and migration is more about the chosen process guiding the journey as opposed to the cloud technologies that are selected for deployment”
But with the vendor FUD in the market about cloud computing models, how does one truly know the path to take? Too much focus is often taken by the project team on the technology wars and not on reaching a complete understanding of the existing infrastructure, application dependencies and key business requirements.
Furthermore, one of the most misunderstood consequences of moving to a cloud model is the fact that the Enterprise will be abandoning many of its organizational and operational practices that have defined its mission in the context of the traditional IT environment. As we transform the infrastructure, how do we transform the organization in order to maximize the deployment of cloud models?
Through hundreds of migrations, my personal experience as a Cloud and Data Center Transformation expert indicates that certain answers to significant questions must be reached before the Enterprise commences the journey toward cloud services, whether provided internally or over the top:
1. How do I link the strategic initiatives of the Enterprise to the IT objectives of transformation?
2. What is the best way for my organization to consume cloud services?
3. Which cloud services should I source externally versus build internally?
4. How will I migrate my applications and data to the selected cloud services?
5. How can I integrate the cloud environment into my current operational platform?
6. How will I manage my service levels?
The lessons learned (painfully, I will admit) is that the Enterprise must utilize a methodology which drives the collection of certain information required to determine the most viable alternative for cloud transformation. I have found the following methodology to be very useful in this process. Discovery Requirements Definition To identify all of the technical and The key to successful cloud adoption and migration is more about the chosen process guiding the journey as opposed to the cloud technologies that are selected for deployment business requirements for the cloud solution
• Discovery Assessment of Current State How are services provided today?
• Infrastructure Discovery What technologies are currently utilized within the environment? What is the current BC/DR solution and what impact will a cloud model have on that solution? How much existing infrastructure does the Enterprise wish to reuse?
• Application Discovery Application footprints, application dependencies and machine-to-machine application flows (Web Tier-App Tier-DB Tier). This is also the correct time to conduct an Application Rationalization process to identify applications that could be retired or upgraded.
• Analyze Solution Alternatives At this point, enough information exists to have meaningful discussions around cloud technologies and models (Let the technology religion wars begin!). Once the discovery and characterization of the application mix is completed the applications can then be individually analyzed to ascertain if they are candidates for retirement, igration to cloud services and/or Data Center optimization.
• For those applications which are candidates for cloud services, it is necessary to further analyze them by Cloud model. This step requires a detailed analysis of multiple factors for the applications including the following:
• Platform requirements
• Storage requirements
• Bandwidth requirements
• Migration considerations
• Sunsetting of applications
• Costs to migrate legacy applications to a virtualized environment
• Solution Recommendation & Roadmap Development Justification and rationalization for the cloud model selected for transformation. This phase also creates the following migration deliverables:
• Application Migration Method of Procedures (MOP) The definition of the move groups and migration waves based on the application dependencies identified in the discovery process.
• Data Migration MOP A MOP focusing on the migration of data to the recommended cloud model.
• Migration Timeline The creation of a Migration Timeline supporting the transformation to the new environment The other huge error often made in cloud migration projects is the failure to define a Common Cloud Management Platform before the final cloud models are selected. The CCMP exposes a set of business and operational management focused services as well as provisioning, orchestration, federation and management services. This is where service catalogs can be developed to facilitate the use of APIs to rapidly turn up new services in the future. Besides business and operational management focused services, the CCMP also includes User Interfaces serving the three main roles defined within the CCMP
• A Service Consumer Portal to be used by Cloud Service Consumers for self-service delivery, provisioning & management (the actual cloud service instances are used via a cloud service specific UI)
• A Service Provider Portal serving Cloud Service Provider internal users & administrators for daily operations ; and
• A Service Development Portal used by Cloud Service Creators.
CCMP functionality is accessible via APIs exposed by the CCMP- internal components. As the name already implies, the CCMP is structured as a platform. Based on the platform nature, the CCMP exposes a set of services which can (and sometimes must) be used within the context of a specific cloud service. The management services exposed by the CCMP to cloud service creators are not to be confused with the cloud services developed by cloud service creators.
At this point, enough information is in hand to facilitate the evaluation of integration approaches as well as a determination of how service levels can be appropriately managed in connection with the recommended solution alternative.
In conclusion, cloud transformation requires careful planning and a roadmap which defines how cloud will create a Value Generating Platform. Transformation not only represents a considerable investment from a technology perspective, but it also drives the development of new processes and roles. As such, cloud migration is an evolution it is not a revolution.
Infrastructure and Application discovery is the key to identifying applications which can be transitioned to cloud services, particularly since certain applications can function better in certain cloud models. The integration of orchestra tion, management, provisioning and federation into existing Enterprise IT infrastructure becomes the key process in terms of successfully operationalizing cloud services. The process also helps to define how the organization will change in order to optimize the transformation.
I also feel that it is important to mention the impact that SDN and NFV may have in the future. As with many new technologies, there is not enough deployment history in hand to provide firm guidance. With the maturation of SDN and NFV, and the desire to utilize container technologies as well as cloud brokerage mechanisms, the analysis will only become more difficult in the future. All of these added layers of abstraction and microcode in effect trade less code complexity for more operational complexity and I don’t think that the market has caught up yet to truly understand the level of operational complexity that they will create.
Unfortunately, in the arena of cloud transformation, there is no one all-powerful solution or approach. There is no silver bullet. Based on my personal experience, the key to successful cloud adoption and migration is more about the chosen process guiding the journey as opposed to the cloud technologies that are selected for deployment.