Seamless Cloud Adoption
By Arun Mehta, Head of Data & Analytics Engineering, Royal Bank of Scotland (RBS)
Cloud is a buzz word, and most of data-intensive enterprises in some shape or form are contemplating moving their data and compute workloads to Cloud. Cloud First is the headline strategy topic in almost every CIO forum. The key set of questions that come into mind are:
• What to move, when to move and how much to move?
• Where to start and how to embark on the Cloud adoption journey with confidence and with a better chance of succeeding?
• How to have positive traction with the stakeholders for sustainable relationship and sponsorship while delivering business benefits?
Below is the suggested and practical guideline for Seamless Cloud Adoption if your organization is thinking or has already embarked on Digital Transformation Journey where Cloud could be a key enabler.
1) Start with a WHY?
Clearly articulate the business outcome expected out of the move. Below are some questions to consider while creating a business case for cloud adoption:
• Will the move expedite/ facilitate the enhanced Revenue Stream, e.g., Moving the analytical loads (customer behavior analysis/ web channel clickstream analysis for upsell/ personalized offering) to cloud for faster time to market, handle uncertain compute elasticity or facilitate leveraging native cloud APIs for machine learning activities in a nimble manner?
• Will the move lead to Better Customer Experience via Digital Transformation Enablement– Implementing Contextual Nudges via digital channels / Enabling next best action via mobile push notification or value-added services?
• Will the move result in Operational Resilience and Cost Optimization (TCO)?
• Will the move result in Risk Reduction (Especially applicable to regulated industries like BFSI and Pharma)?
• Will the move result in Better Employee Engagement/ Experience and Workforce Productivity?
2) Do a LITMUS Test Against Cloud First Decision Tree
The movement to Cloud is possible and may be the best solution strategy, but it’s not a golden bullet. Putting a toll gate/ decision tree to vouch if Cloud is the answer to address the pain area via Cloud solution approaches (PaaS, IaaS or SaaS) gives you fact-based information to decide and puts you in control of the decision.
3) Select the Business Outcome and define the MLP (Minimum Loveable Product)
• Create a backlog of use cases and prioritize the one with the maximum tangible outcome and minimum risk
• Moving the mission-critical application or sensitive mandatory reporting data/ workloads may not be the first candidates in the organization to be prioritized to move to Cloud
The movement to cloud is possible and may be the best solution strategy, but it’s not a golden bullet
• Generally moving Analytical loads for Data Science activities that demand elasticity in the compute may be the initial use cases to be considered for Cloud movement as they come with tangible business benefits and relatively defined data movement
• The guideline here is don’t try to boil the whole ocean. Define the MLP with a hybrid of Cloud Foundation Layers and Use Casescope relating to the movement of associated relevant data, logic and workflow and expected functionality
• Plan to move only the meaningful portion to the cloud. Start with non-sensitive data upload to the cloud to have less initial resistance/ risk
• Lay out the foundation by delivering at least one fully functional cloud food chain covering data ingestion, data pipeline, security pattern, hosting pattern (Object Store of Big Data on Cloud) and associated cloud-native tooling for business outcome
4) Design for Business Outcome in DevOps/ Agile way
• Line up resources with can-do attitude and form a delivery POD with resources mix of the business domain, technology, operations, and Cloud SMEs.
• Use collaboration tools to plan the sprints, user journeys and expected outcomes with clear ownership and timelines
• Don’t try to deliver all in one go, use the iterative approach of delivery, review, and improvement in a product mentality form
• Transformations could be difficult, and in some cases, nerve-wracking and moving to could well be considered as a transformation from the legacy way of thinking. Key here is realizing the fact that expected outcome may not be evident in the first go.
5) Educate & Align Stakeholders
• Typical notion that comes with any cloud movement is that data and control of artifacts will move away in the hand of the third party. That’s where a robust communication must be made to stakeholders about the cloud benefits and security measures being undertaken. In some cases, very basics education of Cloud Tenancy and the associated potential value it can bring to an organization goes a long way.
• Alignment and syndication with Control/ Security group in a proactive manner is also a good strategy to avoid late surprises
• Making security/ control professional the allies and enablers in this journey by explaining each facet of the use case, potential benefit, sponsorship, and security control being designed to mitigate the risk reap in rich dividends
• Always ensure that such meetings are diligently noted with observation, next steps, the risk with their magnitude (High/ Medium/ Low). Avoid discussions and comments made on emotions and encourage people to give fact-based objections/ observations
• The general rule of thumb controls/ security groups critically review cloud proposals with the right intention to ensure companies asset (data/artifacts) from leaking/ hacking. To do so, they must share the magnitude of risk involved (material/ non-material). The go / no go decision generally rests jointly with the business owner and the CIO organization.
6) Assess Business Benefit & Socialize Success
• Do an objective assessment of business outcome following before and after state on laid out KPIs
• Socialize success with relevant sponsors, beneficiaries, and stakeholders at large to improve on the confidence quotient and future sponsorship of similar initiatives
Finally, to sum it up developing and delivering a seamless cloud-enabled business case requires taking your objectives into account, along with the current state of existing applications (application/ architecture), and their constraints. Engaged leadership, frequent communication, and clarity of purpose, along with aggressive but realistic goals and timelines with resources who believe in “Art of possible” make it easier for anentire organization to rally behind the decision to migrate.