How to Start Your Application Modernization Journey
0 likes355 views
BT is modernizing its application estate to simplify its architecture and adopt public cloud technology, enhancing responsiveness to customer demands through a flatter organizational structure and agile practices. Key issues include challenges in data quality and the need for clearer planning and communication to facilitate application decommissioning and refactoring. The transformation utilizes squads and tribes to foster collaboration and accelerate modernization efforts, emphasizing continuous improvement and agility in the development process.
How to Start Your Application Modernization Journey
- 1. How to start your application modernisation journey 1
- 2. 2 Data classification. Published version. Owner’s name. Last 10 years working in transformation roles within Orange, EE and now with BT. Worked as an architect and technical lead within numerous transformation projects as well as a joint venture and for EE during BT purchase Currently leading the modernisation of BT’s application estate aligned to our transformation objectives, driving simplification and rationalisation of applications Steve Hawkins Senior Manager Converged Technology BT’s Modernisation Journey • BT’s Transformation • Setting up the Tribe • Refactoring Squad Structure • Taking your first slice • Scaling the work • Lessons Learnt
- 3. 3 BT’s Transformation • BT is driving towards major Public Cloud adoption • The organisation is changing to a flatter more accountable structure • Adopting modern methods to be more responsive to customer demands • Pivoting to Tribes and squad structures to make faster better informed decisions with empowered teams • Accelerate transition to modern technology • Remove complexity and legacy to reduce time to market In order to remain relevant and competitive BT is modernizing the organization and architecture
- 4. 4 BT’s Transformation in practical terms Simplify the Architecture Modernise the application estate Establish a new Operating Model Introduce Agile working practices Remove Legacy BT
- 5. 5 Data classification. Published version. Owner’s name. Numerous teams across BT wanted to fully understand the relationship between network components, IT applications and product variants in order to remove legacy and cost. Didn’t know how to start modernisation Were hesitant about Public Cloud The quality and completeness of data about our estate needed improving. Problem Statement…… Many people are looking at the same problem space. Previous experience tells us that full closure of all connected elements is difficult to achieve because unexpected connections turn up due to the quality of the data about how our estate is connected together. It can take months of analysis to clearly understand the potential impacts to the IT estate, and to find sources of information. The ability to realise cost savings is needed well-defined and realisable plans. Migrations needed traction and needed to have work aligned across Tribes to realise removal outcomes. Modernisation of the estate needed increased focus as an enabler for the simplification of the estate
- 6. Line of sight to Cost and Plans Work with key finance teams to create a view of IT Cost. Create models to support decision-making using cost, value and efficiencies. Hold the overall plan of which applications of the estate are being decommissioned / moved to cloud and the timing of each. Knowledge Repository Build, maintain and automate a data repository to aggregate application knowledge. Establish ownership and automation for data sources. Report on accuracy of the estate based on key focus areas. Establish a Self Service capability of the data. Produce dashboards, reports and detailed technical insights into the IT estate Establish engagement with external tribes to enable working relationships to drive opportunities for refactoring and decommissioning and build the backlog. Early Deep Dive applications for refactoring or decommissioning Maintain the front door for the tribe. Establish regular communications to stakeholders. Engage with Value Streams to build plans for execution Opportunities Specific squads will be set up to deliver outcomes in line with IT Transformation targets. Own and deliver against a single, prioritised backlog for decommissioning demand in line with KPIs with a single backlog. Establish tiger teams to accelerate decommissioning Decommissioning Outcomes Disaggregate legacy applications into microservices to support Value Stream and Common Component solutions and to enable decommissioning outcomes. Own the refactoring backlog, standards & design patterns. Deliver against a single prioritised backlog for all SIMPLIFY refactoring Build a community, capabilities and patterns for wider refactoring. Establish and lead Public Cloud refactoring aligning to SIMPLIFY objectives. Refactoring Outcomes Tribe Structure What is a tribe in the context of the work ? A model used within Spotify, a Tribe is a collection of squads that work to achieve the same outcomes
- 7. 7 Flow of work • Utilise the EA roadmaps to categorise the estate • Analyse the estate to establish the right group of application • Establish a selection criteria to isolate good candidates • Consider licences, 3rd parties, people costs and where possible business benefits • Build detailed understanding of each application • Execute refactoring – Determine your delivery model When dealing with a large application estate which needed a method to break down the estate into chunks As candidates are reviewed in each of the squads there is a refinement which drops out poor candidates Build Criteria and Decisioning to accelerate the assessment throughput Refactoring candidate App3 App1 App2 Opportunities Line of sight to cost and plans Knowledge Repository Refactoring
- 8. 8 Refactoring Squad Product Owner SA/Engineer SA/Engineer DevOps Engineer DevOps Engineer Product Manager Solution Architect Solution Architect Persistent squad members VMWare team Engineering teams join the squad Engineering teams return to BAU enabled Core squad established with team members that persist across applications Application specific engineers from BAU team VMWare ‘buddy-up’ to provide consultancy and education on the approach BAU Teams return having carried out refactoring Team are educated and experienced in Agile and refactoring Team can refactor and operate using Agile framework Transient squad members
- 9. 9 Refactoring Technology Approach Function 1 Current Platform (Weblogic/Java) Function 2 Function 3 Microservice 1 Microservice 2 Microservice 3 Target Platform Kubernetes Determine the best slice of the application to refactor by analysing how the end to end flow works Determine the outcome value resulting from refactoring Build out microservices that deliver a decoupled and resilient service Take a slice that is functionally discrete and can be easily refactored
- 10. 10 Scaling and CoE for Refactoring Product Owner SA/Engineer SA/Engineer DevOps Engineer DevOps Engineer Product Owner SA/Engineer SA/Engineer Squad 1 Squad 2 Engineering teams join the squad Engineering teams return to BAU enabled Establish multiple squads on a per application basis Establish a CoE squad to work across the organisation and establish best practices, patterns and methods Transient squad members Return to BAU enabled to operate in an Agile framework and utilise CI/CD pipeline deliveries
- 11. 11 Lessons Learned • Don’t be afraid to try different approaches to refactoring to determine what works • Changing the ‘Ways of Working’ is key and needs it’s own focus • Establish a CoE for refactoring to build best practices and methods • Functional alignment of the squads for refactoring brings stronger outcome benefits • Ensure stakeholders continue to be aligned to the objectives of the refactoring • Know the barriers that slow you down and be prepared to challenge them • Use the big imperatives to accelerate the journey