15. The software-defined datacenter….
- All repeated tasks fully automated.
- All configuration in source control.
Goal: Build a new REA datacenter from scratch and deploy
all apps at the push of a button.
16. Star Deploy Star Deploy
• Build Pipelines
• Push-Button Deployment REA Apps
• Build Pipelines
• Push-Button Deployment REA Platform Image REA Platform Image
VMWare
• Automated….
• Creation
• Monitoring Amazon Web Services
• Destruction
Physical Tin
Today - markets, brands, financials, growth profileHistorystartup, listing > bankruptearly growth > leadershipinternationalisation > defocused/stagnation2008: leadership change2009: rebuilding a healthy core. Key: TW (Agile, XD), LM (platform), HW (reliable ops); core group of key staff (~25), lots of sweat and commitment from all staff, lots of contractors.Mid 2010: people (Delivery).Current focus:broadening the value proposition > market maker, not just market participant.optimising operational performance > global operating model.Financial performance
STARTING POINT (2009)Monolithic organisation, thinking internally and moving cautiously Few leaders, low benchstrength, poor employment brandCentralised, plan-do decisionsWaterfall development, ad hoc operationsCluttered design, low delivery outputUnreliable infrastructure, catastrophic riskPoor business systems, lots of manual processes
Focused objectives:Delivery a brand new site with a modern experience, new products, new platformQuickly make tactical site improvements.Stabliise our intrastructureDecided (1) couldn’t be done in-house. Lots of debate internally and with TW. Deal with LM.(2) Became the cataylst for Agile. TW engaged as coach, Agile Adoption Group initiated. Would be disbanded less than a year later after massive success in moving quickly to agile development.(3) Turned into a focused DC rebuild. Too early/risky for wholesale move to cloud – discussed and deferred.
Not long after launch, EVP rose significantly. We were seen as a company that could move quickly, was serious about agile, and looking for talented people to create experiences and solve interesting problems that would be appreciated by millions.Leadership, talent build, restructure teams along business domains (vs technologies) to attack multiple opportunities in parallel, and build domain knowledge so teams could participate more fully in the design of the product, not just execution.
InceptionBusiness mapping, groupsketching.Most potent of all – unlocks the potential for design-level thinking, helping the team uncover needs before moving into potential solutions.
Continuous delivery
Register. Opportunity to guide customer-focused thinking, without telling. What unmet customer need are we solving?Hacking. Get your product or business or design personnel to participate in teams.Showcase and vote. Watch your team start to vote up hack entries that are most likely to have the biggest customer impacts, rather than just the coolest tech stuff.