Cloud orchestration and the future of DevOps

Posted by admin May 20, 2015
Cloud orchestration and the future of DevOps

DevOps has been colorfully described as "the industrialization of IT," with an effect on resource provisioning and organizational culture comparable to that of the assembly line on automobile manufacturing in the early 20th century. DevOps practices theoretically help IT teams increase productive collaboration and move them forward in areas like agile development, continuous integration, delivery and deployment.

Accordingly, they can iterate products and services both quickly and at scale. Chef Software CTO Adam Jacob neatly summed up the ideal result of implementing DevOps at this year's ChefCon, when he talked about "high velocity, high quality output" that, he argued, most organizations were still unfamiliar with. DevOps is undoubtedly challenging from both a technical and cultural standpoint.

In the current context of having to handle potentially thousands of updates each day and survive the ongoing pressure to push new products and services to market, a cloud orchestration solution, like QualiSystems CloudShell, has become essential for sustaining DevOps. At the same time, DevOps is not something that comes in a box or even a software package. There's the underlying need to break down silos between application developers and IT operations if complex, dynamic environments are to mesh properly with deployed software.

"Developers and ops folks are used to being siloed," MRE Consulting CIO Ken Piddington told CA Technologies' Rewrite blog. "Now you're trying to get them on the same page. Whoever is leading a DevOps charge has to help these groups understand that if they work together, the whole organization wins."

DevOps was conceived as a cultural collaboration movement to help accelerate development. DevOps was conceived as a cultural collaboration movement to help accelerate development.

A look at orchestration for DevOps: The need for for more sustainable processes
The rise of software-defined everything, most notably (at least in terms of media attention) SDN and NFV, has opened up the possibility of real infrastructure automation on the road to DevOps. Still, the ideal of sustainably automated infrastructure is a far cry from reality, even within many organizations actively seeking to become more agile.

The number of enterprises planning to adopt DevOps within the next five years climbed steadily from 2013 to 2014, according to one survey. But these plans could run into headwinds such as inefficient manual processes, fragile script-based automation and the general persistence of legacy and physical infrastructure.

"A more sustainable path to DevOps is needed."

In his presentation about DevOps as IT industrialization and the growing importance of orchestration to DevOps, GigaSpaces CTO Nati Shalom contrasted existing approaches to infrastructure management to the ideal of orchestration. Practices that depend upon humans (whether Dev or Ops) are both error-prone and time-consuming:

  • Shalom estimated that four-fifths of outages pertaining to mission-critical services are caused by "people and process issues."
  • Moreover, half of those outages are rooted in configuration and change issues, or botched handoffs.
  • To add to these issues, existing automation is often code-centric, meaning that delays naturally arise when teams have insufficient numbers of programmers.

A more sustainable path to DevOps is needed, one that takes into account all types of infrastructure and is also amenable to the non-programmers (e.g.,vendor-trained IT specialists) that make up the lion-share of IT teams. Many enterprises (61 percent, according to Shalom's numbers) are still at an early stage in terms of laying a foundation for DevOps, with some degree of standardization, consolidation and virtualization already in place but automation and orchestration still in the distance. More than 80 percent are looking to automation and orchestration as ways to improve end user experience and accelerate time to market.

Supporting a DevOps culture ultimately requires an orchestration platform that can handle the provisioning of compute, storage, networking, hypervisors and many other components, on both the physical and virtual fronts, in diverse infrastructure combinations and topologies. CloudShell excels at this task, with its object-oriented architecture and visual tools that enable non-programmers to build and maintain automation effectively. In addition, it includes multi-tenant management such a reservation and scheduling system that helps development, testing and other teams effectively share infrastructure, along with integrated reporting and business intelligence.

The takeaway: DevOps was conceived as a way for teams to apply collaboration along with agile development principles to create an industrialized process for high-quality IT and software outcomes. It ideally removes intra-organizational silos and accelerates time to market. However, it requires significant cultural as well as technical change. Infrastructure orchestration and automation enables DevOps through the establishment of sustainable automation processes, effective provisioning within complex data center and lab environments and support for programmers and non-programmers alike.