August 16, 2013 | Written by: Biswajit Mohapatra
Share this post:
In part one, we discussed the assessment model of the current IT infrastructure and IT environment such as the type of hardware been used, software stack, their current software level and various network zones like intranet, extranet and Internet. We also covered the importance of automated data collection using tools and scripts and get the output of these scripts applied in the following cloud adoption model.
Cloud adoption model
After assessment is completed, the next step is to perform feasibility analysis and develop a strategy for cloud modernization to an IBM cloud environment. A diagrammatic view of the cloud adoption strategy practiced by IBM is depicted below.
A reference model for cloud solution adoption utilizing various services is to be considered while creating strategy for cloud modernization. As part of our work, we considered infrastructure as a service (IaaS), platform as a service (PaaS) and software as a service (SaaS) for cloud modernization. We didn’t take business process as a service (BPaaS) into account incurrent context considering our cloud modernization requirement was not catering to business process optimization aspects.
Identifying candidates for the target cloud
The IaaS adoption model is aimed at moving application imagea as a whole from source to target environment. In our work, we applied the 80-20 rule where 20 percent of applications within the customer enterprise will serve 80 percent of the services and most of the critical applications are associated with strong non-functional requirements (NFR) like performance and growth. Most of the applications will move to target cloud environment AS-IS without much change. This also reduces the risk of uncertainty during cloud migration since change in hostname and IP address increases risk and cost of migration due to dependency on upstream and downstream connections, firewall rules, proxy setup and certificates.
There are several ways to adopt this model such as extending the existing network into the target cloud environment, such as a VLAN extension where the same subnet or gateway can be maintained between the source to target network. You can take a copy of the image by using tools like PlateSpin, VM-Converter, MS-backup and Flash-achieve. Most organizations move their workload in two phases to minimize the risk of migration to cloud by moving to cloud as-is without introducing many changes which is associated with immediate return on investment (ROI) from reduction of hardware footprint and optimizing the environment to achieve efficiency which will provide continuous improvement.
PaaS adoption is associated with a defined catalog of services running for a set of middleware in a pre-defined set of the OS environment. Applications may not move as-is due to incompatibility of the middleware or OS version between the source and target environment. This requires application transformation, which is associated with either upgrading the existing application middleware or OS environment or re-platforming the current application deployment environment. The workload which cannot move to IaaS as-is due to the incompatibility of an existing image in the target hardware needs the following PaaS adoption model:
- At the virtual machine (VM) level – create a VM in the target environment and migrate the component from the source to target unlike the whole image move. Tools like AppZero can be leveraged for component level migration.
- At the application instance/container level — enterprise Java application move from old container to new container such as WebSphere Migration Toolkit, application profile export-import.
- At Database Instance Level – Database workload move from source to target database instances. Tools like DB2Move, DB2load can be leveraged for database instance level migration.
This adoption model is more complex compared to IaaS adoption and it requires extensive application know-how and testing knowledge.
The SaaS adoption model is associated with transformation of source application into off-the-shelf services available in the cloud in a target environment. In the SaaS adoption scenario, the application is neither migrated as-is nor upgraded to the target environment. This model is associated with application rewrite to adopt existing cloud services in a target environment. Building a multi-tenant application or multi-tenant database or adopting cloud services in a target environment are examples of the SaaS adoption model. This is the most complex transformation type of cloud adoption as it calls for application re-engineering for the target environment. All application development and transformation project life-cycle methods need to be followed to drive desired results.
After finalization of cloud adoption, we need to refer to IT infrastructure and application assessment results and develop a roadmap for grouping application workload using the above adoption choice. Candidates get selected based on software compatibility, reference architecture, workload characteristics and application platform dependencies.
The following process can be adopted for candidate selection:
- Separate out and start the “quick win” group that does not require detail level assessment and get buy-in for implementation schedules
- Collect application-specific detailed application data, review app interfaces and app development/deployment in the current environment
- Perform application code inventory for custom applications – for instance, run code counter tool against code repositories
- Understand and estimate the effort and cost for network readiness, system readiness, storage readiness and software licensing readiness
- Determine if any in-flight projects along with resource availability constraints
To summarize this approach, it’s very important to understand current IT environments and target adoption models. The next step is to categorize the workload based on the adoption model, estimate the effort and develop a cost for each adoption model. We need to consider that for any cloud modernization initiative, a big-bang approach is the biggest fallacy. A careful workload evaluation and strategy going into each of the adoption models is the key to success. All cloud modernization solutions should be addressed from an enterprise perspective and define the ROI based on the model adopted and accordingly a roadmap needs to be developed over a period of time.
Biswajit Mohapatra is an IBM Certified Consultant and Global Integrated Delivery Leader for IBM AMS Business Application Modernization (BAM) practice. Biswajit is IBM India Competency Head for Global Specialized Application Modernization (AM) and Conversions & Migration (C&M) Competency. Biswajit has 17 years of multi-functional experience in the IT industry spanning across consulting, defining technology roadmap, solution architecting, offering incubation, technology innovation, solutioning, capacity and capability development, establishing practice based teams, large account management with P&L responsibility, customer/partner/alliance relationship management, heading delivery center of strategic business unit and leading global integrated delivery. Biswajit is experienced in using disciplines and tools of business process redesign, business architecture and technology innovation to help organizations realize their transformation objectives. Biswajit’s current responsibility includes growth of Application Portfolio Rationalization, Application Analysis and Implementation Roadmap, Business Rules Extraction, Porting, Conversion, Restructuring, Re-engineering, Consolidation, Web Enablement, Legacy to SOA and Cloud Modernization Business.
Debasis Roy Choudhuri is an IBM Certified Senior Architect and lead architect of Business Application Modernization, IBM Global delivery. He has 15 years Industry experience, encompassing a wide range of skill set, roles across the verticals. He is specialized in Application Infrastructure Design, Server Consolidation, Workload Migration from legacy system to various virtualized and cloud environment He has an extensive experience with architecture & design, customizations and implementation of various transformation solutions. He was involved several complex Data Center separation engagements since 2006. He is also Certification Review Board member for Architecture Review Board (India). He is currently working for cloud solution engagements for IBM ANZ clients.