Closing the Gap Between Apps and Ops
By Jake Sorofman

Agile techniques, dynamic programming languages, mashups, and frameworks are all part of the same trend of driving time and cost out of the application development process. The problem is that the speed and agility we’ve achieved in the domain of “apps” hasn’t crossed over into the domain of “ops,” which is where application value is realized.

This is because of a longstanding “deployment gap” that exists between apps and ops.

At its most basic level, this gap is a result of conflicting motivations. Apps is about speed, while ops is about control. Apps folks focus on delivering solutions to the lines of business as rapidly as possible. Ops folks focus on operating stability, compliance, and cost control through standardization and stringent change management. The gap between them can inhibit business responsiveness, delay deployments and cause organizations to miss opportunities. It can take months and sometimes even years to deploy an enterprise application — and this stands in the way of business value.

One of the most notable recent solutions to the pain of this deployment gap is software-as-a-service (SaaS), which closes the deployment gap by entirely eliminating the interaction between apps and ops. Unfortunately, the one-size-fits-all approach of SaaS lacks the flexibility of on-premise commercial and custom software solutions; and the idea of multi-tenancy still raises questions about security, privacy, and compliance.

What enterprises want is the simplicity and zero-latency consumption of the SaaS model, together with the flexibility and control of traditional software models — a combination of custom and commercial software, deployed into production without the onerous delay of today’s deployment gap.

The Freedom to Compute: Virtualization and Cloud

Application virtualization can close the deployment gap by separating applications from their operating infrastructure. The virtualized application contains just enough operating system (JeOS), databases, and middleware required to run the software in production. These bits travel with the application package and allow it to run as an image on any operating infrastructure without any manual setup, tuning, configuration or certification. Suddenly, the deployment gap disappears, applications are set free, and deployment cycles are compressed from months to minutes.

Cloud computing has emerged to bring together virtualization and SaaS — along with the notions of grid and utility computing — to create a flexible and scalable environment for the deployment of both custom and commercial applications for on-demand, pay-as-you-go access. Key to cloud computing is the notion of application virtualization, which allows applications to run in virtually any compute environment without any manual set up, configuration or tuning — just plug it in and it works. Releasing applications in this manner opens the door to embrace cloud computing by enabling applications to scale seamlessly from the internal datacenter to the external service provider.

Apps and Ops United: Combining Agility and Control

The reality is that new flexibility for one group always yields a loss of control for another. This is true for virtualization and cloud, which provides great flexibility for application development and lines of business, but often creates anxiety for the IT operations personnel responsible for quality of service and compliance and the architects responsible for the definition and enforcement of standards.

Virtualization and cloud require operations and architecture groups to think about repeatable practices, definition of corporate standards and policies for virtual machine consistency and quality, and patterns for collaboration between apps and ops. This requires an integrated lifecycle approach and new application architectures to support it.

The new architecture for virtualized applications sits between the traditional worlds of application lifecycle management (ALM) and enterprise system management (ESM), providing a bridge that dramatically accelerates production deployment, while also controlling the chaos, inconsistency, and sprawl resulting from the unmanaged proliferation of virtual machines. This architecture facilitates seamless collaboration between apps and ops, providing application development the speed and flexibility they want, while ensuring IT operations has the control and predictability they need.

In this new context, the definition and enforcement of policies becomes the bridge and the basis for trust between apps and ops, ensuring that virtualized applications are “correct by construction” — consistent and always conforming to operating standards before they’re deployed.

This type of application architecture is foundational to virtualization and cloud initiatives — it has crosscutting implications that must be addressed in the earliest stages of implementations. The reality is that those who delay an architectural approach to application virtualization will put initiatives at risk, unable to contain the sprawl and manage the chaos as demand and scale compound. The lesson is that you can’t simply “back into” this sort of management foundation over time — organizations who delay a lifecycle approach will find it too late to take corrective action once issues start to appear.

This approach closes the deployment gap, finally aligning apps and ops by combining speed and flexibility with control and predictability. It becomes the bridge between apps and ops — and the bridge between application innovation and the realization of business value.

Jake Sorofman is vice president of marketing for rPath, the pioneer and leader in technology for virtualizing software applications and managing the complete lifecycle of virtual appliances and application images for cloud and virtualized environments. Learn more about rPath at http://www.rpath.com, and contact Jake at jsorofman@rpath.com.