10 Key Steps for Moving from Legacy Infrastructure to the Cloud
- 1. Top 10 Best Practices for Transforming Legacy IT to the Cloud September 2016 James Bond
- 2. 2 Top 10 Best Practices for Transforming Legacy IT to the Cloud 1. Virtualization 2. Automation and Orchestration 3. Cloud Design 4. Cloud Management Platform 5. Deployment 6. Brokering 7. Operations 8. Staffing 9. Security 10. Cloud Native Applications AGENDA
- 3. Virtualization * Focus on server Virtualization. Network, storage, data center virtualization beyond scope of this presentation and the time allotted for this webcast
- 4. 4 Standardize, Consolidate, Governance Virtualize and Automate Onramp to Cloud Private Cloud Baseline Enterprise Cloud Hybrid Cloud Multiple Cloud Providers Broker Arbitration, Aggregation Across Numerous Providers Essential for all IT and data center operations Critical path for data center modernization and IT operational efficiency Initial cloud infrastructure and internal IaaS service Additional providers, PaaS and SaaS Service Brokering Transitioning from Legacy IT to the Cloud
- 5. 5 Virtualization by itself does not equal cloud …but it is a good start to modernizing legacy traditional data centers Note: Hypervisors are used to virtualize physical servers. Do not confuse a software hypervisor for a full cloud management platform
- 7. 7 Automation and Orchestration • Transition all manual software installation processes • Consolidate or incorporate all manual decisions or business processes • Adopt a “Relentless Pursuit of Automation” mentality Adopt overall orchestration platform Automation vs. Orchestration • Automation is task based such as a scripted software installation; static sequential • Orchestration is dynamic, bi-directional workflow with numerous task automation; query/if-then logic, parallel paths, reverse/back-out transactions; integration with third-party tools (e.g. Puppet, Chef, etc.)
- 8. Cloud Design
- 9. 9 “All Clouds Will be Hybrid” Information Applications Infrastructure Traditional IT Private cloud Virtual Private cloud Public cloud Traditional IT Evolving current state The future hybrid IT Multiple cloud deployment models (managed separately) Dedicated, physical, homogenous Integrating multiple cloud models in a single hybrid cloud system and customer experience Traditional Private cloud Virtual Private cloud Public cloud
- 10. 10 Application Portability Assume workloads / applications will change cloud provider/platform in future Applications/Workloa ds Existing cloud services HP Public Cloud HP CloudAgile Partners Traditional IT Private cloud Public cloud
- 12. 12 Management of the Cloud Cloud Management Platform Unified portal, automation, orchestration, billing, reporting, operations, and API
- 13. 13 Cloud Management Functional Layers Evaluate and buy a cloud management platform (with ongoing support and upgrades). Do not develop your own management platform—it will have limited functionality, will be very costly, and huge distraction from your core business
- 14. Deployment
- 15. 15 Standardize, Consolidate, Governance Virtualize and Automate Onramp to Cloud Private Cloud Baseline Enterprise Cloud Hybrid Cloud Multiple Cloud Providers Broker Arbitration, Aggregation Across Numerous Providers Essential for all IT and data center operations Critical path for data center modernization and IT operational efficiency Initial cloud infrastructure and internal IaaS service Additional providers, PaaS and SaaS Service Brokering Transitioning from Legacy IT to the Cloud DEPLOY Deploy initial basic IaaS private cloud within existing data center(s) MATURE EXPAND
- 16. 16 The Evolution from Traditional Data Centers to Cloud
- 17. Brokering
- 18. 18 Emergence of the Internal Service Provider Public Cloud SaaS Private Cloud IaaS PaaS Public Cloud Private Cloud SaaS Today’s IT Reality • LOB becoming more technology aware • Alternatives exist in the market for direct consumption • Technology is viewed in terms of service value from the user • Users & developers want simple • Security, compliance, availability Service Catalo g Dev Platfor m • Compliance • Governance • Competence • Data sovereignty • Budget constraints • IP protection • Culture • Agility • Secure data transfer Considerations Internal Service Provider – providing the right destination for the right application
- 19. 19 Existing cloud services Traditional IT Private cloud Public cloud Cloud Management System Multi-Provider, Multi-Cloud Brokering Cloud Management Platform performs aggregation and arbitration to internal and/or external cloud providers
- 20. 20 Cloud Brokering • Consolidation of SLA, billing, utilization reporting across all providers • Security and event monitoring and alerting • Multi-tenant real-time dashboards and reporting • Requires integration/APIs between cloud broker and all providers Aggregation Arbitration • Static rules/logic to determine which cloud provider is best fit for each application/workload that is ordered • Criteria set by customer and broker such as cost, SLA, geographic location, security compliance • Future: Dynamic logic engine to use real-time analytics (price, performance, SLA, capacity, utilization) to determine the best fit cloud provider • Future: Combine dynamic arbitration with application/workload portability to automatically move workloads between cloud providers to for optimal performance, cost, availability, etc. (think: electricity grid) Private cloud Traditional IT Virtual Private Cloud Public cloud
- 21. Operations
- 22. 22 Operations • Design cloud infrastructure for 100% availability; 99.9- something is no longer acceptable • Design for zero system scheduled or routine (even for upgrades) maintenance outages • Use live failover, scale-out nodes, and resilient cloud native apps/services for 100% up-time by design • Utilize real-time replication of data rather than batch backups whenever possible; this also facilitate granular (low/quick) recovery point objectives Availability Plan for Resiliency – not for Recovery
- 23. Staffing
- 24. 24 Staffing • One of top problems for organizations deploying cloud are existing IT staff bad habits, failure to adapt/change, legacy approach • Existing IT personnel were hired and typically serving traditional IT roles (and paid accordingly); therefore often not suited to design/build your cloud • Many traditional IT staff are organized into teams based on technology such as Linux and Windows OS/servers, SAN storage, networking, monitoring, security Use Caution Leveraging Existing IT Personnel To Build Your Clou • Use external expertise for cloud design, transition planning, initial implementation • Augment your existing staff with external cloud expertise during initial cloud operational period while legacy staff is trained. • Reorganize traditional IT staff into service-oriented teams. Cloud orchestration/automation and development/ops are critical cross-functional skills. • Don’t forget your developers. Traditional developers may not really Recommendations
- 25. Security
- 26. 26 Security • Orchestration = consistent monitored processes/provisioning of cloud services = easier detection of the abnormal • Utilize orchestration to enable immediate continue monitoring, system configuration scans, updates • Cloud can be more secure than traditional IT; however, the impact of failure or security breach can have broader impact Automation / Orchestration Improves Security • Replace manual security processes to assess and approve every new server or application (resulting in deployment delays) • Pre-certify all operating system and server/virtual machine templates so that the orchestration systems can deploy new cloud service in an automated manner on demand • Pre-certify all networks segments, VLANs, load balancer configurations, IP address ranges, etc. so that software defines networking and orchestration tools can automatically deploy/configure cloud services on Pre-Certification
- 28. 28 New Approach for Application Delivery Application Development and Delivery
- 29. 29 Faster application deployment enables accelerated innovation Developer Centric • Agility is paramount • 30x increase in application releases • “Even when delivering at cadences of less than a week, 20% of organizations want to go even faster” 1 • Developers need flexibility • “Expects to deploy 50%-70% of code to cloud environments by 2015”1 • Want open flexible architectures for application portability and prevent lock-in 1 Forrester Thought Leader Paper commissioned by HP, “Better outcomes, faster results. Continuous delivery and the race for better business performance”, December 2013 2010 4 releases/year/app 2015 36 releases/year/app 2020 120 releases/year/app
- 30. 30 Cloud Native Application Attributes Open Standards and Open Source Reduces vendor lock-in and speeds your access to innovation Elastic, Distributed, Resilient, and Composable Accelerates IT productivity, performance and availability Continuous Development and Delivery Rapid delivery of updates and new features through automation Micro-Services, Application Components, APIs Enables scale-out, resilience, active node failover and recovery
- 32. Summary
- 33. 33 Top 10 Best Practices for Transforming Legacy IT to the CloudVirtualization: Start with virtualization, but by itself this is not cloud. Don’t confuse hypervisors with cloud management systems. Automation and Orchestration: Automate all manual processes – “relentless pursuit of automation”. Understand automation vs. orchestration. Cloud Design: Most clouds will be hybrid. All workloads should be agnostic of cloud and assume cloud provider will change in future. Cloud Management Platform: Implement hybrid cloud management system early. Developing your own is too costly and disruptive. Deployment: Start with a basic private cloud Infrastructure / IaaS services within internal data center; then mature to PaaS and expand to multiple providers/internal service brokering. Brokering: Internal service brokering is a new maturing capability in the industry—aggregation and arbitration are static and limited capabilities today. Service catalog portals that provision services to multiple cloud providers is easy--this does not equal a true cloud brokering platform with dynamic arbitration and full aggregation. Operations: Target 100% availability with no service outage maintenance windows, real-time replication and failover rather than backup/recovery approach. Staffing: Many existing internal IT personnel are not suited to design/build your new cloud. Hire external consultants and restructure existing staff removing silos based on technology and align to service-orientation. Security: Automation = consistent quality and easier detection of abnormal. Clouds can be more secure than traditional IT; however, impact of failure can have broader impact than traditional IT. Continuous and immediate monitoring. Pre- certification of “everything”. Cloud Native Applications: Porting traditional apps to the cloud only moves or repeats the problems/limitations. Cloud native apps are designed for cloud: resilient, composable, and micro-services. Auto scaling, self healing, and designed for failure through retry, throttling, queuing and admin notification rather than user errors.
- 34. Thank you Updated industry news, analysis, or links to purchase the book: go to facebook.com/enterprisecloud
Editor's Notes
- #30: Applications are transforming at a rapid pace. Application deployment has greatly accelerated over the past few years. As you can see from this diagram, which comes from a Forrester report, in 2020 enterprises are looking to deploy 120 releases/year/app, up from 4 releases/year/app in 2010 This is a 30x increase. Developers are expecting to deploy the majority of applications to the cloud. And they also want to be able to move applications from one environment to another so they can avoid vendor lock-in. These changes have a huge impact on IT. IT must make significant changes to support this accelerated release cycle and to meet these developer requirements