Within the cloud delivery models Platform as a Service (PaaS), the model providing the facilities for full life cycle application development, has a smaller overall market share than IaaS or SaaS. However, the PaaS growth rate is accelerating with Gartner predicting a 26% PaaS growth rate to reach $15 billion by the end of 2018.
Private PaaS, containerization and data-intensive applications are fueling the PaaS growth rate. PaaS service offerings have matured and many of the latest platform announcements by the large cloud vendors have shifted from IaaS to PaaS.
This post looks at PaaS benefits, a business case and potential workloads for PaaS. Perficient has partnerships with all of the leading cloud and pure-play PaaS vendors and as such we work on a great deal of PaaS implementations. These implementations vary from operations focused lift-and-shift migrations to greenfield digital product development.
Interviews with client’s and reviews of customer case studies where used to compile the following list of PaaS benefits:
• Significant application lifecycle velocity improvements – release engineering, deployment and life-cycle management
• Application portability across IaaS providers, private and hybrid cloud
• Application scalability and availability – auto-scaling, elastic infrastructure, operational services
• Automated provisioning and builds – CI/CD, DevOps
• Reduction of operating expense – a shift to open source, operational efficiencies, platform consolidation
• Application services marketplace – database, caching, security, logging, API management, etc.
• Consistency of architecture – Spring and Docker ecosystems, microservice patterns, etc.
• Security and compliance – automation, scans, rapid patches, stable-documented environment
• Simplified and centralized administration using containers and container orchestration
• Operational services – security, logging, dynamic routing, load balancing, etc.
• Application monitoring and availability management
Many PaaS case studies point to a business imperative to change the application development approach to achieve greater velocity and operational scalability. The well-known public case studies for Netflix and Google for example chronical their moves to technology such as cloud computing, DevOps, and containers. They, among others, invented new application and platform approaches to achieve the scale mandated by their business model.
But many companies now feel the threat of digital disruption or see the opportunity for digital products to increase revenue. We are seeing companies implement PaaS due to a business imperative to shift to digital products. In fact, Gartner says by 2022 PaaS will be used “in over 33% of new application designs oriented toward digital business solutions.”
Formal cost justification of PaaS solutions is often tied to specific products most often customer-facing digital products or platforms such as e-commerce. Systems-of-engagement benefit the most from improvements in development lifecycle velocity.
Systems-of-record typically change less frequently and are more problematic to move to PaaS. A PaaS purchase can be easily tied to a specific product since the pricing model is usually subscription or usage based. It’s easy to start small with greenfield applications and grow the platform as business value is added and measured.
Another common use case for PaaS is an operational focused lift-and-shift of legacy applications (e.g. a Java application server farm) to PaaS. These efforts focus on containerizing applications to retire server farms, software licenses and the costs associated with them. There are also tangible labor costs and outage reductions, and performance improvements that supports the business case for PaaS.
It is best to consider both software development (full-lifecycle) and operational benefits together for the business case and also, more importantly, for the platform architecture requirements. PaaS efforts lead by operational teams often fail to fully consider product development requirements.
Operations lead PaaS migrations can lean towards a container orchestration solution versus one that also encompasses application services (such as CI/CD, database, cache and API management). A collaborative approach to migrating applications to PaaS leads to a more comprehensive solution and encourages a culture shift to a DevOps operational style.
It is often desirable to change some application level components during brownfield application migration such as moving from commercial to open source software, which may require application changes. These costs must be factored into the cost-benefit analysis.
There are many less obvious benefits we have seen with PaaS. For example, security is improved by automation (such as scans during deploy), platform consistency and the ability to rapidly patch vulnerabilities. Platform consistency can also greatly reduce the time spent on compliancy for regulated companies and government agencies. Containerized software and open source allows software to be more readily shared across business entities and government agencies.
We have seen improvements in engineering staff capabilities and reductions in turnover attributed to PaaS platforms. And, PaaS can improve availability promoting a shift in staff focus from operational administration to proactive application health management.
Perficient has partnerships with all the major PaaS vendors and provides services for PaaS migrations from planning and architecture to implementation should you need any help.