We were recently approached by a large retailer who needed help with their application deployments. One of the options that came up was manifest application deployment, whereby the application would be paired with an XML manifest that describes the application resources, payload, and assemblies. Since this manifest is a kept in the companies source control all changes are tracked and version-ed. Various stake-holders of the application are responsible for keeping the manifest intact and valid, including parameters for data file locations, database queries that need to get executed during deployment, and services that need to get installed and updated. An example manifest would look something like this:
Read More
Patrick McKenzie, while playing Bingo with Devops, documents his experience and insight while working for a Japanese mega-corporation. Familiar with the too well known bubble gum, duct tape, and praying methodology, Patrick explains that without engineering being an integral part of the organization you will not likely know much about the other side.
Read More
Using Application Ratings to Foster DevOps
Last week I had the opportunity to have a hallway conversation with a CTO of a large software company about how the interactions between development and operations teams are changing. I asked his opinion on measuring the benefit of that collaboration. He noted while they did not have direct metrics for their teams, they found that having a rating system for applications under development did effect how developers approached their projects.
Their rating system incorporated the expected operational characteristics of an application. For example, applications that must be as resilient as possible had a different set of ratings than applications with asynchronous or long running transactions. He said that once their developers understood the rating system, operational requirements were considered more throughly during the design phase. Additionally, they were more receptive to using existing operational knowledge as they were building the application. The system helped operations teams because it clearly prioritized the business importance of specific operational characteristics, which enabled better proactive problem analysis and prevention.
read more at sys-con.com site
