A consistent cloud development and deployment model

Problem

To build enterprise class solutions that leverage virtualization and cloud computing technologies, an organization’s existing solution development lifecycle (SDLC) must be aligned to an automated cloud deployment lifecycle, which supports repeatable processes, dynamic provisioning and capabilities to meter and charge-back resource consumption.

Solution

Define the organization’s cloud software development and deployment  lifecycle model (CDLC). Typically, this should include the cloud activities to be performed during the phases of design, testing and quality assurance (QA), staging, production and post production. Identify automation tools to support CDLC phases. Implement templates to standardize for consistency, repeatability, and reproducibility. Align the existing project management and SDLC phases to CDLC.

Click to Enlarge

Application

• Repeatable, reproducible cloud targeted application development and deployment
• Consumption-based metering and billing of infrastructure, applications and services

Impacts

• Alignment between various IT SDLC processes and CDLC
• Complexity of business solutions may require close collaboration between teams
• Solution level metering and billing could become tricky. CDLC activities may direct more attention towards optimizing cost rather than the quality of the business solutions

Anti-Patterns

• Traditional SDLC interlaced with project management controls
• Non-repeatable customized cloud deployment process on a project-by-project basis, most often by siloed teams

Related Patterns

• VM Blueprinting Pattern
• Platform VM Pattern
• Application Virtualization Pattern
• VM Factory Pattern
• Stateless VM Pattern

References

• Opscode Chef
• Puppet
• rPath
• ControlTier
• Func, Fabric, Jumpstart, Kickstart
• Eucalyptus, Xen, VMware

MomentumSI Solutions

• MomentumSI Platform Engineering Maturity Model (PEMM)